My grandmother was one of four wives in a Yoruba household. She started by hawking before she eventually secured a stall of her own. From the proceeds of that stall, she raised her children, including my father, the eldest across all four wives.

My father went on to become professionally successful, in a way that gave his own children a materially different starting point from the one he had. That success extended sideways too, helping several of his siblings toward better circumstances of their own. Further down the same line, one member of the family is now senior in the US software industry. None of that was mechanically caused by one trading stall. But the surplus from that stall financed something, and what it financed compounded well past the size of the original business.

She is one version of a pattern repeated across Nigeria under different names: hawker, trader, POS agent, WhatsApp merchant.

Ask a conventional lender what two decades of that kind of operating history is worth, and the honest answer may still be: we cannot price it.

No audited accounts. Limited conventional collateral. Thin formal credit history. Business and household cash flows intertwined.

Twenty years of demonstrated capital management can still arrive at an institutional credit committee looking surprisingly close to a blank page.

Some of that intertwining looks, from a lender’s vantage, like a weakness. Household withdrawals reduce the cash available for the next inventory cycle and complicate any assessment of debt capacity. But viewed on a longer horizon, a school-fee payment is not simply lost working capital. It may also be an investment in human capital whose economic return appears in a child’s earning capacity years later rather than in next quarter’s turnover. The same cash flow looks different depending on which unit of analysis is applied to it.

That gap is important because it reveals something about what finance recognises as evidence.

Nigeria has roughly 40 million MSMEs, the large majority informal. Women own or lead a substantial share, particularly across trade, food and other microenterprise activities. Within that population are businesses that have survived for years and businesses that will disappear within months. There are capable operators and poor ones, scalable enterprises and businesses that will remain small regardless of how much financing becomes available.

Competence and scalability are different variables.

So the interesting question was never really whether she should be celebrated as an entrepreneur. It is whether the financial system can reliably distinguish operators like her from everybody else.

The Record That Existed But Couldn’t Be Read

Historically, that was difficult.

Her operating record existed, but much of it was fragmented across cash transactions, handwritten ledgers, suppliers, cooperatives, trade associations and counterparties. A supplier extending sixty days’ credit knew something about her reliability. A trade association knew whether she paid her dues and showed up when the market needed collective bargaining with a landlord or a local government levy officer. An ajo contribution history knew whether she honoured commitments under pressure. None of that information was economically meaningless. It was simply scattered across people and institutions that had no reason or means to standardise it, verify it centrally, or feed it into anything resembling a credit model.

A bank cannot easily underwrite a memory distributed across people who never speak to each other.

Legible, Not More Competent

Digitisation is beginning to change that.

POS transactions, bank transfers, mobile money, supplier payments and digital-commerce activity increasingly leave persistent records. A merchant processing hundreds of transactions can now generate observable evidence of revenue consistency, seasonality, growth and repayment behaviour without producing an audited financial statement.

This creates an important distinction.

Informality no longer necessarily means opacity.

A business can remain organisationally informal while becoming increasingly financially observable.

Nigerian fintechs are already exploiting that distinction. Moniepoint, for example, uses transaction and account behaviour to extend credit to merchants that conventional underwriting may struggle to assess. Its lending activity provides evidence of the broader mechanism: digital operating behaviour can be converted into underwriting signals.

Digitisation made more of that behaviour legible. Legibility creates the possibility of pricing risk.

The first opportunity is better working-capital financing for businesses that may have been commercially proven but institutionally difficult to assess.

Share

The Portfolio Is the Asset

A larger capital-markets question sits behind it.

Consider an illustrative ₦1 million facility to one merchant. For a large institutional investor, that exposure is economically insignificant.

Now consider 10,000 similarly underwritten facilities.

That is a ₦10 billion portfolio.

The individual entrepreneur may be too small to justify institutional attention. A sufficiently large, diversified pool of exposures to entrepreneurs like her is a different financial object.

This changes how we should think about the fintech or payment platform originating the loans. The platform is not only a lender or distribution channel. It can also become a manufacturing engine for another financial asset: a diversified portfolio produced from thousands of individually small credit decisions.

The potential chain looks something like this:

merchant activity → behavioural data → underwriting → credit → repayment history → diversified receivables portfolio → warehouse funding → institutional takeout

Eventually, sufficiently mature portfolios might even support securitisation.

Nigeria is not there yet.

The country has a legal framework for securitisation and structured-finance precedents in other asset categories. But institutional-scale securitisation of alternative-data merchant and MSME credit remains underdeveloped.

The constraint is not simply capital.

Institutional investors need assets they can diligence, compare, monitor and price.

That requires consistent underwriting standards, several years of performance data across different economic conditions, predictable servicing, credible recovery processes, appropriate credit enhancement and legal structures capable of separating the portfolio from the financial condition of its originator.

Those functions are still developing.

India’s microfinance market provides one indication of where the path can lead. Granular loans to individually small borrowers have been aggregated into structured portfolios capable of attracting larger pools of capital. Nigeria’s market, regulatory environment and borrower characteristics are different, so the precedent does not establish that the same outcome will follow here. It establishes that individually tiny exposures do not have to remain institutionally irrelevant forever.

For decades, informal operators like her may have been generating commercially useful information that formal finance could barely process. Increasing digitisation means more of that information now has a format. Fintech underwriting can turn the format into a risk signal. Repeated lending can turn those signals into performance histories. At sufficient scale, those histories can begin to support portfolios that entirely different pools of capital can evaluate.

The first opportunity is therefore obvious: finance good informal businesses better.

The larger opportunity is to complete the financial machinery that allows thousands of those individually small exposures to be underwritten consistently, aggregated and eventually priced as one institutional asset.

Nigeria’s market woman may never have been too small for institutional capital.

The unit of analysis was.

Drake’s deal with Universal Music Group, completed in 2021 and confirmed by UMG’s own chairman on a 2022 earnings call, was described as an “expansive, multi-faceted” arrangement covering recordings, publishing, merchandise, and visual media. Industry estimates put the figure north of $400 million. The exact number was never confirmed, and the industry didn’t wait for it. Nobody needed a number that size to be exact to treat it as proof of value, the biggest bag an artist can command, the scoreboard settled in the artist’s favor.

The size of a check doesn’t tell you who won the negotiation. It tells you how much was sold.

A $400 million deal does not necessarily create $400 million of new wealth. It may simply convert part of an artist’s future wealth into cash today. If the rights exchanged are worth more than the check, the headline number can rise while the artist’s economic position gets worse. The relevant metric is therefore not deal value. It is value retained.

Drake is not merely a compelling case study. He is a credible contender for the most commercially dominant recording artist of the streaming era, the first artist to hold the top three positions on the Billboard 200 simultaneously in the chart’s seventy-year history, and his counterparty is UMG, the largest music company in the world. That combination, extreme demonstrated demand meeting the largest label balance sheet there is, makes his situation an unusually clean live test of how much bargaining power an artist at the outer edge of scale can actually convert into ownership and better financing terms.

The Bundle, Split in Two

The prior piece in this series established that a major-label deal has never priced one thing. It bundles capital, distribution, marketing, and a claim on long-duration rights, and two of those four components, distribution and marketing, have been commoditizing for a decade. Any artist can reach every streaming platform for an annual fee. Audience-owned channels now move attention that a label’s promotional budget can’t easily match, and even the labels concede the point. A music-technology executive told The Hollywood Reporter in 2025 that the majors are evolving “more toward services companies.”

What’s left unbundled, and unexamined, are the other two pieces, capital and the label’s willingness to absorb the risk that the music simply doesn’t work. That second piece is worth assessing, because it isn’t really about marketing or distribution. It’s insurance.

A label advance guarantees an artist money regardless of what happens next. If the album flops, the artist keeps the cheque and the label eats the loss. In exchange, the label receives a claim on the artist’s rights that runs for years, sometimes decades, and that claim doesn’t shrink if the ‘flop-risk’ it was intended to cover never materializes. For a young artist with no other income, no catalog, and a real chance the whole thing goes nowhere, that trade can be rational. It’s the same logic behind why anyone buys insurance priced above its statistically fair value: when you can’t absorb the loss yourself, certainty is worth paying for even at an unfavorable price.

The open question is what happens to that trade once the risk it was pricing has fallen materially, while the price hasn’t moved with it.

What Taylor and Chord Together Prove

Two developments outside Drake’s own situation now let that question be asked with more precision than it could have been asked five years ago.

The first is Taylor Swift’s move to Republic in 2018. Her early deal with Big Machine, signed at fifteen, surrendered the masters to her first six albums in exchange for the standard early-career bundle, money, promotion, infrastructure she had no way to build herself. By the time she renegotiated as a global superstar with a demonstrated, repeatable audience, she stayed inside the major-label system on different terms: she would own every master she recorded going forward. This doesn’t prove that superstar leverage alone caused the shift. Universal’s desire to sign her, the broader evolution of artist bargaining power, and her own experience with Big Machine all played a part too. What it proves is more precise and useful. Needing a major label no longer has to mean giving it your masters.

The second development is what’s happened to the market for seasoned royalty income itself. In April 2026, Chord Music Partners, a platform tied to Universal through Dundee Partners, borrowed $500 million against royalty income from a large, diversified portfolio of songs, paying investors an interest rate of 5.56%, the tightest pricing on record for this kind of deal. A ratings agency, KBRA, has now rated close to $13 billion of these royalty-backed loans since 2020, and the number of companies doing this kind of deal has doubled since 2023. None of that tells you what it would cost to borrow against one artist’s future output alone. Lending against a single artist is a narrower, riskier bet than lending against thousands of songs from many artists, and it should cost more. But it does tell you that a seasoned, proven income stream now has a real, quoted price attached to it when used as collateral. Owning the rights and borrowing against them are no longer the same decision.

Put together, those two facts describe an artist enterprise where more of the bundle can now be separately priced than at any previous point in the modern music business. Whether it makes sense to keep buying all of it together is now more of a math problem than a loyalty question.

The Point Where the Insurance Stops Making Sense

Here’s the shape of that math problem, stripped down to numbers simple enough to hold in your head, and stated up front as a thought experiment rather than an attempt to reconstruct anyone’s actual contract terms.

An artist needs $100 million today. There are two ways to get it.

Under Structure A, the bundled path, a label pays the $100 million now, guaranteed, and takes half of whatever the artist’s future rights turn out to be worth. If the music flops, the label absorbs the loss and the artist keeps the cash. If the music succeeds, the label’s share rises right along with it. There is no ceiling on what half of a runaway outcome is worth.

Under Structure B, the unbundled path, the artist borrows the $100 million against seasoned catalog income already coming in, at an interest rate that has to sit meaningfully above the 5.56% a diversified catalog can borrow at, because lending against one artist alone is a narrower, riskier bet than lending against a large, diversified portfolio of songs. Call it something in the high single digits. The artist also pays separately for the distribution and marketing a label would otherwise have supplied, whether as a flat fee or a negotiated revenue share. The important difference is that these costs can be contractually bounded. Debt carries a defined repayment obligation. Services can be purchased for fees or a capped participation that doesn’t hand the provider a permanent claim on the underlying rights. Their economics don’t have to compound indefinitely alongside the value of the artist’s rights the way an equity-style claim does.

The two structures cross at a specific point. If the expected value of the artist’s future rights is modest, half of it is a small number, smaller than years of debt service plus service fees. The bundle wins there, and buying the insurance is the right call. But as the expected value of those future rights grows, half of it grows right alongside it, while the debt and the service costs stay within the bounds set at the start. There is a level of demonstrated, repeatable commercial output past which giving up half of an increasingly large number costs more than paying a bounded price for the pieces you actually still need.

This is the entire point, and it doesn’t require knowing what Drake’s contract says. A label sells a bounded amount of capital in exchange for a claim on an unbounded outcome. That trade is easiest to justify when the outcome is genuinely uncertain. It becomes progressively harder to justify as the expected value of the rights being surrendered rises and alternative financing becomes available.

The Honest Counterargument

UMG has priced Drake before. Its executives are not naive about audience data, streaming payouts, or what an artist with his track record is capable of producing. The size of a reported $400 million figure could just as easily be read as evidence the market already repriced his risk efficiently, rather than evidence of anything left on the table. It’s possible there is no mispricing here at all, and that a bundled deal at superstar scale already reflects everything this essay has just laid out.

There’s no way to settle that from outside the contract, and pretending otherwise would be the kind of overclaim this essay is trying to avoid. What can be said with more confidence is narrower and still matters. Five years ago, that external benchmark was far thinner. Now seasoned royalty income has an increasingly observable institutional price. Services have an observable, separately purchasable cost. The bundle isn’t automatically wrong at any given size. It’s testable in a way it wasn’t before, and that changes the information available in the negotiation, whether or not it changes the final number.

Drake’s own contract status remains publicly unclear as of this writing. That uncertainty is beside the point here on purpose. The argument doesn’t depend on how his situation resolves. It would hold whether he renews with UMG next month or walks away entirely, because the question was never really about Drake’s next cheque. It’s about whether the market now carries enough information to ask, of any artist at his scale, whether the policy still fits the risk.

The logic isn’t unique to music. Startups sell equity when revenue is uncertain and they have no collateral to borrow against, and that’s rational early and expensive later, which is why mature companies lean on retained cash flow and debt instead of repeatedly selling ownership once they’re able to. An emerging artist surrenders long-duration rights for the same reason a young company sells equity: uncertain output, no financing alternative. The real milestone isn’t independence. It’s creditworthiness, the point at which an asset that once needed someone else’s balance sheet develops one of its own.

The most sophisticated version of Drake’s next move might not be independence at all. It might be re-signing with UMG, on terms that price distribution, marketing, capital, and risk separately rather than as one bundled premium, because the label, priced correctly, may still be the cheapest and most capable counterparty for at least some of what he needs. That outcome and the exit outcome are not opposites. They’re both downstream of the same question: not how big is the cheque, but what did it cost to get it.

The artist became the platform. Whether the platform still needs the insurance it’s paying for is a number someone should now be able to check.

This follows an earlier piece, The Artist Has Become the Platform, on why the asset needed pricing in the first place.

The Artist Has Become The Platform

The AI economy is usually discussed as a race to build models. That framing almost guarantees Africa loses.

Frontier models require enormous amounts of compute, reliable power, specialised chips, deep technical talent and billions of dollars of patient capital. Those advantages are already concentrated in a handful of American and Chinese companies and the supply chains, cloud platforms and research pipelines built around them. Africa is unlikely to win by trying to recreate Silicon Valley’s AI stack locally.

There may be another way in.

Compute is concentrated. Data is distributed

As AI models become more capable, increasingly important inputs include information they cannot easily obtain from the public internet: proprietary workflows, specialist knowledge, underrepresented languages, real-world operating data, human-generated code, cultural material and high-quality examples against which their performance can be tested.

Africa has a lot of that. Sub-Saharan Africa alone moved more than $1.4 trillion through mobile money in 2025, roughly two-thirds of the world’s total mobile money transaction value, according to GSMA’s latest industry report. Behind that aggregate are continuously generated records of financial behaviour across markets where mobile money plays a role that conventional banking data alone cannot fully describe.

Africa has no monopoly on this opportunity. India, Brazil and other emerging markets have large technology workforces, substantial local datasets and established data-services industries already positioned to serve many of the same buyers. A model builder indifferent to which market supplies the relevant signal will buy from whoever offers the best combination of quality, price and speed. What resists substitution is the specific texture of a market: the way Lagos traders code-switch between languages, the rhythm of a Kenyan chama savings cycle, the particular shape of a naira devaluation working through a small business’s books. Generic “emerging-market” data can become a commodity. Precisely located operating knowledge is harder to substitute.

The internet gave AI its first corpus

The first generation of large language models benefited from an extraordinary historical accident. Humanity had spent decades putting books, websites, software, discussions, images and other knowledge online. Model developers could ingest enormous amounts of it at relatively low marginal cost.

The next frontier is harder. AI systems increasingly need to reason reliably, perform specialised tasks and operate inside particular professions, industries, languages and real-world environments. More generic internet data does not necessarily solve those problems.

A model may understand the textbook principles of credit and still struggle with lending behaviour in an economy experiencing severe currency volatility. It may speak English and still misunderstand Nigerian English. It may understand the engineering of an electricity network while having little exposure to how businesses actually operate around unreliable supply. It may generate something labelled “Afrobeats” without understanding the musical, linguistic and cultural structures that distinguish it.

These are information gaps, and information gaps can become economic assets.

Many of the characteristics that make African markets difficult to operate in also make them informationally unusual: informal commerce, multilingual communication, mobile-money-heavy financial systems, volatile currencies, legacy technology, intermittent infrastructure, distinctive cultural systems and regulatory environments that do not map neatly onto Western equivalents.

A frontier AI company can buy more compute. It cannot instantly manufacture twenty years of operating experience in an African market. It cannot recreate genuine Yoruba conversations by translating English. Nor can it synthetically invent the historical record of how businesses adapted to infrastructure failures, currency shocks or informal distribution systems and assume that the result represents reality.

This becomes more important as AI moves from knowing about the world to acting within it. An AI system performing work in Lagos, Nairobi or Accra eventually has to understand the realities of Lagos, Nairobi and Accra.

Some of the information generated by those environments may therefore acquire economic value precisely because it is difficult to reproduce elsewhere.

But possessing valuable information and capturing its economic value are very different things.

Africa has seen this movie before.

The worst outcome would be another raw-material economy

Across many of its most important global value chains, Africa has followed a familiar pattern: extract something valuable, export it in relatively raw form, allow somebody elsewhere to refine it, manufacture around it, control distribution and own the customer, then import the higher-value product.

Oil is the obvious example. Cocoa-producing economies capture only part of the value ultimately embedded in branded chocolate. Mineral-rich countries export ores while much greater value can accumulate in processing, battery materials, components and finished technologies.

The recurring problem sits in where value-chain ownership, transformation and bargaining power land, not in whether the raw material was valuable to begin with.

AI could reproduce that structure with information. African companies provide the data. African workers label it. Foreign intermediaries aggregate and structure it. Technology companies convert it into better models. Those models are then sold globally, including back into African markets.

The commodity has changed. The economics of extraction have not.

If that becomes Africa’s principal role in AI, we will have digitised the commodity-export model rather than escaped it.

That is why simply declaring that “African data is valuable” misses the point. Data is only the beginning of the value chain.

From raw data to AI infrastructure

Raw data is frequently messy, legally uncertain, technically unusable and commercially difficult to evaluate. Where the underlying information is personal rather than institutional (payments, health or behavioural data, for example), privacy and data-protection regimes also constrain how it can be processed, transferred and commercialised. The rights question therefore extends beyond who possesses the data to what they are legally permitted to do with it. Turning raw information into an AI asset looks something like:

Raw Information → Rights → Curation → Machine Readiness → Demonstrated Utility → AI Asset

Each stage can increase usefulness, scarcity and bargaining power. Information whose ownership cannot be established may have little commercial value. The same information with clear rights, appropriate privacy treatment, useful metadata and demonstrated relevance to a model capability is a different asset.

But the progression does not necessarily end with a cleaned dataset.

Proprietary information can potentially support benchmarks, evaluation suites, specialist reference systems, preference datasets and other tools that measure or improve how AI performs within a particular environment. A business might simply license years of information about how an industry operates. Or that information, combined with domain expertise, might be used to create a benchmark for determining whether an AI system can actually perform important tasks within that industry.

The first monetises information. The second begins converting information into AI performance infrastructure.

Music offers an intuitive example. A catalogue has historically been valued primarily around human consumption: streams, performances, synchronisation, mechanicals and other royalty-generating uses. Generative AI creates potential additional demand for rights-cleared musical information. But the opportunity need not stop at licensing songs. Music and the expertise surrounding it can potentially contribute to datasets, evaluation tools and other systems for determining whether models actually understand particular genres, languages or musical structures.

South Africa’s Lelapa AI has already built a working version of this move, on a small scale. Its Esethu Framework — developed with transcription company Way With Words and a University of Pretoria research group — gives local language communities formal governance over how their speech and text data is used, while its licensing model is designed to reinvest commercial value into further dataset creation rather than simply extracting the current one. The first release, an open-source isiXhosa speech corpus, offers an early example of the raw-data-to-AI-infrastructure move already being attempted, rather than merely theorised.

The same principle can apply to software, professional knowledge, healthcare, agriculture, finance, industrial processes and other information-rich domains.

That suggests a progression:

Raw Data → AI Inputs → AI Performance Infrastructure

The further an economy moves along that chain, the greater its potential to capture something resembling intellectual-property rents rather than commodity margins.

Follow the bottleneck

There is an important reason not to build this thesis around today’s demand for proprietary datasets: technology changes where scarcity sits.

Synthetic data may replace some human-generated datasets. Models may become more data-efficient. Information commanding a premium today may eventually become commoditised.

The opportunity does not necessarily disappear. The bottleneck moves.

If raw data becomes easier to obtain, reliable provenance becomes more valuable. If synthetic examples become abundant, trustworthy evaluation against reality becomes more important. If general models become commoditised, domain knowledge that determines whether they work inside specialised environments can become the constraint. As AI systems perform more real-world work, continuously generated feedback from that work may become more useful than historical training corpora.

The strategic objective is therefore not to identify one permanently scarce form of data. It is to understand where the information bottleneck is moving and own assets on the scarce side of it.

This also means not all African data is valuable. Most data is not inherently valuable, and rarity alone proves little. A commercially meaningful AI asset needs some combination of four characteristics: scarcity, because equivalent information is difficult to obtain elsewhere; utility, because it materially improves or tests a valuable AI capability; control, because someone possesses sufficient rights to permit the relevant use; and defensibility, because competitors cannot cheaply manufacture an adequate substitute.

Quality, freshness, longitudinal depth and domain expertise can strengthen those characteristics. Some African datasets will satisfy them. Many will not. Some will be valuable but legally unusable; others unique but irrelevant to model performance. The opportunity requires asset selection, not a continental rush to monetise databases.

And the strongest assets may ultimately not be databases at all.

A static dataset is finite. A system that continuously generates difficult-to-replicate information is different. Businesses, professional networks, creative communities and institutions produce new domain knowledge through ordinary activity every day. Their strategic asset may eventually be less the historical database than the continuing information supply chain.

That changes the economics from a one-off data sale towards something closer to recurring intellectual property.

The eventual winners may therefore not be whoever possesses the largest dataset today. They may be whoever controls systems that continually convert proprietary human activity into useful machine intelligence.

Rethinking Africa’s place in the AI stack

The conventional AI stack runs roughly:

Chips → Compute → Models → Applications

That is useful for understanding where capital is concentrated. It is incomplete for understanding where countries outside that concentration might participate.

Put another chain alongside it:

Proprietary Information → AI Inputs → AI Performance Infrastructure

Now the geography changes.

Compute rewards enormous capital investment and physical concentration. Information originates everywhere. The strategic question is whether its owners recognise its value early enough, and control enough of its transformation, to capture it.

This also complicates the increasingly popular idea of “sovereign AI.” Economic sovereignty does not necessarily require every country to build a frontier model. For many countries, that would be an extraordinarily expensive way to enter a race whose leaders already possess massive scale advantages.

Sovereignty can also mean controlling scarce assets on which other people’s systems depend.

For Africa, the objective should therefore be clear:

Do not export African data as raw material. Productise it before it leaves.

That does not mean hoarding information or pursuing digital protectionism. It means understanding rights before commercialisation, building the technical and institutional capabilities that make information more valuable, retaining ownership where appropriate, moving towards higher-value AI inputs where the economics support it, and preserving participation in the recurring value generated from continuously refreshed information.

Africa does not need to own every layer of the AI economy. It does need to become much more deliberate about which layers it allows others to own around African assets.

Technology waves create wealth across different parts of their value chains. The strategic task is to identify where a country’s actual comparative advantages sit rather than imitate the layers where others already possess overwhelming advantages.

For Africa, one candidate is becoming increasingly interesting: the proprietary information required to make artificial intelligence genuinely understand and operate across the rest of the world.

Africa has spent decades generating that information without thinking of it as an economic asset. AI may finally give some of it a price.

But price alone is not the objective. Africa already knows how the story ends when it owns the raw material while somebody else owns the refinery.

This time, the opportunity is to own more of the refinery too.

An origination exercise for an international music-rights investor put me in a room with an operator who has spent two decades inside African music deals. I laid out the thesis I’d carried in: find catalogues that still sit directly with the rights-holder, negotiate direct, avoid the layers. He let me finish. Most of the catalogue worth having, he said, isn’t sitting untouched anymore. Distributors and label-services companies have spent years advancing money to commercially interesting artists and independent labels, and in exchange they’ve secured a position on the revenue those catalogues throw off. His point was that the clean deal, the one where you find an undercapitalised rights-holder and negotiate a straightforward acquisition, may be less available than the standard origination thesis assumes.

That’s one operator’s read on the market. It also turns out to describe, fairly accurately, how a meaningful part of the market actually works.

What’s actually happening to the catalogue

Nigerian industry reporting describes the pattern plainly. Label-services firms offer the artist an advance and license the artist’s music for an agreed number of years, while recouping through streaming revenue and possibly publishing, and a number of them have started looking into quasi-360 deals that give them a better chance at recouping their outlay through sync, neighbouring rights, performance royalties, mechanical and more. The firms doing this in the Nigerian market alone run from local operators to Platoon, EMPIRE Distribution, OneRPM, Vydia, Believe, Ditto, Africori and more. These arrangements aren’t limited to individual performers. Distributors and label-services companies contract with independent labels on similar terms, advancing against a slate rather than a single artist.

This isn’t an obscure financing structure. It has become an important part of how capital reaches independent artists and labels in Nigeria.

It’s worth being precise about what these arrangements do and don’t do. They don’t necessarily transfer the underlying copyright. The relevant master may remain with the artist or the independent label that signed it, while composition rights sit with a publisher, an administrator, or the writer directly, under a separate agreement entirely. What changes is the contractual architecture around the cash flow: who collects it, who participates in it, what gets recouped first, and, in broader arrangements, which additional revenue streams fall within the intermediary’s economics. Hear “intermediated” and assume “acquired,” and the diligence is compromised before it starts.

Commercially interesting African music can now sit behind a distributor or label-services company that has already advanced capital and holds a contractual position around the resulting revenues, regardless of whether the underlying rights sit with the artist, an independent label, or another rights-owning entity. The catalogue hasn’t necessarily disappeared. The route to its economics has changed.

Why that changes what you’re actually buying

A standard catalogue-acquisition model assumes ownership, economic participation, and practical control of the cash flow all sit in one place. In an intermediated market they don’t. The copyright might sit with the rights-holder. Economic participation in the revenue might sit partly with the distributor, under a recoupment schedule with years left to run. Collection, who the DSP actually pays, might sit with a third party. Exploitation authority, who can license the catalogue for sync or a new use, might sit with a fourth.

Skip past that distinction and the risk isn’t just overpaying. It’s misidentifying the asset. You can spend months negotiating what you assume is a clean catalogue acquisition and discover, near signing, that the copyright you can acquire remains subject to someone else’s contractual economics for years to come, or that the most attractive cash-flow interest in the deal sits with an intermediary whose position you cannot acquire without its consent.

What you’re actually screening for

The instinct is to open diligence by asking who owns the copyright. That skips a prior question: what exactly is the investable asset. A master copyright, a composition, a royalty interest, a contractual economic participation, a receivable, some other contractual right, none of these are interchangeable, and a single catalogue can contain several of them stacked on top of one another.

From there: who owns it. Who else participates economically in the revenue, and on what terms. Who controls collection, meaning who actually receives and accounts for the money. Who controls exploitation, meaning who can make the licensing decisions that generate new revenue in the first place. What’s transferable, and whose consent that transfer requires. For how long, including termination, reversion, and renewal. And what has to be recouped or discharged before the buyer’s own economics begin.

Run a catalogue through that sequence and you learn quickly whether you’re looking at a clean acquisition, an income stream layered under someone else’s advance, or a position that can’t move at all without a third party’s sign-off.

Where the capital actually goes

Once the catalogue and its economics stop being the same thing by default, the opportunity separates into three levels.

Rights. Acquiring an identifiable copyright, royalty interest, or other transferable economic interest directly. This is the deal everyone pitches, and the one the standard model was built for. It’s also a pool that gets smaller once existing distribution, label-services, and financing arrangements across a given catalogue are properly mapped.

Portfolio. A distributor holding a hundred advances against a hundred rights-holders has something that starts to resemble a portfolio of contract-backed cash flows recouping on different timelines. Whether that’s financeable, through a portfolio facility, a forward-flow arrangement, or an outright purchase of the position, comes down to the less glamorous questions: assignment rights, remaining term, revenue volatility, who controls the collection account, concentration across artists and labels, recoupment status, and what happens to the position on termination or a change of control at the distributor. Where the underlying rights and contract terms allow it, this is a distinct financing trade in its own right. Where they don’t, the size of the aggregate cash flow stops mattering.

Platform. Investing in, or acquiring, the institution that repeatedly originates, administers, and monetises these positions, rather than any single catalogue or portfolio it currently holds.

These are not three ways to finance the same asset. They’re three different underwriting propositions. Rights requires conviction in the asset. Portfolio requires conviction in the cash flows. Platform requires conviction in the machine.

A platform doing this repeatedly can also accumulate something a buyer of isolated catalogues never sees: which genres and artist profiles actually recoup, how quickly revenue arrives and decays, where consumption concentrates geographically, how advance size correlates with eventual return. Over enough transactions, that becomes underwriting data an outside acquirer analysing isolated catalogues may struggle to replicate. The machine doesn’t just originate assets. It learns from originating them.

The clearest public evidence sits at the platform level

Warner Music invested in Africori, then Africa’s largest independent distributor, in 2020, took a majority stake in 2022, and completed its acquisition on 12 February 2025, by which point Africori represented more than 7,000 artists. At the 2020 investment, Africori had worked with more than 6,500 artists and 700 labels; by the 2022 stake, roughly 7,000 artists and 850 clients.

Warner described the original investment as giving it access to Africori’s largest catalog and A&R network. The significance wasn’t simply exposure to existing music. It was access to the infrastructure and relationships through which more of it could be found and distributed.

What Warner acquired, in other words, was the company sitting across thousands of artist and label relationships, along with its distribution infrastructure, its rights-management operation, and the A&R capability that keeps finding the next act worth advancing. That’s the platform trade, playing out in public, at a scale most catalogue-by-catalogue origination strategies would take years to approximate.

The paradox underneath it

The same intermediary layer that can make a pristine catalogue acquisition harder to find can also make African music cash flows more legible to institutional capital. A distributor with a working recoupment schedule, reasonably clean metadata, and an actual collection track record has already produced documentation an institutional investor would otherwise have to build from nothing, in a market where historical paperwork is often incomplete and cross-border licensing stays genuinely fragmented. Payment history, administration records, and observable recoupment behaviour amount to underwriting infrastructure, even as the layers built on top of them create contractual complexity. Neither cancels the other. It’s not a substitute for legal, financial, or commercial diligence. It’s a reason diligence has something real to work with. Not every catalogue needs an intermediary to become financeable, but where one exists with a genuine track record, it changes what diligence starts from.

As existing revenue streams become better administered and genuinely new ones, including potentially AI-related licensing, develop, the institutional value of clean rights records and clear licensing authority could reasonably increase, because each revenue line is only collectable by whoever already controls licensing on the relevant catalogue. That’s a hypothesis about where value accrues next, not a premium anyone has priced yet. It doesn’t replace the question of who owns what today.

Where this leaves the diligence conversation

None of this suggests African catalogue value has vanished into a handful of distributor balance sheets. The evidence supports meaningful intermediation in specific markets, not a market that’s fully locked up. But it does suggest the obvious deal was never the only deal. For institutional capital approaching African music rights, the intermediary an investor is looking past on the way to the rights-holder may itself be the more interesting asset.

Before asking who owns the catalogue, establish what the investable asset actually is.

Then ask where the opportunity sits: Rights, Portfolio, or Platform. At the first level, you’re underwriting the asset. At the second, the cash flows. At the third, you’re underwriting the machine that keeps finding the next asset before you do.

PenCom spent the first half of 2026 loosening the rules. A waiver for pension funds to buy into the Dangote refinery IPO. A roadmap promising deeper pension participation in the capital market. The regulator that spent two decades essentially telling PFAs to stay in government paper is now actively inviting them out of it.

The money mostly stayed put. As of the May 2026 PenCom data released in late June, Nigeria’s pension industry held N31.3 trillion, and more than half of it, N17.48 trillion, sat in Federal Government securities, with the rules doing far less of that work than the numbers suggest. At a Monetary Policy Rate of 26.5% and OMO stop rates running 20 to 22% across tenors this year, government paper is still simply the best risk-adjusted trade available to anyone managing other people’s retirement savings. The gate opened. Most of the traffic just didn’t change direction.

The Mop-Up Is the Real Story

The mechanics sitting underneath that OMO stop rate start with oil money. The Central Bank takes in oil dollars, converts them to newly created naira at its own rate to pay government, and holds the dollars back, releasing them in small doses through its FX auctions. That pumps naira into the system faster than it releases the dollars that would balance it out. So the CBN mops up the excess naira by selling OMO bills. To get banks and pension funds to absorb that much paper, it has to pay for it: 20%, sometimes 22%, for instruments as short as eight days. Henry Boyo tracked this printing-and-mop-up cycle for years before he passed, and Samson Esemuede at Zrosk Investment Management has tracked the same money-supply mechanics more recently, though neither focuses on pension funds specifically.

That OMO rate isn’t what PFAs earn on their government paper, and the distinction matters. OMO bills are a Central Bank instrument, sold mainly to banks and authorized dealers to sterilize the naira the CBN prints. Pension funds hold FGN Bonds and Nigerian Treasury Bills instead, issued through the Debt Management Office on a separate track. Those yields have been easing all year as inflation cools: 91-day NTBs closed July at 16.3%, 182-day at 16.5%, the 364-day at around 17.4%. With headline inflation at 15.9% in June, that leaves PFAs earning a real return of about 0.4% on the paper making up more than half their book.

Structured venture debt at 16 to 18% is no longer the clearly lower-yielding option next to that. On a pure return basis the two trades are close to a wash. PFAs holding government paper anyway are now paying for certainty directly, in yield given up, rather than collecting a premium for taking the safer trade.

It helps to size how far downhill Nigeria has rolled. The OECD’s 2025 Africa Capital Markets Report, using end-2023 data, puts the regional pension average at 44.4% in bills and bonds, and Nigeria already sat above that average. A more balanced mix, the kind Kenyan and Canadian pension reformers cite, runs closer to 20 to 40% in government bonds, with real weight in equities, infrastructure, and private credit too. Nigeria’s PFAs, at 55.8%, aren’t just skewed by regional standards. They’re skewed against an already-skewed region.

PFAs have started diversifying this year, just not in the direction that matters here. Domestic equity holdings jumped as the Nigerian Exchange rallied through 2025 into 2026, a genuine shift in the portfolio. But buying more NGX-listed blue chips is still a bet on liquid, exit-any-day paper, not the same decision as writing a venture debt facility or a structured SME credit line, where the fund is locked in for years against a single borrower’s outcome. The diversification story holds up. It just hasn’t reached the part of the market that needs patient, illiquid capital to grow.

It’s also the shape you’d expect if compressing yields on safe paper were pushing PFAs toward risk generally. Money moves into the most liquid risk asset available first, and equities are exactly that. A real NTB return of 0.4% hasn’t been enough to send PFAs any further down that road, into the illiquid end, yet.

Run the numbers on what certainty is costing right now. A pension fund holding N100 billion in 91-day NTBs at 16.3% earns about N16.3 billion a year, nominally. Set against June’s inflation print of 15.9%, that’s a real return of roughly N400 million, about 0.4% of the position. Move the same N100 billion into structured venture debt at a comparable 17% instead, and the nominal return is nearly identical, but the money is also financing companies that hire people, build products, and eventually generate the tax receipts and FX earnings that ease the mop-up pressure in the first place. Treat this as a scenario, not a forecast: no PFA holds this much in a single instrument, and real allocation decisions weigh liquidity, tenor, and regulatory limits the illustration skips over. What the scenario highlights is that PFAs aren’t collecting a meaningful yield premium for choosing certainty anymore. But they’re still choosing it anyway.

Diaspora Nigerians sent home $21.8 billion in 2025, flat against the year before, stable enough now to plan around. Still, almost none of it finds its way into equity or venture debt for a Nigerian business. It goes to households and obligations, the same social capital channel that pension funds mirror at the institutional level. More capital in the loop, a bigger diaspora flow, a bigger pension pool, doesn’t escape the pattern. It merely gives the pattern more fuel.

Certainty Has a Price, and Nigerians Pay It at Every Level

Step down from the trillions to a single household, and the same instinct shows up in different clothes. Nigerian money moves fast for a wedding levy, a burial contribution, a milestone birthday. It moves slowly, reluctantly, and suspiciously for a friend’s business idea, even when that friend is closer than most people getting an envelope at an owambe.

The usual explanation is trust radius: people give to family because they know them, and refuse strangers because they don’t. That doesn’t survive contact with how hard it still is to raise money from family too. Ask any founder who’s tried to raise a seed round from an uncle. The friction shows up after the money changes hands, in what the backer worries happens next, not in how far it travelled to get there.

Give N2,000,000 to a burial levy and the return is instant and guaranteed. Gratitude, social standing, a debt repaid in kind when it’s your turn. Even the deferred version, the owambe reciprocity of attending everyone’s party so they’ll attend yours, runs on an unbroken social contract. Give the same N2,000,000 to a cousin’s business and the payout is uncertain, binary, and deferred on every side of it. If it works, you get your money back, maybe with pride at having spotted talent early. If it fails, you lose the money and your social standing, because a failed business reads as bad judgment on someone’s part, and no cultural script turns that loss into social credit the way a funeral levy does. No rational person makes that trade at the margin without unusual conviction in the specific person standing in front of them.

This runs on mental accounting: two ledgers, two different rules for what counts as a safe bet. It’s the same ledger a pension fund’s investment committee runs when it picks a 20% OMO bill over a 17% venture debt facility with a founder attached. Certain and immediate beats uncertain and deferred, whether the actor is a village elder or a fund manager with a Bloomberg terminal.

Share

The Position

Nigeria doesn’t have a capital shortage. N31 trillion in pension assets and $21.8 billion a year in remittances say otherwise, loudly. It has a system where the certain, extractive-adjacent choice keeps beating the productive, uncertain one at every level, from the household to the regulator. Deregulating the rulebook can’t address that, because the rulebook was never the only thing holding the money in place.

My view is that two things would move the needle, both with working models to build from already.

The first is changing the risk math on productive lending itself, so venture debt and structured credit can compete with government paper on certainty, not just yield, since the yield gap has already mostly closed. DFI-backed first-loss facilities do exactly this. A development finance institution absorbs the first tranche of default risk, so a commercial PFA only takes exposure once that cushion is used up, and picking the productive trade stops requiring a leap of faith. It’s a term sheet mechanic, not a policy speech, and it’s the difference between a fund manager saying no and saying yes.

The second solution sits closer to home, in a system Nigerians already use every week. Esusu and ajo run on a guarantee, not a gamble. A rotating contribution society works because your turn is guaranteed, just delayed, and the group enforces the rotation so the deferral never turns into a loss. That’s a working example of turning a deferred payout into a certain one through structure instead of personal relationship. The lesson worth porting to entrepreneurial capital isn’t asking people to tolerate more risk. It’s building the rotation calendar’s equivalent for a business investment, something that makes the process certain even though the business outcome can’t be.

Picture a $50,000 diaspora cheque, pooled from ten diaspora executives contributing $5,000 each, routed through a designated representative drawn from the pool rather than straight into a founder’s personal account. The funding terms get papered upfront in a standardized agreement, with schedules that flex to the specifics of the deal, not settled over a phone call. Standardizing the template this way keeps drafting friction low and is what eventually lets disbursement run on automation instead of someone checking each condition by hand. The money sits in a wallet with joint signatories, representative and founder, both required to authorize a release, so no single party can move it alone. Payment conditions, a revenue milestone, a delivery date, an expense verified against invoice or receipt, trigger each disbursement, so the founder isn’t asking on goodwill and the backers aren’t withholding out of suspicion. Every trigger and transfer sits on an auditable, immutable record any backer can check without asking permission. None of this removes the business risk. The venture can still fail. What it removes is the procedural uncertainty, whether the money was spent as agreed and whether any record exists if it wasn’t, and with it the sting that makes failure so costly: if the venture doesn’t work out, a backer can point to a documented process instead of a private judgment call, and that record keeps the loss from curdling into a story about their own foolishness for trusting a stranger, or a relative, with money. It’s the same logic as the DFI first-loss facility, running at retail scale for a $50,000 pooled cheque instead of a N100 billion pension allocation, minus the administrator the pension version needs at that size.

None of these require Nigerians to trust strangers more, just mechanisms good enough that trust becomes less necessary. Until the return math changes or the trust mechanics catch up, PenCom can loosen the mandate all it wants and the money will keep finding its way back to the safest trade in the room.

The next time someone points to N31 trillion in pension assets or $21.8 billion in remittances as proof Nigeria’s capital problem is nearly solved, ask a sharper question: solved for whom, moving toward what. The money was never missing. It was always exactly where its incentives told it to be.

Kimi K3 charges $3 per million input tokens and $15 per million output tokens. GPT-5.6 Sol charges $5 and $30. Claude Opus 4.8 charges $5 and $25. K3 is an open-weight model from a Chinese lab most Western allocators had never heard of eighteen months ago, and its full weights aren’t even public yet: Moonshot won’t release them until July 27. It’s already undercutting both US flagships on price, sharply on the output side, while ranking fourth among all frontier models on independent benchmarks, behind only Claude Fable 5 and GPT-5.6 Sol itself, and ahead of Claude Opus 4.8. That capability is coming out of a lab operating at a fraction of the capital base American labs have spent building the infrastructure behind their own models.

Reflect on that gap before anything else. It matters less as a fact about Kimi K3 and more as evidence about the story everyone else has been telling.

The bet underneath the bet

Start with what’s really being financed. The big four hyperscalers are on track to spend somewhere between $600 billion and $700 billion on AI infrastructure in 2026 alone. Chips, data centres, power generation, grid capacity. Up from roughly $400 billion the year before. Goldman Sachs puts the 2025 to 2027 hyperscaler capex cycle north of $1.1 trillion.

Capital intensity has crossed 45 to 50% of revenue for some of these companies, a ratio that used to belong to utilities and heavy industry, not software firms. Debt issuance to fund it is projected at $1.5 trillion over the next few years, because none of this can be paid for out of cash flow anymore.

That capital base only makes sense if AI stays hard enough, expensive enough, and proprietary enough that whoever spends the money to reach the next capability tier can charge for the privilege before everyone else catches up. AI already works. That part of the bet is settled. What’s still open is whether working AI stays scarce enough to be worth what was paid to build it.

This is not a new kind of mistake. Railways rewired entire economies in the nineteenth century and still bankrupted most of the companies that built them. Overbuilding and rate wars pushed freight and fares down for shippers and passengers: real consumer surplus, genuinely valuable. The operators competed away their own margins funding the capacity that made those low prices possible, and went into receivership doing it. A large share of the companies that built the internet suffered the same fate. A technology can be a total success and still be a bad place to have parked your capital, because success and captured value are two different questions with two different answers.

What Kimi K3 really threatens

Chinese AI labs have been closing the capability gap for a while. DeepSeek’s release in January 2025 wiped $590 billion off Nvidia’s market cap in a single trading day, and that was a much smaller model than K3. What’s different about K3 is that it isn’t cheap in the way earlier Chinese models were cheap. It’s priced close to the frontier, performing close to the frontier, and going out the door as open weights, free for anyone to download, host, and build on. But at 2.8 trillion parameters, K3 is not something a developer runs on a laptop. So, what it realistically enables is a commodity hosting layer: any cloud provider or inference host can run it and sell access, without paying Moonshot anything for the privilege.. That’s a different kind of threat than “China builds a discount version.” It’s “China removes the thing you were charging for.”

Dean Ball called open-weight releases “inherently decelerationist” this week, and he’s half right in an interesting way. Ball is head of strategic futures at OpenAI, having just left the White House, where he was the chief drafter of America’s AI Action Plan. That makes him an unusually well-placed observer, and also someone whose employer holds the single largest financial stake in frontier-model scarcity staying intact. He’s decelerating the wrong thing. What open models slow down is the return to the private capital that financed frontier development. What they accelerate is how fast that capability spreads to everyone else, including the enterprises and developers who no longer need to pay a premium to a US lab for something a free download does almost as well. Ball himself admits he’s puzzled why Beijing allows models this capable to leave the building for free. The likeliest answer isn’t ideology. A lab with a computing disadvantage, competing against companies with better chips and more of them, doesn’t win by matching the leader’s price. It wins by undercutting it, hard enough that the leader’s price stops mattering.

That’s a specific and legible strategy, not confusion or generosity. If Chinese firms keep publishing capable models as free downloads on Hugging Face and equivalent repositories, they give up direct rent on the model layer, the actual software doing the thinking, in exchange for something more durable: faster domestic AI adoption, less Chinese dependence on American providers, and a shift in where the economic value settles, toward the manufacturing, robotics, energy hardware, and physical deployment layers where China already has stronger positions than it does in chip design. America is trying to own intelligence. China’s strongest AI labs are behaving as if the better trade is making intelligence too cheap to own.

There’s a real counterargument here, and it deserves a paragraph rather than a footnote. Jevons paradox says cheaper inputs can cause total consumption to rise faster than the price falls. Cheap intelligence could make demand explode so completely that aggregate compute spending keeps climbing even as the price per unit of intelligence collapses. In that world, model margins compress, but Nvidia, the power companies, and the data centre operators still do extremely well, because someone still has to run all this compute regardless of who wrote the software. That distinction matters for anyone pricing this risk. The exposure isn’t uniform across the AI stack. It’s concentrated at the model layer, where the pricing power was always going to be hardest to hold onto.

That same concentration cuts the other way for a different set of people. An African founder running AI workloads has spent two years paying dollar-denominated API margins to a US lab, on top of naira or cedi revenue and every other FX squeeze already on the P&L. Once K3’s weights are public on July 27, and as more near-frontier open-weight models follow the same path, self-hosting or routing to a cheaper open-weight backend becomes a live option for workloads that used to require a frontier API call. For a founder whose AI spend is a meaningful share of monthly burn, that’s a real opex line moving in the right direction. The exact size of the saving depends on the workload and the founder’s own numbers, but it’s the kind of compression that buys extra runway without cutting anything else. The sharper implication sits underneath that saving, though. Founders who built “AI-powered X” and treated API access itself as the differentiator now have a shrinking moat, because the thing they were charging a premium for is becoming a commodity faster than most expected when raising their last round. What was defensible eighteen months ago, access to good-enough intelligence, is turning into table stakes. Defensibility has to move to data, distribution, or workflow lock-in that a cheaper model can’t replicate just by existing.

Where this lands on African capital

This is the part that turns it from a Silicon Valley story into a LumiBrief one.

Call it the Capital Transmission Mechanism: global liquidity conditions set by US monetary policy, US risk appetite, and US capital markets don’t stay contained to US portfolios. They move downstream into how much capital DFIs, pension allocators, and global LPs are willing to commit to frontier markets, and how fast. When US financial conditions tighten, African fund commitments tighten with a lag, not because anything changed on the ground in Lagos or Nairobi, but because the LP sitting in New York or Brussels just got more conservative.

I traced this mechanism directly earlier this year, when Iran-war shock psychology helped drive an 87% collapse in African VC fundraising that had almost nothing to do with African fundamentals. Series A volume down 69% year-on-year. Series B rounds essentially at zero. Equity’s share of deal value falling from 76% to 43% while DFI-backed debt rose 165% to fill the gap. The trigger was a war half a continent away from any of the companies whose funding rounds it shrank. The mechanism doesn’t care what triggers it. A war, a rate shock, a US tech repricing: they all move through the same pipe.

A model-layer repricing in US AI equity is a plausible trigger of exactly that kind, and it moves through more than one channel. The most direct one is risk appetite. If investors start pricing in what K3 implies, that the AI capex bet assumed a durability of returns a fast-improving open-weight competitor is actively eroding, a correction in AI-linked tech valuations doesn’t need to be catastrophic to matter. A 15 to 20% repricing across the AI-heavy names carrying an outsized share of US market capitalisation would be enough to tighten the risk appetite DFI investment committees weigh before approving a next fund’s first close or re-upping into a new vintage. DFIs move on US macro signal faster than commercial LPs do, because their mandates are built around exactly this kind of systemic-risk sensitivity.

The second channel runs through the currency, and it’s the one allocators tend to underweight. Tighter US financial conditions generally mean a stronger dollar, and a stronger dollar is bad news for the naira, the cedi, the shilling, and every other African currency already managing its own pressures. A US equity repricing large enough to tighten LP risk appetite is also, most likely, a US equity repricing that comes with dollar strength, which squeezes African portfolio companies on imported costs and debt service at the same time their funders are getting more cautious. The two channels reinforce each other instead of running independently.

None of this requires African GPs to have any AI exposure at all. The transmission doesn’t run through African portfolios. It runs through the LPs deciding whether to commit to the next fund, through the currency their portfolio companies transact in, and through to founders too, further downstream: a slower cycle of new fund closes means fewer priced rounds and smaller checks in the exact window a founder is trying to raise, regardless of what their own numbers look like.

The practical implication isn’t “watch AI stocks” as a general instruction. Everyone already does that badly. It’s narrower. GPs currently in fundraise should treat a US AI-equity repricing event as a leading indicator for LP commitment timing, arriving faster than most African macro data would suggest and independent of anything happening in the underlying African economy. The variable worth watching isn’t whether AI works. It’s whether the companies that financed it get paid back on the timeline their capital structures assumed, because when that question gets repriced in New York, the answer shows up in Accra and Nairobi a quarter or two later, in the form of a slower first close.

The AI investment thesis and the AI technology thesis have quietly come apart from each other, and most allocators are still pricing risk as if they’re the same question. They aren’t. One asks whether artificial intelligence will create enormous economic value. It probably will. Kimi K3 is itself proof the technology keeps getting more capable, more available, and cheaper to access. The other asks whether the specific companies that borrowed $1.5 trillion to build it will be the ones who get paid. Those are different bets, and an allocator who has only priced the first one is carrying risk they haven’t named.

The greatest threat to the AI investment boom was never that AI would fail. It’s that intelligence becomes abundant before its financiers get paid back, and that bill doesn’t stay in Silicon Valley. It travels through the same liquidity pipes that already decide, every quarter, how much capital reaches a fund closing in Lagos.

If you’re in a fundraise this year, or sitting on an LP commitment decision, this is worth watching more closely than the headlines suggest. I’ll be tracking it.

Iya Bose (“Bose’s Mum,” the standard way an older market woman is addressed, using her child’s name) has run a grains stall at Mile 12 for close to two decades. She knows which retailers pay on collection day and which ones need three days’ grace before a bad week turns into a bad month. She knows this the way a person knows a family member’s moods, not from a spreadsheet, from watching. When a retailer’s order slows down two weeks before Ramadan, she already knows why. She already knows whether to extend credit anyway, and how much.

No fintech founder built her that judgment. She built it herself, transaction by transaction, over years nobody else was tracking.

The founders who eventually made money in Nigeria’s payments boom didn’t replace Iya Bose. They built underneath her. First a POS terminal, then a settlement account, then eventually a credit line, the payment and settlement pipes, the “rails,” that sit underneath a transaction once it’s agreed, rather than anything to do with deciding who gets credit in the first place. She kept doing exactly what she was already doing, just faster, and with someone else’s balance sheet behind her instead of her own memory alone. The founders who tried the opposite move, going around her straight to her retailers, mostly spent eighteen months rebuilding a version of the trust she already had. Badly, and at their own expense.

I first saw the shape of this argument somewhere other than my own notebook, a sharp, specific observation on a timeline about why the trader down the road trusts the person collecting her daily contribution more than she’ll ever trust an app. I went looking for it again while writing this and couldn’t find it. What follows is my attempt to work out why it was right.

The Trust Radius

Every pitch deck that quotes “40%-plus of Nigeria’s GDP is informal” treats that number as a gap, an unbanked population waiting for the right app. Buried inside that framing is an assumption nobody says out loud. That the informal economy is informal because nobody has built anything for it yet.

The informal economy already runs on infrastructure: Iya Bose’s memory, and a hundred thousand people like her, each holding a working credit model inside a relationship instead of inside software.

Ask what any lender needs to underwrite a loan. A reliable read on cash flow, and some way to make sure the money comes back. Iya Bose has both, and her version updates faster than any bank’s. She watches a retailer’s order size drop before a risk model in a Lagos head office would ever see it. What she doesn’t have is a way to turn that knowledge into capital at the scale her retailers need. That was always the real gap, and it sits in the pipes underneath trust that already existed — not in the product on top of it.

None of this works if the person holding the trust has a reason to resist formalizing her role instead of joining it. So name the incentive directly rather than assume she’ll cooperate. Take the alajo collector, the person who has for generations walked Yoruba markets daily gathering contributions into a rotating savings pool. She has always earned a fee for that exact function, standard practice puts it at roughly one day’s contribution per thirty-day cycle, around 3.3%. It’s a paid role already, just an informal one, and it’s also a role with real exposure. Lagos courts have prosecuted collectors over fraudulent thrift cards, and every market has a story about one who vanished with a month’s contributions. LibertyPay and a few others now hand the collector a small POS terminal, and contributions move through it directly instead of through her hands, whether by card, transfer, or USSD, which removes the thing that actually creates the risk: cash sitting with one person between collection and payout. The same shift shows up even without a fintech company involved. Plenty of ajo groups now run over WhatsApp, contributions by transfer instead of cash-in-hand, with someone in the group still doing exactly what the alajo always did: tracking who’s paid, chasing stragglers, deciding when the pot moves. Where that person gets removed entirely, in the fully automated version of digital ajo with no organizer at all, participation and trust both tend to suffer. She’s the reason the system works. Nobody is asking her to give up the thing that matters to her position. She is still the one whose word decides who’s good for it. Formalize the fee that already exists. Don’t touch the judgment that earns it.

Where founders get burned is the opposite move, trying to reach her customers directly and cut her out. It’s the more expensive path every time someone’s tried it, because you’re not just building a product. You’re rebuilding, from zero, twenty years of memory someone else already had for free.

A claim this clean should worry you a little. If “trust conversion, not digital adoption, is the binding constraint” can explain every outcome after the fact, it isn’t an argument. It’s a horoscope. So ask what would break it, and whether anything already has.

Moniepoint’s own numbers hand over the answer. Through 2025 the company built a book that, by its own account, ran to over ₦1 trillion in credit to roughly 70,000 businesses, an average loan size of about ₦14.3 million, and borrowers who kept up repayments showed an average 36% jump in transaction value afterward. That’s the trust-radius model working as described: short tenors, repayment tied directly to daily POS data, underwriting that’s really just formalized memory of cash flow.

Then look at where it didn’t hold. By early 2026 the same institution was carrying close to ₦7 billion in non-performing loans, concentrated in two facilities that looked nothing like the model above. A ₦5 billion working-capital loan to Alerzo and a ₦2.4 billion facility to the ShopRite Nigeria franchise, both written with corporate-style, multi-month tenors instead of the daily-repayment structure tied to transaction data. Both went bad. ShopRite’s stores shut down entirely.

It’s the boundary condition — and a useful one. The pattern holds where credit is priced against a trust-holder’s daily, observed cash flow, and it stops holding the moment a lender extends the same logic to a borrower who was never inside that trust radius, meaning the specific stretch of relationships whose day-to-day risk someone can actually judge from having watched it. A large corporate account underwritten like a bank loan sits outside that radius even when the same institution is writing the check. And a founder building an app-first product in a category with no Iya Bose yet to formalize sits outside this claim entirely too. This is narrower than “digital-first always loses,” and being narrower is what makes it testable.

Share

Where EETAM Stops

EETAM does one job well. It tells you how big a market really is once you strip out the four things founders love to skip, how much of that population is economically active, how many can be reached digitally, how many can afford the product, how often they’d realistically use it. Run those filters on a claim like “Nigeria’s informal SME credit gap is a $30 billion opportunity,” at 45% economic activity, 55% digital reach, 60% affordability, and 70% frequency, multiplied rather than added since a business that fails even one test isn’t addressable at all, and that collapses to roughly $3.1 billion. That’s the number that should anchor a term sheet, not the headline.

EETAM was never built to answer why the market stayed unaddressed once it’s been sized correctly. For Iya Bose’s segment, the four filters above already point to yes: she’s economically active, reachable, able to pay, transacting constantly, in cash, at high frequency. None of that was ever the binding constraint. The constraint sits one layer beneath the sizing math entirely, in whether a lender can turn what Iya Bose already knows into a number a risk model will accept. Call it trust conversion. EETAM tells you the market is real. It doesn’t tell you why the market stayed unaddressed anyway, and that second question is the one this piece is answering, because a founder who runs the EETAM math correctly and still ships an app-first onboarding flow has solved the wrong variable precisely.

There’s a second-order effect worth stating, because it’s the part that makes this more than a story about one trader. Give Iya Bose working capital and she doesn’t just hold it. She buys more stock. And because she already extends short-term credit to her own retail customers, on exactly the same trust logic a lender is now extending to her, that capital moves one more step down the chain without anyone underwriting it twice. Her retailers get more room to restock. Their customers see fuller shelves. One loan, priced against one trust-holder’s judgment, re-lends itself through relationships that already existed and cost the lender nothing extra to reach. That’s a different kind of return than a single-borrower loan produces. The same naira turns over more times, at more points in the chain, before it leaves the market. It’s also the real efficiency case for funding the person who already holds the trust rather than trying to underwrite every trader beneath her one loan at a time. You reach the whole chain by pricing the node at the top of it correctly, once.

The founders who win the next wave in Nigeria’s informal economy will be the ones who can name, specifically, who already holds the trust in a given trade, a super-agent, a market association head, an alajo collector, and who build settlement and credit rails for that person to run ten times their current volume without asking them to change how they work. This also flips who the real competitive risk is. It’s the well-capitalized founder who tries to disintermediate Iya Bose, fails, and spends eighteen months rebuilding, from a standing start, the one asset that took her two decades to earn and can’t be shortcut with funding.

The Position

Sizing a market correctly answers one question. Whether that market can actually be reached is a separate one, and in large parts of Nigeria’s informal economy, that second answer was never digital adoption. The boundary case shows exactly where it stops being true, the moment credit gets written against someone who was never inside the trust radius to begin with. Everyone who has made real money underwriting the informal economy so far did it by finding the person the market already trusted, and building underneath her.

For anyone underwriting a bet in this space rather than building one, the diligence question is the same one, asked in reverse. Don’t start with whether the product is good. Ask who already holds the trust in that specific transaction, whether the model is priced against that person’s daily, observed behavior or against something one step removed from it, and what happens to repayment the moment that distance grows. Run the Alerzo and ShopRite loans through that question and they fail it before a single naira goes out the door, which is exactly the diagnostic value of asking it first instead of after the write-off.

The next founder chasing a piece of that opportunity should stop asking what app the market needs. They should go find their own Iya Bose, sit with her for a week, and ask who she already trusts to hold money on her behalf. Build for that person. Everyone else is still knocking on the wrong door.

In June 2021, a month after his fund paid $140 million for the Red Hot Chili Peppers’ catalog, Merck Mercuriadis told *Variety* there was no bubble in music rights. “A bubble is when somebody is overpaying for something that doesn’t have the sort of metrics that these investments have,” he said. By late 2023, Hipgnosis Songs Fund’s board was issuing what its chair called a “health warning” about its own asset values. The independent valuer resigned. A reassessment sliced 26% off the fund’s stated worth in a single report, and currency effects pushed the real number to 33%. Shares fell to an all-time low.

Give Mercuriadis this much: music catalogs really do have measurable metrics. His fund just measured the wrong ones. It had priced 65,000 songs on the assumption that earnings would keep climbing the way a growing company’s revenue does. What a later valuer found instead was a portfolio doing what royalty income actually does: rising fast in the first few years after release, then settling into a long, slow decline. Hipgnosis had underwritten a startup. What it bought was a bond.

This should bother anyone allocating capital into African creative IP right now. In the same period Hipgnosis was unwinding, money was moving into African music platforms on the opposite assumption, and nobody called it a bubble. Kupanda Holdings, backed by TPG Growth, took equity in Mavin Global in 2019, ahead of Universal later buying a majority stake. Warner acquired the distribution platform Africori outright. Afreximbank took an equity position in the direct-to-fan platform CREAM around the same time the bank committed a reported $1 billion to African creative industries more broadly. Every one of those deals was priced like a growth business: market share, user acquisition, the bet that the platform becomes more valuable as more artists and fans route through it. Nobody ran a decay curve on CREAM’s user base the way Shot Tower ran one on Hipgnosis’s back catalog. Nor should they have. The two are not the same kind of asset, and treating them as though they are is exactly the mistake that gutted Hipgnosis.

The distinction matters beyond any single fund’s balance sheet. African creative IP is one of the few asset classes where the continent is a genuine net exporter of globally demanded product: the songs travel, the streams accumulate in London and Paris and Nairobi, the value is real. Almost none of it gets collateralized locally, because almost none of it has been sorted cleanly enough to price. The same category-confusion shows up wherever African assets meet institutional capital: plenty of demand, mismatched underwriting discipline. Get the categorization right on catalog versus platform, and you get financeable, legible assets. Get it wrong, and the value stays real but unbankable — a worse outcome than a mispriced deal, since at least a mispriced deal shows up on someone’s balance sheet.

The word “music” is hiding two different balance sheets.

A song catalog is a right to collect income from work that already exists. Nobody is writing new hooks for it. Its earnings peak somewhere in the first 2 to 5 years after release, then decline toward a long, flat tail, predictably enough that buyers price it the way a bond desk prices a long-duration instrument, off a discount rate rather than a growth story. A platform is an operating business: a label, a distributor, a rights-management tool, a direct-to-fan app. It has a product roadmap, a user base it’s trying to grow, and a moat it’s trying to widen. What it will become matters far more than what it already collects. One is an annuity you buy because the cash flow is dependable. The other is a startup you buy because the cash flow doesn’t exist yet and you’re betting it will.

Confuse the two and the error compounds regardless of direction. Overpay for a catalog using platform-style growth assumptions and you get Hipgnosis, a board explaining to shareholders why the songs are worth a third less than the fund said they were. Run the same mistake the other way, treating a platform’s growth story with catalog-style skepticism about decay, and you pass on deals like CREAM and Mavin, real equity gains that had nothing to do with royalty curves at all.

Share

Run the catalog side of the math first, because it’s the more expensive place to be wrong

By 2026, independent catalog deals were closing at roughly 12 to 18 times net publisher’s share, down from 18 to 25 times at the 2021 peak, 3 to 6 turns of compression driven mainly by higher interest rates repricing every long-duration income stream. There’s a direct relationship between the multiple a buyer pays and the discount rate baked into it: roughly 10x implies a 14% discount rate, 16x implies about 9%, and 22x implies something closer to 7%. That’s bond territory, reserved for blue-chip, decades-old catalogs with the flattest, most proven decay curves. When Kroll valued the legacy Hipgnosis portfolio for a securitization, it assumed a long-term growth rate of about 2%, modest and decay-aware, a long way from Mercuriadis’s original growth story. Shot Tower’s earlier, harsher review had used a 9.63% discount rate and arrived at a valuation $690 million below what the fund had claimed the year before.

Put a number on what that gap means for a single deal. Take a catalog earning $500,000 a year in net publisher’s share. Price it the way Hipgnosis priced its early acquisitions, an 18x multiple implying roughly a 9% discount rate and an optimistic growth assumption, and you get a $9 million valuation. Price the same catalog the way the market prices a disciplined 2026 deal, 13x, a 12% discount rate, decay properly modeled, and you get $6.5 million. Same songs, same royalty statements, a $2.5 million gap that exists entirely inside the assumption about how fast the earnings fade. That gap is where funds get built or wrecked, and it has nothing to do with whether the songs are any good.

Platform deals don’t run on that math at all, and shouldn’t. CREAM’s value to Afreximbank isn’t a multiple of trailing royalty income: it’s a bet on whether African artists route more of their direct-to-fan revenue through the platform over time, the same logic that prices any consumer infrastructure company. Mavin’s jump from a Kupanda equity check to a majority Universal acquisition tracked the label’s growing catalog *and* its distribution reach, its A&R pipeline, its ability to keep producing hits rather than just collecting royalties on old ones. In other words, a platform priced like a platform. Underwriting Mavin as a growth business was the right call; applying that same growth logic to a decaying catalog is the actual error.

A third posture sits between those two, and it changes the math for anyone underwriting a catalog as pure annuity. Some buyers don’t just collect the decay — they work against it. Hipgnosis’s own model priced in what it called “Song Management”: the deliberate pursuit of sync placements, brand campaigns, and cultural moments that pull a catalog’s earnings back toward peak rather than letting them slide down the curve untouched. A song 15 years past release with a flat, predictable tail can spike hard off a single film placement, a viral sample, or a brand deal, and if the buyer is actively hunting those opportunities rather than passively collecting royalty statements, the decay curve isn’t fixed. It’s a variable the buyer is trying to bend. The underwriting question shifts accordingly, from what a catalog earns on autopilot to what it earns in the hands of a buyer whose job is to keep finding it new life. A passive holder prices the tail as given. An active manager prices their own ability to reset it, and pays a premium for the option, much as a value-add real estate investor pays more for a tired building they believe they can re-lease than a passive landlord would pay for the same building’s existing rent roll. Get that distinction wrong, in either direction, and you either overpay for management skill nobody actually has, or underpay for a catalog someone else is about to revive.

The question that sorts a given African music asset into the right lane, annuity or startup, isn’t genre or geography. It’s whether the rights are clean enough to trade at all. A catalog with unregistered splits, disputed authorship, or a chain of title that breaks somewhere between the songwriter and the current claimant doesn’t behave like a bond or a startup. It behaves like an asset nobody can safely price, because nobody can confirm what they’d actually own if they bought it. Clean title, a legible cap table on the platform side, and rights that can actually transfer without a lawsuit attached — those three things determine whether an asset can be priced on its economics at all, before anyone gets to argue about which economics apply.

PwC has put Nigeria’s music industry at roughly $19 billion. Spotify alone paid Nigerian artists $37.5 million in royalties in 2024, more than double what it paid the year before and a genuinely good number for the industry. Sit those two figures next to each other and the gap is the whole argument: an industry-size estimate and the actual, statement-backed cash flow a lender could underwrite today are not the same number, and treating them as though they describe one undifferentiated pool of value waiting for capital is how allocators end up writing a single check into “African music IP” — one discount rate, one exit assumption, one story about where the growth comes from — running the Hipgnosis mistake at a smaller scale, just with less press coverage when it comes due.

Two disciplines are required here, and most allocators only have one built out. Catalog acquisition is fixed-income analysis that happens to involve music, full stop, no matter how the deck describes it. An allocator who can only do that kind of analysis well should stay in that lane rather than let a single valuation model wander across both catalog and platform, much as Hipgnosis let a growth story wander onto a spreadsheet built for annuities, and paid for it in a board resignation and a third of its NAV.

Which lane a specific catalog falls into gets decided before any money moves, and it gets decided by lawyers, not by market analysts. Whether a set of rights produces a flat, predictable decay curve or an unpriceable mess depends on whether the chain of title holds: whether every songwriter, every sample clearance, every prior assignment is documented cleanly enough to survive due diligence. Usually that question gets asked after the check clears, when it should be the first one. It’s the subject of the next piece.

That’s the part with real stakes attached. Kobalt raised $266.5 million and Concord raised $850 million in the last few years by securitizing music rights into bonds. That packaged royalty income into something a debt investor could underwrite with confidence, precisely because the income was legible enough to model. Nothing comparable exists yet at scale for African catalogs. The underlying songs don’t earn any less; too few of them have been sorted cleanly enough, on clean title, clear posture, active versus passive, to package that way. This is a capital-formation gap before it’s a valuation gap. Every African catalog and platform priced correctly, in the right lane, with the right discipline, is one step closer to being an asset a debt investor could actually collateralize rather than one only a patient equity check can touch. The songs don’t know what kind of asset they are. The people pricing them have to.

Two companies signed a strategic partnership eighteen months ago. There was a press release, a photo of executives shaking hands, a paragraph about synergies neither side ever bothered to define. Neither company can now tell you the exact date it ended, because it never formally ended. It just stopped getting calendar time. Nobody called a meeting to announce the funeral, because nobody had signed anything expensive enough to require one.

Six months earlier, a different pair of companies had finished the opposite kind of deal. Fifty pages, three law firms, a shareholders’ agreement with reserved matters and drag-along rights, a full board structure built out before either side had sold a single unit. It took four months and cost each side the better part of ₦20 million in fees. The joint venture company has generated ₦0 in revenue in the fourteen months since incorporation, because whether customers would actually pay for what the two companies proposed to build got tested after the entity existed, not before.

Both partnerships failed. Neither failure looked like the other, and that’s the part worth sitting with.

Two Failure Modes, One Cause

The instinct is to treat these as opposite problems needing opposite fixes — do less paperwork, or do more diligence. That’s the wrong diagnosis. Both deals failed for the same reason. Neither one tied the size of the commitment to the amount of evidence anyone had that the commercial idea worked.

The MOU costs almost nothing to sign. No resource changes hands, no deadline forces a decision, no metric can be missed because none was set. That’s exactly why it carries no weight — with regulators, with boards, or with the other side. Economists have a name for this: cheap talk. A signal only tells you something if it costs the sender something to send. Both companies got to tell their boards they’d “signed a strategic partnership” without either side risking anything that would sting to lose. So neither side prioritized it over whatever was actually on fire that week, and it quietly starved.

Nigerian courts generally treat this the same way, as a matter of doctrine. An MOU is presumed non-binding unless the parties can show an actual intention to create legal relations, and most MOUs are drafted specifically to avoid showing exactly that.

The joint venture agreement is the mirror image. It’s expensive enough to be a real signal, which is exactly the problem, because it got spent before there was anything to signal. Board seats, profit shares, exclusivity terms, and governance rights all got negotiated against a commercial thesis that existed only as a slide deck. Setting up the entity itself, board resolutions, CAC registration, a shareholders’ agreement, typically runs six to twelve weeks once any regulatory sign-off is involved. That’s six to twelve weeks both sides could have spent finding out whether anyone would pay for the thing they were about to spend that time structuring. The lawyers did their job well. The job itself was premature.

The Proof Gate

This is the model I’ve been using on recent partnership work, stripped of client specifics. Treat the first phase of any partnership as a call option, not a commitment.

A call option gives you the right, not the obligation, to buy (or buy into) something later at a set price, and you pay a small premium now for that right. Applied to a partnership, that means something specific. Instead of negotiating the full structure up front (equity splits, board seats, exclusivity, exit terms) before anyone knows whether the underlying commercial idea works, you pay a small premium instead. A short, cheap, tightly scoped pilot, in exchange for the right to build the full structure later, once there’s evidence.

The premium has to be real, though. An MOU’s premium is zero by design; a pilot’s can’t be. A pilot earns that distinction three ways, and none of them are optional. Phase 1 still carries a minimum viable set of protective terms — who owns the IP generated during the test, confidentiality, a defined exclusivity window so the counterparty can’t shop your integration to a competitor while “testing” it, a clean no-fault exit. Five pages, not fifty, but real. Before Phase 1 even starts, both sides agree a named metric and a deadline, not “let’s see how it goes” but ninety days and a specific paying customer or transaction volume, a number written down in advance. And a go/no-go date already sits on both calendars, so the review isn’t something either side has to chase.

The go/no-go date is what prevents the invisible death. An MOU dies by attrition because no date forces anyone to look at it. A Proof Gate dies, or graduates, on a specific Tuesday, in front of everyone who signed off on it, on both sides, whether the news is good or bad.

Run the arithmetic on what each path actually costs. The dead JV costs ₦15–25 million in fees plus three to six months of senior time on both sides, sunk, producing zero revenue and a governance structure nobody needed yet. The dead MOU costs next to nothing directly, but it burns something more expensive to rebuild — the credibility of “strategic partnership” as a phrase, inside both organizations, the next time someone proposes one. A Proof Gate that fails costs a small, bounded loss, plus a clean, dated answer that lets both sides move on without the slow bleed.

Apply the same ninety days the other way, and the picture changes. A scoped pilot at ₦2–4 million in legal cost, tested against a named metric, produces something a slide deck never can. A real number. Walk into the full JV negotiation with ₦18 million in verified transaction volume from the pilot window, and the ₦20 million structure is now being priced against realized cash flow, not a forecast either side can quietly walk back later. Board seats and profit shares argued over proven revenue settle faster and hold better than the same terms argued over a projection, because nobody on either side can claim the number was always going to be different.

There’s a second thing the pilot buys that the arithmetic above doesn’t capture, and it matters more than the revenue number: attribution. Walk into almost any partnership pre-pilot and ask each side privately who is actually driving the value, and you’ll get two different answers, both confident, both unverified. Take a fintech and a telco co-selling a savings product. The fintech assumes its underwriting model is the reason anyone deposits money. The telco assumes its distribution is the reason anyone signs up at all. Both walk into the term sheet believing they’re the senior partner. Ninety days in, the transaction data settles what the debate couldn’t. Say the telco’s channel produced 3,200 of the pilot’s 4,000 signups, but the fintech’s underwriting produced 71% of actual deposit volume once those accounts went live. That’s not a 50/50 partnership, whatever the original deck assumed, and Phase 2 terms (profit share, board composition, whose brand leads the marketing) should be built on that split, regardless of the goodwill estimate both sides walked in with. The same pattern shows up in music. A label and a distributor structuring a joint release each tend to credit their own marketing spend or catalogue reach for a record’s performance. A tracked pilot period settles that argument too. If 60% of streams trace to playlist placements the distributor controlled and only 15% to the label’s own campaign, that figure, not either side’s sense of its own importance, is what should set the next contract’s terms. Before the Proof Gate, both parties negotiate from the assumption that they’re the primary value driver, because neither has evidence otherwise. After it, the evidence exists, and there’s simply less left to argue about.

Who Gets to Insist on This

One honest caveat, because the model isn’t neutral to power. Whoever has more resource leverage, better alternatives, or lower switching cost gets to decide how a deal is shaped, not whoever has the better argument for phasing. A smaller company proposing “let’s test first” to a much larger counterparty can read, to that counterparty, as a lack of conviction rather than discipline — and the larger side may simply walk toward someone who’ll commit on their terms.

The way around this isn’t abandoning the model. It’s making sure the smaller side is offering something genuinely scarce inside the pilot (proprietary data, a licence, a distribution position the larger company can’t easily replicate elsewhere), so the Proof Gate reads as confidence in the asset, not hedging on the relationship. Scarcity is what buys the right to propose the sequencing. Without it, the larger party sets the terms regardless of what any deal-structuring model recommends.

There’s also a version of this that doesn’t tolerate phasing at all. Anything gated by a regulator before commercial activity can begin (a banking licence, a change-of-control approval) collapses the option value. You can’t run a cheap test of something that legally cannot happen without the full structure first. And any deal where the first exposure is itself the risk (handing over a proprietary valuation model, a royalty ledger, source code) doesn’t have a cheap version of “testing the water,” because the water is the asset. The Proof Gate is a default, not a law.

A second caveat matters just as much, and it’s not small. The protective terms inside a Nigerian Proof Gate are worth less as courtroom weapons than they look on paper. Commercial disputes in Nigerian courts commonly take up to three years to reach a first-instance judgment, which means an exclusivity clause is functionally unenforceable within the pilot’s own ninety-day window; nobody is getting an injunction before the test period ends anyway. The real deterrent isn’t the clause, it’s reputational. A counterparty who breaches a Proof Gate mid-pilot burns the relationship, and in small, overlapping markets like Lagos venture capital or Nigerian entertainment, that story travels faster than any suit would resolve. The clause is for the record; the market is the enforcement.

The real shift this argues for is sequencing, testing the partnership against a paying customer before testing it against a board committee. Most partnership work today optimizes for internal approval first: get the board comfortable, get the structure signed, get the press release out. Market validation happens afterward, if at all. The Proof Gate reverses that order, which is a harder discipline than it sounds. Internal approval is comfortable. Getting a stranger to pay for something takes actual proof.

For a platform company sitting on data, distribution reach, or a proprietary tech stack that a technology partner wants access to, this cuts the other way too. A naked pilot without IP and exclusivity terms is just a slower way to get raided, watching a counterparty extract the value of the integration and walk before Phase 2 ever gets negotiated. Size the structure to the cost of being wrong, and gate it to a date that forces an honest answer. The whole model, in one line.

Most partnerships don’t fail because the idea was bad. They fail because nobody designed a cheap way to find out early — so the expensive version got built on a guess, or the cheap version never had to answer for itself. Fix the sequencing, and the deal tells you the truth before it costs you anything real to hear it.

None of this is really about one JV or one MOU. [Africa’s private capital market moved US$5.1 billion across 530 deals in 2025] (AVCA), deal volume up 8% year on year even as total deal value fell 5%, a shift toward smaller, more disciplined transactions rather than a shortage of appetite. Africa-focused fund managers raising new vehicles saw fundraising drop 34% year on year in the same period. That’s the environment every partnership on the continent gets structured in now: less capital, chasing fewer positions, held by LPs asking harder questions about what actually converted into cash rather than what was announced. A ₦15–25 million dead JV or a slow-bleeding MOU isn’t just a bad quarter for two companies. Multiply that pattern across a market this much more expensive to raise into, and it’s the same story that shows up at portfolio level as a weak distributions number and a GP who can’t fully account for where the last fund’s capital went. The continent has plenty of partnership ideas. It doesn’t have enough cheap, honest ways to find out which ones were ever real before the capital gets spent finding out the hard way.

On Tuesday, Flutterwave announced a Series E. The headline number was $3.2 billion. Four years ago, in February 2022, the company raised $250 million at just over $3 billion. Do the arithmetic and the conclusion writes itself: flat. Maybe up a touch, depending on which outlet’s figure you use. Some say $3.2 billion. Bloomberg’s reporting puts it at $3.3 billion. Either way, one of Africa’s most valuable startups is still worth, four years later, almost exactly what it was worth in February 2022.

The crypto press covered this as a Ripple story, mostly. RLUSD landing on one of Africa’s largest payment rails, the XRP Ledger getting a live enterprise use case, the institutional case for stablecoins moving forward another notch. Read past the headline, though, and there’s a quieter story sitting underneath the valuation line that nobody seems to be asking about: Flutterwave just held a number it was priced at four years ago, through a global rate-hiking cycle, naira devaluation, and a continent-wide repricing of every 2021-22 vintage startup.

Is “flat” actually a good outcome, or the most interesting number in the entire announcement?

What We Called In January

In January, after Flutterwave’s all-stock acquisition of Mono, I wrote a forensic piece on this publication arguing that the all-stock structure was not strategic. It was arithmetic. Mono’s investors took Flutterwave shares instead of cash for a $25-40 million deal, and the reason, I argued, traced back to a cash position under pressure. Working from disclosed funding ($250 million raised in February 2022), a reconstructed burn rate, and the absence of any new round in four years, the model put Flutterwave’s runway at roughly ten months from January 2026. The company would need to raise by late 2026 at the latest, and likely had already started.

The size was wrong, and the reason it was wrong is worth more than the original guess would have been.

What the essay got right: the need, and the timing. Flutterwave was in raise mode. We now know that by April 2026, the company was deep enough into fundraising conversations that a Tinubu aide felt confident enough to publicly announce a $75 million government investment via the Ministry of Finance Incorporated, tied to a planned $250 million IPO. Flutterwave denied it — but the denial, read carefully, only disputed the $250 million IPO figure specifically. The company never addressed its broader relationship with the federal government, and Agboola had joined President Tinubu’s UK state delegation weeks earlier. Two months after that, the Series E closed. The runway pressure the January model predicted was real; the company was managing exactly the kind of capital-raising process the model anticipated, on roughly the timeline it anticipated.

What the essay got wrong: the magnitude. The January model assigned 60% probability to a down-round of 25-33%, landing somewhere around $2-2.5 billion. That was the central, highest-confidence scenario. It didn’t happen. Flutterwave priced flat to slightly up.

Why the Down-Round Didn’t Come

A markdown is not just a number. It’s an action that specific people have to take, and those people have incentives that a pure cash-flow model doesn’t capture.

Flutterwave’s existing investor base — Tiger Global, B Capital, Avenir Growth, and others from the 2021-22 rounds — are sitting on positions marked at or near $3 billion. A new round at $2 billion doesn’t just reprice Flutterwave. It forces every one of those funds to take a markdown on their own books, in their own LP reporting, in the same reporting cycle. Tiger Global, in particular, led both Flutterwave’s 2021 Series C and Mono’s 2021 Series A. A flat Flutterwave print protects two marks at once.

The January model gave this one line and moved on. It deserved the center of the analysis. A markdown requires someone with leverage over price to want it to happen, and in Flutterwave’s case, the people with the greatest incentive to avoid one were also among the stakeholders best positioned to help the company avoid it. That’s not fraud, and it’s not even unusual. It’s just how syndicates with overlapping cap tables behave when an existing portfolio company needs fresh capital. Defending a mark is what the relationship is built to do, the same way a bank’s credit committee is built to protect its existing loan book before it underwrites a new one.

The second piece worth separating out: not all capital prices the same way. The 2022 Series D was led by B Capital Group, a financial growth-equity investor, with a syndicate of seven additional named funds, full disclosure of round size ($250 million), and a publicly stated thesis about Flutterwave’s growth trajectory. That is what a priced, competitive, third-party-tested round looks like.

The Series E is a different animal. One named investor: Ripple. Agboola confirmed to TechCabal that it was “an actual cash investment” and that Ripple is “now an equity shareholder of the company,” so this is real primary equity, not a stock swap like Mono. But Agboola also told Bloomberg directly that he would not disclose the amount invested or the size of Ripple’s resulting stake. No lead financial investor. No syndicate. No round size. What’s disclosed is a strategic partnership: Ripple’s RLUSD stablecoin and the XRP Ledger get embedded into Flutterwave’s payment rails and remittance corridors.

A strategic infrastructure investor is not pricing the same thing a growth-equity fund prices. Ripple is buying distribution — a settlement rail across one of the largest payment networks in Africa, plus a stablecoin embedding deal that B Capital was never going to write a check for. When the thing being purchased is partly strategic access rather than a pure bet on enterprise value, the valuation attached to that access tells you less about fair market price than a syndicate-led round would. You can pay a premium for a relationship. You cannot easily pay a premium for nothing, which is what a financial investor, with no use for the rails, is implicitly checking when they price a deal.

Put those two things together. An existing syndicate with every incentive to avoid a markdown, and a single strategic check that was never required to compete for the deal. Neither proves the $3.2 billion is wrong. Together, they mean this round provides a weaker signal of market-clearing value than a competitive, disclosed, multi-investor round would.

Share

What the Round Size Might Actually Be

Flutterwave hasn’t disclosed how much Ripple invested. We have one useful data point, even though it comes from an entirely separate, denied transaction. In April, reports surfaced that the Nigerian government, through the Ministry of Finance Incorporated, was preparing a $75 million investment as part of a $250 million IPO raise. Flutterwave denied that specific deal. The estimate that follows rests on a thin evidentiary base: the $75 million figure came from a transaction Flutterwave denied, with a different investor and likely a different rationale, and there’s no public evidence the Ripple check was sized anywhere near it. It’s simply the only externally reported reference point available for what kind of capital Flutterwave was discussing around this time, and it’s a reasonable floor for the Ripple check on that basis alone.

Ripple is a strategic infrastructure investor, not a financial investor maximizing ownership. Strategic checks of this kind, paired with a commercial integration rather than a pure capital play, tend to land at the lower end of what’s plausible, because the investor is optimizing for the partnership, not the stake. Working estimate: $100 million. Using Flutterwave’s own $3.2 billion figure as the post-money valuation (Bloomberg’s $3.3 billion would shave the resulting stake slightly smaller, not larger), that implies new investors taking roughly 3% of the company. A modest, controlled dilution event, consistent with a company managing its cap table rather than recapitalizing under pressure.

That estimate carries through to the runway picture as a single committed number, with the assumptions shown.

The January model assumed $3.5 million in average monthly burn, based on the deployment pace from 2022 onward. That assumption needs updating on two fronts. First, Agboola’s own H1 2025 letter stated that monthly margins had doubled compared to the 2024 average, attributing it explicitly to cost discipline. That’s a company-disclosed, sourced signal of real burn reduction, not a presumption. I’m using a 25% cut rather than taking “margins doubled” at face value, because a literal reading of that disclosure would imply a steeper reduction than 25%, and the more conservative figure avoids building the model’s most optimistic interpretation into the result. That brings baseline burn to roughly $2.6 million a month. Second, Mono adds its own cost: roughly $1.7 million a year, or about $140,000 a month, per the January estimate. Combined working burn rate: $2.75 million a month.

The January model put Flutterwave’s cash position at roughly $35.5 million as of January 2026. Five months of burn at the new, reduced rate brings that to roughly $22 million by the time the Series E closed in June. Add the $100 million raise: a post-close cash position of approximately $122 million.

At $122 million and $2.75 million a month, that’s just over 44 months of pure runway, into early 2030. But companies don’t wait until the tank is empty. The January model itself noted that companies typically begin fundraising with 12 to 18 months of runway still in hand, and Nigerian and African late-stage fintechs, operating with thinner banking relationships and slower DFI disbursement cycles than their Silicon Valley counterparts, tend to sit at the cautious end of that range rather than the aggressive one. Strip out a 15-month buffer and the next capital event, whether that’s an IPO or a Series F, lands around November 2028. Call it late 2028, roughly 29 months from the round that just closed.

That figure lines up with something else worth naming: 24 to 36 months between late-stage rounds is the normal cadence for a disciplined-growth fintech right now, a sharp change from the 12 to 18 month gaps that defined the 2021-22 boom. A bottom-up cash model and a top-down read of how the market currently prices capital events both land in the same place. That convergence is what makes the estimate worth stating outright instead of hiding inside a four-year range.

None of this is precision dressed up as fact. The $100 million is an estimate built on one denied government figure and a judgment about how strategic investors size their checks. The burn figure leans on a company disclosure that didn’t include hard numbers. Two assumptions stacked on top of each other will always carry more error than either alone. A stated assumption that can be argued with does more work than a range wide enough to be true regardless of what actually happens.

The Pattern, Not the Exception

This isn’t unique to Flutterwave. Across the continent, the type of capital funding startups has shifted hard in the same direction. African startups raised $705 million in Q1 2026. Debt, structured instruments, and strategic capital accounted for roughly $490 million of that, well over half. A year earlier, the split ran the other way, with priced equity taking the lion’s share. Whether it’s venture debt, a strategic partnership, or a single undisclosed equity check from an infrastructure player, the common thread is the same: capital that doesn’t have to clear the test a competitive, disclosed, financial-investor-led round clears.

Flutterwave isn’t an outlier in this pattern. It’s the largest, most visible example of it. The company everyone is watching, doing exactly what smaller, less-watched African fintechs and logistics startups have been doing in funding rounds all year.

What This Means for the Marks You’re Holding

If you’re a GP holding African fintech exposure from the 2021-22 vintage, the temptation right now is to point at Flutterwave and say: see, the sector held. Be careful with that read. A held headline number, sourced from a single undisclosed strategic check inside a syndicate with every reason to defend its own marks, tells you who was writing the check and why. It’s weaker evidence of what an uninvolved buyer, with no stablecoin rails to gain, would actually pay.

The practical implication: every “flat round” headline from here forward deserves the same three questions asked of this one. Who’s writing the check, and what do they get beyond equity? Is the round size disclosed, or only the resulting valuation? And who in the existing cap table benefits from the number landing exactly where it landed? A number that survives all three questions is a real signal. A number that doesn’t is a press release with good PR.

Flutterwave is not in trouble. Not in any way shoe or form. The product is real, the revenue is real and growing, and the Ripple partnership may well be strategically sound on its own terms. But “held its valuation” and “the market re-tested the price” are two different claims, and this round is much weaker evidence for the second one than the headline number implies.

The same forensic question shows up across every asset class I underwrite, including ones a long way from payments infrastructure. Is this cash flow legible, disclosed, and independently tested, or is it being managed around a number someone needs to hold? Music catalogs. Contract receivables. Royalty streams sitting on African artists’ balance sheets that look, on paper, exactly like Flutterwave’s GMV. Large, real, and almost impossible to lend against until someone is willing to disclose what’s actually inside the number.

It’s the same question, asked of a different balance sheet.

I’m back from Hajj — grateful for the time away and glad to be returning to your inbox. This week’s piece picks up something I’ve been watching closely: what Drake’s recent release reveals about a structural shift in music IP that African artists and investors are about to face head-on.

On May 15, 2026, one artist held the first, second, and third positions on the Billboard 200 at the same time — the first time that had happened in the chart’s seventy-year history.

No conventional press cycle preceded it. No traditional radio build-up. Three albums, released in the same window, driven by an internal team operating around an artist-controlled imprint.

First-week numbers: 463,000 units for the lead project, 687,000 combined across all three. Streaming did the heavy lifting. Spotify confirmed Drake as the most-streamed artist, with ICEMAN as the most-streamed album on the platform in a single day in 2026 (Billboard).

The conversation that followed focused on the obvious questions.

Is Drake leaving Universal? Was the triple release designed to satisfy outstanding contractual deliverables? Does he sign with Sony? Does he renew with UMG? Does he go independent through OVO?

Those are interesting questions. They are not the important one.

The important question is what the rollout demonstrated, regardless of why it happened.

Share

The Bundle That Built the Industry

For most of the modern music business, a major record label bundled four functions into one institution.

It provided capital: advances against future royalties. It controlled distribution: first physical, then digital access to market. It handled marketing: radio, press, playlist relationships, promotional spend, audience conversion. And it retained catalog rights: the master recordings and long-term royalty streams that continued generating income long after the original advance was recouped.

An artist who wanted access to one function usually had to accept all four. That was the deal. The bundle was the leverage. Signing meant exchanging long-term asset ownership for short-term infrastructure, cash, and access.

Technology has been dismantling that structure for more than a decade. Distribution is now largely commoditised — any artist with a laptop can reach every major streaming platform globally for a modest annual fee. Marketing has shifted toward audience-owned channels, where artists with large direct followings can activate attention at a cost no traditional promotional team can easily match. A music-technology executive put it plainly to The Hollywood Reporter in late 2025: “The majors are going to evolve more toward services companies.”

That is the unbundling. The label is no longer the only institution capable of providing the functions it once monopolised. It remains the most convenient capital provider and, once rights are signed over, the default long-term catalog owner. But distribution and marketing have been pulled away from the old package.

Drake’s rollout shows what that looks like at the highest level. A team capable of executing a simultaneous three-album global release, sustaining attention across all three projects, and generating 687,000 first-week units without a conventional promotional cycle is not simply an artist relying on label infrastructure. It is an operating platform.

The labels understand this. UMG has moved deeper into the distribution infrastructure independent artists rely on, including assets connected to Downtown Music such as CDBaby, FUGA, Songtrust, and others. DistroKid, which reportedly processes a major share of new global music uploads, has been linked to a potential sale at a valuation around $2 billion. An artist who goes “independent” through one of those pipes may still be routing revenue through the same major-label ecosystem they thought they had escaped.

The labels’ second response will be contractual. Expect more “services” deals and broader participation structures that capture touring, merchandise, endorsement, brand, and ancillary revenue alongside recordings — expanding the label’s economic claim across an artist’s full commercial footprint rather than just the recordings. The exit route is being purchased. The bundle is also being widened. Artists and advisers who assume unbundling automatically favours talent are reading only half the board.

Two Assets, One Conversation

Most commentary treats “Drake’s value” as one number. It is not.

There are at least two separate assets.

The first is the catalog: historical recordings, publishing rights, streaming income, and royalty flows that accrue whether Drake releases new music or not. Before the 2022 UMG deal, Drake’s catalog was reportedly generating around $50 million annually for the label. That is an income-producing asset. It can be modelled, discounted, financed, and traded.

The second is the operating platform: the audience relationship, brand equity, OVO infrastructure, rollout capability, internal creative engine, and ability to create new commercially viable recordings at scale. That is the machine that produced the May 2026 numbers.

The difference matters.

A catalog is a cash flow pool. A platform is a cash flow machine. The first supports debt-style underwriting — royalty-backed facilities, structured credit, securitisation, predictable amortisation against known income streams. The second looks more like equity underwriting. It depends on execution, management depth, cultural relevance, release cadence, brand extension, and the continued ability to turn attention into revenue.

Conflating them is how bad deals get written. Price a platform like a catalog and you underwrite future upside too cheaply. Price a catalog like a platform and you overpay for optionality that never materialises.

That is the real billion-dollar question.

Industry speculation around Drake’s next deal often drifts toward a $1 billion+ headline figure (“B’s On The Table”). But that number only becomes economically defensible if the asset bundle extends well beyond recorded music. A serious counterparty would need to see something closer to this:

• Historical catalog income at a current run-rate of roughly $50–70 million annually

• Future recording rights across a defined release term

• OVO Sound economics and roster participation

• Publishing rights or participation

• Name, image, and likeness rights

• Brand licensing and merchandising

• Possibly some structured participation in touring or live-event economics

At a 15–18x multiple, a $50–70 million annual catalog income stream can theoretically support a valuation range around $750 million to $1.25 billion at the optimistic end — assuming durability, clean rights, and a buyer willing to pay for long-term scarcity. Add a 30–50% premium for platform-related assets — OVO infrastructure, brand licensing, NIL rights, future recording participation — and the upper end of the range closes in on $2 billion. But, applying a catalog-style premium to platform assets is a shorthand, not a methodology — platform underwriting is a separate discipline.

That is why the headline figure is rarely the economics. What closes is not “Drake is worth $1 billion.” What closes is a schedule of assets, rights, term, cash flows, control, recoupment, and future participation. The marketable number comes later.

Frank Ocean provides the cautionary precedent. Ocean executed a similar move in 2016: a contract-fulfilling visual release, followed immediately by Blonde as an independent project. In one move, he increased his revenue participation from 14% to 70% on that album. The ownership logic made sense. The commercial result was real. But his recorded output since then has been thin — closer to a short album’s worth of singles than a sustained independent operating machine. The independence worked, but the platform stalled. Those are different outcomes.

Independence is not a business model. It is a legal and commercial position. It only becomes more valuable than the surrendered catalog if the artist continues producing at the level that justified the freedom. And that is an underwriting question.

What African Capital Should Be Asking

This is not really about Drake. Drake is the data point. The argument is about a structural shift in how music IP is created, financed, controlled, and monetized — and that shift is arriving in African music at the exact moment institutional capital is beginning to take the asset class seriously.

According to Chart Masters estimates from January 2025, Wizkid generated around $1 million per month from Spotify. Burna Boy generated approximately $782,148. Tems, approximately $660,210. These are gross platform payout estimates, not artist take-home. An artist on a standard major-label deal receiving 14–20% of recording income does not receive the headline Spotify number. The larger share flows through the rights structure. The catalog producing that income belongs to whoever secured the rights when the deal was signed.

UMG’s majority investment in Mavin Global in February 2024 and Warner’s Lagos creative hub are not isolated events. The majors understand the direction of travel. Their response is to acquire infrastructure, catalog access, A&R capability, and market position before the next generation of African artists reaches the leverage point where independence becomes viable. The African artist approaching a major-label deal today may be signing at the moment just before the catalog’s long-term value becomes obvious to the market. Terms that look generous today may look extractive by 2032.

For institutional investors, the sharper question is this: when you invest in African music IP, are you buying a catalog, an operating platform, or both — and do you know how to underwrite the difference?

Most funds evaluating Afrobeats catalog acquisitions can (or at least should be able to) handle the first version. Model historical streaming income, apply an appropriate discount rate, stress-test concentration risk, adjust for FX where dollar-denominated royalties meet naira-based operating costs, arrive at a defensible valuation range. That is catalog underwriting.

Platform underwriting is different. It asks whether the artist, label, or creative company can continue creating new assets with repeatable commercial performance. The inputs are cultural traction, team depth, release consistency, brand extension potential, touring economics, audience ownership, management quality, and contractual control. Those variables do not sit comfortably inside a standard DCF. The funds that develop a credible method for underwriting that second asset will have a real edge in the next phase of African music IP investment. Right now, very few have it.

Before any African music IP transaction closes, three questions should be answered with precision.

First: who owns the masters, and under what reversion terms — if any — can those rights return to the artist? The headline deal value rarely answers this.

Second, and most important: is the artist’s operating platform being valued separately from existing catalog cash flows? If the platform is the appreciating asset — the machine capable of repeatedly generating new commercial output — pricing it as a catalog premium is the category error that will define which funds made money in this cycle and which didn’t.

Third: what is the chain-of-title across publishing, mechanical rights, neighbouring rights, producer contributions, samples, splits, and collection mandates? A catalog with unresolved ownership documentation trades at a discount. In some cases it is not investable at all.

The Label as Menu

The major label is not disappearing. It is becoming something else — less a single bundled counterparty, more a menu: capital provider, distribution utility, marketing services firm, rights administrator, catalog owner, strategic partner.

Artists with sufficient audience scale will increasingly choose which parts of the menu they need. That is not a death sentence for the labels. It is a compression of margin on deals where the artist no longer needs the full bundle. The catalog assets labels already hold from older full-bundle agreements may become more valuable precisely because those deals were signed before the unbundling accelerated. The next generation of deals will be negotiated differently.

For African artists, labels, publishers, and investors, the window to negotiate from that position is now — before the infrastructure consolidates further, before the next generation of Afrobeats catalogs is locked into terms that look generous only because the market has not yet repriced them.

The artist has become the platform. The question is who owns it, who finances it, and on what terms.

The midweek diagnostic works through both: the platform valuation framework this shorthand can’t fully capture, and the chain-of-title questions that determine whether a catalog is investable at all.

NASCON Allied Industries — the salt and seasoning manufacturer that anchors much of Nigerian household cooking — closed Q1 2026 with receivables sitting at more than 150% of its quarterly revenue. Read that number twice. NASCON is profitable, well-capitalised, part of one of Nigeria’s largest manufacturing groups. International Breweries, owned by AB InBev, is sitting at 60.7%. Champion Breweries at 46.7%. As I’m watching this from a buyer-side seat, what’s interesting isn’t that the receivables grew. It’s why, and what it says about where the working capital in the Nigerian economy is actually being warehoused.

The bank balance sheets that used to fund distributor working capital are paying down hard. FMCG firms collectively repaid roughly ₦1.2 trillion of debt across 2025. Nestlé Nigeria alone cut borrowings from ₦653 billion to ₦476 billion. Meanwhile the new capital rotating into African credit markets — TLG Capital’s $200 million Africa Growth Impact Fund, FCMB Asset Management’s pension-funded private debt vehicle, the AVCA-reported $2 billion private debt pipeline set to deploy by 2027 — is being directed predominantly into corporate term loans, mezzanine, and growth equity. The receivable piling up on Nestlé’s balance sheet has no institutional buyer at scale because the underwriting capability for it was never built locally at sufficient scale.

The Nestoil syndicate impairment essay two weeks ago named the supply-side of this story — Nigerian banks reallocating away from corporate underwriting after absorbing roughly ₦3.2 trillion in oil and gas-linked losses. This piece names the demand-side mirror. What happens to the asset classes a market can’t underwrite, once the capability for the wrong one has consumed all the available capital.

Share

The Working Capital Migrated. Nobody Marked Where It Went

Through 2023 and 2024, naira reforms and monetary tightening repriced bank corporate credit violently. The monetary policy rate — MPR, what banks pay to borrow from the central bank — climbed past 27%. The cash reserve ratio sat at 50%, meaning banks had to keep half of deposits sterilised at the central bank rather than lending them out. FMCG companies with FX-denominated debt absorbed severe finance cost hits. Servicing those debt stacks was destroying earnings faster than operations could compensate, so in 2025 the deleveraging began.

Two different bank lending books are shrinking simultaneously and for different reasons. The ₦1.2 trillion repaid by FMCG manufacturers is corporate debt the manufacturers themselves took on for capex, FX hedging, and balance sheet purposes. That book shrank because the interest cost became unbearable. A separate book — working capital loans to distributors — is also shrinking. The second book is the one this essay’s argument hinges on.

The demand side never recovered with the deleveraging. Nigeria’s Purchasing Managers’ Index — PMI, the headline forward indicator of business activity — slipped to 49.4 in April 2026, the first contraction reading after sixteen consecutive months of expansion. Distributors who used to pay upfront for inventory cannot. The choice manufacturers face is ugly and simple. Tighten credit to distributors and watch volumes collapse. Or extend it and absorb the working capital onto their own balance sheet. They’ve chosen the second.

Follow the cash through three parties. Nestlé manufactures and ships to a distributor — a mid-sized business that buys in bulk and resells to thousands of retailers across a region. The distributor collects from retailers over 30 to 60 days. Until recently, the distributor bridged that gap with a working capital loan from a bank — First Bank’s FMCG Key Distributorship Finance is one example. Nestlé got paid in 14 days. The bank earned interest on the bridge. Everyone’s economics worked.

What broke is the middle link. With MPR above 27% and naira lending rates following, the working capital loan that used to cost the distributor around 20% now costs 30–35% — beyond what FMCG distribution margins can absorb. Distributors asked the manufacturers for longer payment terms. To defend volume, the manufacturers said yes. Now Nestlé ships on day 1 and collects on day 60. For those 60 days, Nestlé is performing much of the credit function the bank used to perform — without the underwriting tooling and without explicit pricing of the balance sheet burden it has absorbed.

The receivable hasn’t disappeared. It has migrated from one balance sheet to another. Before, it sat as a loan asset on the bank’s books, with the distributor carrying it as a liability. Now the bank’s loan book has neither side of that entry, the distributor’s borrowing line is smaller, and Nestlé carries a large trade receivable on its own balance sheet. The migration is economically inefficient on its new home because Nestlé is holding working capital at its full cost of capital and earning nothing for it directly. A credit specialist holding the same receivable would carry it at a lower cost of capital and get paid a spread for performing the function. Nestlé pays the holding cost primarily to preserve distribution continuity and volume; a specialist pays a smaller holding cost and earns an explicit return for warehousing the risk itself. The asymmetry is the whole point.

NASCON’s 150%-of-revenue receivables ratio is the visible expression of that migration. International Breweries and Champion confirm it as a pattern. A skeptical reader could argue the receivables growth reflects benign factors — strategic payment-term extensions, accounting policy changes, revenue mix shifts. The simultaneous PMI contraction, sector-wide debt repayment, and named-firm consistency make the distress interpretation the better one. The squeeze is real; it just hasn’t fully translated into reported impairment yet. The Camel logic this publication has worked through before applies in reverse here — capital-efficient design is what manufacturers thought they were building when they deleveraged. Instead they’ve absorbed the kind of balance sheet risk the framework warns against, because no credit specialist was available to take it at workable pricing.

Private Credit Came to Africa. It Brought the Wrong Playbook

The new private credit capital arriving in Africa is good news. It is also, predominantly, the wrong product for the gap that actually exists. Three reasons, all observable from fund mechanics themselves.

Ticket size first. A $200 million fund needs to deploy in $5–20 million tickets to keep diligence costs proportionate and portfolio concentration manageable. Receivables financing structures clear at $500,000 to $5 million per facility in this market. The math doesn’t work for the fund’s GP economics, so the fund pushes upmarket into corporate term loans where it can write cheques that justify the diligence cost.

Tenor is the second mismatch. TLG’s stated product offers seven-year tenors with three-year grace periods. That’s growth capital. Receivables financing is 30 to 120-day revolving credit, with the borrowing base — the pool of eligible receivables backing the facility — recalculated continuously. A fund designed for one cannot pivot to the other mid-life.

The third reason runs deeper. Receivables underwriting requires assessing the buyer of the receivable — the FMCG manufacturer or telco offtaker who owes the money — rather than the seller who is borrowing against it. African credit markets, both bank and private, have spent decades building underwriting capability for sponsor-cashflow risk — the same capability that produced Nestoil and the ₦3.2 trillion of bank impairment that followed. The asset-class-specific capability for receivables — whether trade, contract, or IP-linked — has not been built at scale. With limited exceptions worth naming.

Afreximbank has been promoting factoring across the continent for over a decade. South Africa hosts most of the developed factoring volume. The Nigerian SEC’s 2025 rules on private debt issuance opened a regulated channel for receivables-backed instruments. Bibby Financial Services launched an Africa platform. None of this is invisible. It just hasn’t reached the scale that would absorb the receivable volume now sitting on FMCG balance sheets.

OmniPay — the financial arm of the OmniRetail B2B distribution platform — reportedly processes around $95 million in monthly transaction volume, extends credit at roughly $4 million monthly against the same receivable universe, and reports default rates below 1%. Set those numbers against TradeDepot’s $110 million raise anchored on BNPL for five million SMEs, and Alerzo’s 2021 lending product — both reportedly suffered significant losses on credit and paused programs to regroup. Same retailers. Same receivables. Same macro environment. Radically different default outcomes.

Almost nobody outside operating B2B circles is reading this gap as a credit signal. The most plausible structural explanation is the underwriting architecture itself. OmniPay sees transaction flow on its own platform, documents receivable ageing in real time, and underwrites the offtaker risk — Tolaram’s MultiPro subsidiary, which moves through OmniRetail’s distribution — rather than the retailer-as-borrower risk the failed lending programs were chasing. The distinction between underwriting the credit of the entity buying the goods versus the entity selling them remains the architectural difference the rest of the market is missing.

The Underwriting Is Knowable. The Talent Isn’t

A receivables facility is straightforward at the core. Somebody buys the right to collect a future payment. Advances most of the cash now. Absorbs the risk that the payment comes in late or short. Everything else is architectural detail around that core trade.

Applied to the FMCG case, Nestlé — not the distributor — becomes the originator. Nestlé sells its trade receivables to a specialty credit fund at a discount to face value, gets cash on day 1, and the fund collects from the distributor on day 60 and earns the discount as its return. Nestlé’s balance sheet returns to the old shape — cash, not receivables. The credit function moves to a specialist whose cost of capital is lower than Nestlé’s and who gets paid for performing it. The same architecture applied to a music catalog works in the same direction — the artist or label sells its forward royalty receivables to a specialist who advances cash today and collects from the streaming platforms and PROs over the licensing tail. Different asset class, same structural move.

Advance rates typically run 70–85% of eligible receivable face value, depending on debtor concentration, dilution experience, and obligor credit quality. A dilution reserve absorbs returns, allowances, and disputed invoices. Concentration limits cap exposure per debtor and per industry. Pricing builds off MPR plus a spread that reflects underwriting capability and operational cost rather than equity-style risk premium. Currency exposure is layered carefully — naira receivables funded with naira capital, FX overlay only where the underlying receivable is export-linked.

Priced correctly, this structure generates 18–24% naira returns on an asset with verifiable cash flow performance. The OmniPay default experience — sub-1% against $4 million monthly deployment — shows what the asset class can achieve when underwritten properly. That’s a better risk-adjusted return than the corporate term-loan paper the private credit funds are buying, if the underwriting and servicing infrastructure exists to capture it.

Where structures fail in practice is at four specific points. How eligibility is defined — does a delivered-but-disputed invoice count, and under what cure terms? How dilution is reserved. How perfection of security is established under Nigerian law, against the backdrop of the Secured Transactions in Movable Assets Act and the National Collateral Registry — and how assignment is notified to the underlying obligor to defeat set-off and netting risk if the distributor pays the wrong party in ignorance. And how cash trap mechanics actually operate when payment is deferred — who controls the collection account, what triggers a sweep, how the waterfall runs when receivables come in late or short. Each is solvable. None is solved by capital arriving.

The Underwriting Talent Question Is the Deeper Story

The Nestoil essay named that Nigerian bank corporate-underwriting capacity is reallocating. This essay names what the new capital is walking past on its way to replacing what’s reallocating away. The reallocation question isn’t only where the bank capacity goes — it’s what the next generation of African private credit chooses to underwrite, and whether that choice differs in kind from what the banks were doing before. On the current trajectory, it largely doesn’t.

The African VC DPI gap — distributions to paid-in capital, the cash returned to investors after fund fees — and the African private credit underwriting gap are likely related phenomena. Pattern observation more than proven mechanism, but consistent across enough asset classes to be worth naming. Capital deployment is outpacing the development of asset-class-specific underwriting capability. Whether the asset is a software company’s exit multiple, a music catalog’s streaming royalty stream, or a distributor’s receivable book, the question is identical. Who has built the capability to value this thing properly, and what happens when the capital arrives before the capability does? Both markets are answering that question in real time. The answer matters for which capital pools end up profitable a decade from now — and for the founders, artists, and operators whose assets are being valued by buyers who may or may not know how to value them.

I see this from two sides — corporate buyer-side underwriting on one, IP and contract receivables on the other. The pattern is the same. The asset class itself doesn’t matter as much as whether the institutional capability exists to value it.

What This Means

The private credit narrative being told about Africa right now is a capital-supply story. The data supports a different one — an underwriting-capability story. Capital is necessary but not sufficient. The funds, advisors, and operators who will produce real returns from African capital markets over the next decade won’t be the ones that arrived with the biggest cheques. They’ll be the ones that built — or hired, or partnered for — the specific capability to value the asset classes the continent actually generates in volume.

This question runs wider than credit. Anyone deploying capital, time, or attention against an asset whose value isn’t yet legible faces the same problem. The artist whose catalog earnings are growing but whose label can’t structure against them. The founder whose unit economics are real but whose investors can’t price what comes after profitability. The CFO watching receivables accumulate that no bank will fund at workable terms. Each is a different version of the same gap.

Nestlé doesn’t want to be a bank to its distributors. The receivable will eventually be warehoused, financed, and priced by somebody. The question is whether that somebody is a Nigerian institution that built the capability locally, or a foreign specialty fund that priced in the capability shortage as a risk premium and clipped the spread for itself.

A short note before I close. The next L.U.M.I. Brief lands Saturday 6 June. I’ll be in Mecca for Hajj for the two weeks between. The next piece will pick up where this one ends — the same architectural question applied to music IP specifically. Catalog earnings sitting on artist, label, and manager balance sheets the way these receivables are sitting on Nestlé’s — visible to anyone who looks, unstructured at scale because the capability hasn’t been built. I’ll write that one when I’m back.

Until then, the data this essay rests on is public. NASCON’s receivables ratio is in their Q1 statement. The OmniPay figures have been reported. The FMCG deleveraging is in audited filings. Pick any one of those threads and pull on it. The structural picture this essay names becomes more visible the closer you look at any single piece of it.

— Lumi

Two weeks ago a corporate buyer’s CFO called me about a target they were two months into diligencing. The enterprise value model was finished, the synergy case was lined up, and the board was ready to move.

They wanted me to sign off on the loan documentation.

By the end of the call we were having a different conversation. The target wasn’t really a company being acquired. It was a balance sheet of legacy debt with an operating business sitting on top of it. Three syndicated facilities, all originated through one of the Nigerian banks now absorbing 2025 impairment charges. The buyer thought they were paying for revenue and customers. They were also paying — without knowing it — for what the seller’s lender had been deferring.

That afternoon repriced the deal by 18%. The company hadn’t changed. The underwriting market the company was sitting inside had changed, and nobody on the deal team had noticed.

Nestoil is the proximate event. The indigenous oil and gas major defaulted on roughly $2 billion of syndicated debt to a consortium of Nigerian and African banks. In October 2025 the lenders secured a Mareva injunction — a freezing order that locks a borrower’s assets across multiple institutions pending recovery — covering Nestoil’s accounts, properties, and oil cargoes across more than 20 financial institutions (Nairametrics, November 2025). The company is now in receivership and litigation.

The press is reading it as a banking-stability story. Three tier-one banks suspending dividends. About N2.16 trillion of impairment across five lenders (TheCable, May 2026). Mareva injunctions, receivership disputes in the Federal High Court.

Read at the underwriting layer, that framing misses, to my mind, the actual event. Roughly $2 billion of corporate lending demand that used to live inside Nigerian bank syndicates is unlikely to return there at pre-2025 volume in the next 24 to 36 months. Where it migrates, and on what terms, decides the next decade of African corporate finance.

Four people who should be paying attention

The GP whose 2022 portfolio company took on a Nigerian bank facility. That facility comes up for refinancing into a market where the original lender is operationally constrained for the next 18 to 24 months. Most replacement lenders I’m seeing price wider, with covenants the original facility never carried. The exit window narrows to the buyers who can absorb that capital stack.

The founder raising debt right now. AVCA data places the Kenyan venture debt market at roughly $498 million deployed in the most recent reporting year, against about $160 million in Nigeria. Comparability between the two markets isn’t exact, but the directional gap is what matters here, and that gap was already built into how the two markets are organised — I wrote about why in the DFI letterhead piece. With Nigerian bank corporate underwriting paused, the gap likely widens before it closes. The capital that fills it underwrites against your contracted revenue, not your relationship with a credit officer.

The strategic acquirer or institutional buyer sitting on dry powder. A window has opened. African corporate assets are being priced by sellers and by their existing lenders, both of whom are working off frameworks that no longer apply. Buyers who can underwrite intangibles, synergy, and jurisdiction-specific enforceability are pricing into a market the legacy banks are not currently positioned to serve.

The banker reading the 2025 results. The CBN’s forbearance unwind directive of March 12, 2026 ended the regulatory mechanism that let Nigerian banks defer loss recognition on legacy oil and gas exposures (Nairametrics, March 2026). The N21 trillion in sector exposure at end-2024 (TheCable) is now being marked. The post-forbearance market that emerges is likely to look meaningfully different from the one that preceded it.

A wider rewiring is underway. The global capital transmission system that funded the last cycle is breaking down. The work I covered on chain-of-title forensics in African music IP applied this same underwriting discipline to one of those untouched asset classes. And resilience-first venture design is among the few designs built to survive a transition like this. The Nestoil event is one node in that rewiring. It happens to be the noisiest one.

Share

What the legacy model priced

Three things.

Relationships. The Common Terms Agreement governing the largest exposed loan was signed in 2022 (Nairametrics, November 2025). A Common Terms Agreement, or CTA, is the master contract that ties multiple lenders into a single syndicated facility — it’s where the covenants, default triggers, and inter-creditor mechanics actually live. The credit history likely sat on a banking relationship that pre-dated the 2022 CTA by years. Pricing was relationship-credit, not asset-credit. Migrating a relationship-priced loan book to a lender who doesn’t have the relationship rarely works cleanly.

Oil price assumptions. The original financing for the underlying asset was structured in 2012 around oil-price stability and predictable production. When the Forcados pipeline was bombed and the terminal went offline for sixteen months between 2016 and 2017, the syndicate didn’t reprice. Forbearance held the position for the next eight years. The loan was held at a value the asset’s cash flow couldn’t actually support.

Forbearance as a substitute for workout. The CBN’s forbearance regime let banks treat impaired loans as performing through restructuring. What it did not do was force a workout. The N3.2 trillion sector-wide impairment that hit in 2025 (Businessday, May 2026) is the price of nine years of deferred loss recognition.

What the emerging model has to price

Three different things.

Cash flows, not relationships. A music catalogue throws off royalty cheques on a specific schedule, in specific currencies, from specific counterparties. A contracted revenue book has concentration risk you can quantify per customer. A target’s working capital has FX exposure you can model per leg. Underwriting these means opening the cash flow line by line — the kind of diligence framework Nigerian bank syndicates didn’t generally need to build under the incumbent underwriting structure.

Strategic synergy. A target’s revenue is worth more inside a buyer’s distribution network than it is on a standalone enterprise value model. The corporate transaction work I’m running right now lives in this gap. The asset prices differently inside the buyer’s stack. That delta is the underwriting margin.

Jurisdiction-specific enforceability. Where a loan is enforced shapes what recovery costs and how long it takes. Mareva injunctions, receivership proceedings, eight banks in the Federal High Court in Abuja arguing about a Common Terms Agreement (The Economic Times, November 2025) — these are the post-default mechanics. The legacy model priced enforceability as a standard clause. The cash-flow-priced model has to price it as a line item.

Two scenarios from current work

The framework is easier to see in deals that are running right now. Both of the next two are drawn from buyer-side underwriting work currently on my desk — composite, with identifying details stripped out. The first picks up where the opening of this piece left off. The second extends the chain-of-title forensics work into a wider asset-class scope.

The corporate buyer.

Back to the CFO’s deal. The strategic acquirer was modelling the target on enterprise value multiples. The target’s loan book carried three syndicated facilities, all originated under a CTA that ran cross-default triggers at the parent guarantor level. Cross-default triggers mean an event of default on one facility automatically calls every other facility tied into the same CTA — the legacy lender hadn’t pulled them, but they were live.

After the March 2026 CBN directive, the practical discretion lenders had under forbearance narrowed sharply. The buyer had to underwrite to the assumption that anything that could be called could be called. The transaction price moved 18%. The target wasn’t operationally weaker. The underwriting market the target was sitting inside had repriced underneath the deal, and the buyer was the one absorbing the markdown the legacy lender had been deferring.

The 18% delta is, in my read, what the prior-cycle lending architecture didn’t price. The buyer paid it on day one of the new regime.

The institutional buyer.

An international institutional buyer underwriting a portfolio of African catalogue rights. Same diligence pattern as the chain-of-title work — every track traced back to its actual rights chain, every revenue stream traced to its actual counterparty.

The buyer’s credit committee asked the question Nigerian bank syndicates rarely asked of any oil and gas exposure. What does the cash flow actually look like, line by line, by counterparty, by jurisdiction, over a five-year horizon?

The answer required a 22-page underwriting memo. Most of it was about enforceability and counterparty risk in two specific Nigerian and Kenyan jurisdictions. A legacy bank syndicate would likely have priced the same exposure on a relationship-credit basis with a one-page covenant package and called it project finance.

The pattern across both, as I’m seeing it: the underwriting capacity that lived inside Nigerian bank syndicates is migrating to underwriters who price against actual cash flows, against strategic fit, against the specific jurisdiction the deal is enforced in. The migration is happening at deal volume, not in theory.

What the next 24 to 36 months decide

The reason this window matters is that the categories being decided in it are categories that, once set, are slow to revisit. Underwriting frameworks calcify around the deals that get done in their early years. Asset classes that get underwritten in the next 24 to 36 months become the ones that have established pricing benchmarks, recognised diligence patterns, and proven enforcement precedents. Asset classes that don’t, won’t — for a long time after.

Three things get decided in this window.

Who builds the next underwriting market is the first. The Nigerian bank syndicate model isn’t returning at pre-2025 volume in any near-term timeframe I can see. The capital that returns comes through different vehicles — international strategic acquirers, IP-focused funds, regional credit platforms, dedicated new-economy buyers. Some are already deploying; a meaningful share are still being structured.

Which asset classes get underwritten is the second. The asset classes likely to get underwritten next are the ones the legacy banking framework couldn’t price. Music catalogues. Contract receivables. Strategic acquisition targets. Hybrid debt-equity instruments tied to specific cash flow architectures. Indigenous oil and gas project finance is in multi-year retrenchment.

The third is which framework survives. The legacy framework — relationship-priced, forbearance-tolerant, project-finance-templated — is the one that produced the N21 trillion sector loan book now absorbing N3.2 trillion in 2025 impairments. The framework retires alongside that loan book. What replaces it prices the cash flow line by line, prices the buyer’s strategic fit into the asset, and prices enforceability against the specific jurisdiction the deal sits in. None of these are new in global capital markets. They are new at any meaningful scale in Nigerian and broader African corporate underwriting.

The reframe, in my read: the 2026 Nigerian corporate-underwriting reallocation isn’t the end of African corporate underwriting. It’s the beginning of the underwriting market that runs the next decade. The banks remain operationally strong — Q1 2026 numbers across the major players make that clear. What’s happening is narrower and more important: the category of large-ticket corporate underwriting that ran through Nigerian bank syndicates in the last cycle is being rebuilt, and most of the rebuilding is happening outside those syndicates.

The buyers and underwriters who price correctly into this window will be the ones intermediating the deals in 2030. The ones still working off the old framework will spend that decade explaining their losses.

The work I’m running right now sits on both sides of this. Strategic transaction underwriting for corporate buyers absorbing existing capital stacks. Asset-class-specific underwriting for institutional buyers pricing intangibles, contract receivables, and hybrid instruments where the legacy bank framework didn’t engage with the underlying cash flow.

The work isn’t about replacing the bank. It’s about pricing what the bank didn’t price and isn’t, in the near term, returning to.

If you’re sitting on either side of a transaction the legacy underwriting model would have intermediated and now does not, the conversation is worth having sooner rather than later.

[lumi@lumimustapha.com — Get in touch if you’d like a free 30-min Ecosystem Intelligence Briefing, by application only]

​​​​​​​

The Saturday essay named the diligence wall. This piece walks through what actually happens at it — the playbook international institutional investors and strategic acquirers run when they diligence an African music catalog they are considering acquiring or financing against. I have run versions of this playbook on more buy-side mandates than I can quickly count over the past three years, both directly and through Technolawgical Partners’ deal practice. The structure below is the operational version of the four-test framework, written for the seller-side reader who wants to know what the buyer will be looking at, in what order, and using what tools.

The thesis is simple. The same playbook the buyer will run on you is the playbook you can run on yourself first. Catalogs that arrive at diligence already remediated close at headline value, in weeks rather than months, with terms structured around opportunity rather than risk allocation. Catalogs that arrive raw close at a fraction of headline, after extended legal workstreams the seller usually ends up paying for through the deal economics. The difference is preparation.

Phase One — Pre-Mandate Catalog Mapping (Weeks 1–2)

Before any diligence team opens a data room, the lead deal lawyer does a structural map of the catalog. The map is not about valuation. It is about what is being sold and who actually owns it.

Three documents get pulled first. The full track schedule, with songwriter and performer credits per work. The corporate ownership chart of the entity holding the catalog. The list of distribution and administration relationships currently in place — which distributor is uploading the catalog to which platforms, which collection societies are registered, which sub-publishers are administering sync.

A competent buyer-side lawyer can usually tell within the first 48 hours of receiving these three documents whether the catalog sits in the bottom quartile (significant chain-of-title gaps, expect heavy remediation), the middle 60% (typical African catalog with some structural defects, expect moderate remediation), or the top 15% (institutionally-grade documentation, ready for facility deployment). The category determines the rest of the workstream timeline and shapes the deal economics.

What the seller can do in this phase: produce these three documents proactively, at the same level of detail the buyer will demand. If you cannot produce a clean track schedule with songwriter and performer credits per work, you do not yet have a catalog institutional capital can underwrite. You only have a body of revenue.

Read more

The deal had been moving for four months. A strategic acquirer building an African catalog allocation had cleared every commercial gate. The seller, a credible Nigerian label with twelve years of operating history and a catalog the buyer’s own appraiser had valued at $38M, had signed the term sheet, opened the data room, walked the buy-side team through the artist roster in person.

Then the diligence team mapped the chain of title.

Across the catalog’s 800-odd works, the label legally owned masters on roughly 60%, publishing rights on 15%, and held only economic rights — not legal title — on the rest. The other rights sat with three offshore publishers, a defunct collection society, and in twelve cases with the original songwriter’s estate, which had never signed the assignment deed everyone in the room had assumed existed for a decade.

The deal closed at $14M against the clean subset. The remaining $24M of value did not vanish. It just became unbankable until two years of forensic legal work resolved who actually owned what.

I have been on the buyer side of this exact scenario more than once over the past three years, running diligence on African music IP for international institutional investors and strategic acquirers. Across Technolawgical Partners’ deal practice, which has been involved in a substantial body of African music IP, fund formation, and creative-economy transactions, the pattern is now stable enough to name as a structural feature of the asset class rather than an unfortunate exception. The $290M untapped African music IP lending market this publication sized in January is not gated by bank willingness, by rate compression, or by direct payment infrastructure. It is gated by chain of title.

The first $50M of that market gets unlocked by lawyers, not by lenders.

What This Means For The People Reading

Label principals holding catalogs argued at 8–12x earnings — including in this publication — are sitting on assets where the realisable institutional value, post-diligence, often lands closer to 2–4x. Same multiple gap LumiBrief has been calling for two years. Different reason. The gap is documentation, not market perception.

Banks that built credit models around the direct payment thesis are finding that Letters of Direction cannot be issued on works where the borrower’s standing to issue them is ambiguous. The January essay’s 950 basis-point rate compression assumes the letter executes. In the buy-side mandates I have run, more than three-quarters of pitched catalogs fail at the standing test before the credit team gets near the rate math.

IP funds and DFI cultural finance teams quoting the $6–10B continental catalog value figure are working off an aspirational ceiling. The realisable institutional-grade slice today sits closer to $80–150M. Not because the music isn’t worth the larger number. Because what is enforceable today is a fraction of what is claimed.

Founders and creators signing artist-label-publisher deals in 2024, 2025, 2026 are creating tomorrow’s problem. Every standard-form African contract written in this window with ambiguous master, publishing, neighbouring, and sync allocations is an unfinanceable asset waiting to be discovered in someone’s diligence eight years from now.

This essay is the seventh diagnostic in a sustained argument running through this publication over the past eighteen months. The macro layer named the eurodollar machine that funded Africa’s tech boom breaking down. The market-comparison layer named what Nairobi got right while the rest of the continent struggled — and why the enclave development model (versus national, regional or continental) model is optimal for African development. The public markets layer named the pricing failure that compounds the macro: African private rounds anchored to public market comps that no longer exist on the tape. Each diagnosed a structural defect in how capital is priced or transmitted. This one names the same defect inside a different asset class — one this publication has argued can hedge the FX dynamics that make everything else mispriced. The argument holds. The next layer down is the one that determines whether any of it can actually be financed.

The Four Rights Problem

Every recorded song is not one asset. It is at least four.

Master rights — the recording itself, typically held by the label. Publishing rights — the underlying composition, melody and lyrics, held by the songwriter or assigned to a publisher. Neighbouring rights — performance rights collected by collection societies (in Nigeria, COSON or MCSN; in Kenya, MCSK; in South Africa, SAMRO). Sync rights — licensing for film, television, advertising, often administered by a sub-publisher.

In a properly structured Western catalog, ownership of all four flows back through documented assignment deeds to a single beneficial owner with clear standing to license, monetise, and pledge the asset. The diligence work to confirm this exists has been industrialised over forty years.

In a typical African catalog, the four rights are owned by four different entities. Three of them may not have written contracts that name the others. The economic owner — the artist or label receiving net royalties — and the legal owner, whichever party can produce a registered assignment, are frequently distinct. The party with standing to issue payment instructions on each revenue stream depends on which right the revenue is being paid against, which the platform itself may not have documented correctly at upload.

The Diligence Wall

When an institutional lender or IP fund attempts to finance an African catalog at scale, it runs four tests. In that order. Test 4 only matters if Tests 1, 2, and 3 pass.

Test 1 — Chain of Title. Can the borrower produce signed assignment deeds for every work in the catalog, demonstrating legal ownership flows from songwriter or performer to current claimed owner with no gaps? Failure looks like the opening scenario: 60% clean masters, 15% clean publishing, 25% nominal claim with no executed paperwork.

Test 2 — Standing. Does the borrower have legal standing to issue payment instructions on each work? On masters, yes — assuming Test 1 passes for that work. On publishing, often no; the publisher does. On neighbouring rights, the collection society does. Failure looks like a borrower who controls 100% of the artist relationship but only 40% of the rights that produce the revenue.

Test 3 — Jurisdiction and holding entity. Is the holding entity domiciled in a jurisdiction where the assignment is enforceable, the payment instruction is irrevocable, and security can be perfected? A Nigerian-domiciled holding entity with Nigerian-law-governed assignment deeds faces a different test than a Mauritius or Cayman holding with English-law deeds. Failure looks like a catalog held by an artist’s personal-name sole proprietorship — perfectly clean rights, no entity structure capable of receiving an institutional facility.

Test 4 — Currency basis and counterparty. Is the catalog’s revenue paid in hard currency, and through what counterparty does the payment intercept actually run?

This last test is where the January essay needs a refinement worth handling directly, because it matters operationally.

That essay framed the Letter of Direction at the platform tier — issued to Spotify, Apple, YouTube. That holds for major label and major publisher arrangements, where the platform deals with the rights holder directly. Across most of the African catalog universe, the operative counterparty for most catalogs is not the platform. It is the distributor — DistroKid, TuneCore, AWAL, Believe Music, Empire, CD Baby, Africori, Mdundo. The distributor receives the platform remittance, takes its cut, and routes the net to the rights holder.

The structural mechanism of payment intercept holds either way. The credit profile of the intercept point shifts. A Letter of Direction to Spotify is enforceable against an investment-grade balance sheet. A Letter of Direction to Africori — which since the Warner Music majority acquisition in 2022 carries a meaningful corporate backstop — sits closer to that profile than most observers realise. A Letter of Direction to a smaller independent distributor with no parent backing is a different credit instrument entirely. The diligence work has to test the distributor’s contract with the platform, not just the rights holder’s contract with the distributor. Most of the time, on the catalogs I have diligenced, the second test passes and the first does not.

The point is that direct payment is a real mechanism. The counterparty grade varies more than the original framing implied. The Wednesday companion walks through the distributor-tier specifics and how Letters of Direction get drafted differently for distributor counterparties than for platform counterparties.

Share

What Actually Happens In The Room

Three composite scenarios drawn from buy-side mandates I have run, with identifying details removed and numbers approximate but representative.

The label that discovered it owned less than it sold. Nigerian label, twelve-year operating history, around 600 works claimed in the data room. Pitched the catalog at $35M to a regional credit fund. Diligence found assignment deeds for 380 works on master rights only. Publishing assignments for 90 works. The remaining 130 works had artist contracts that referenced “standard publishing splits” in commercial terms but had no executed publishing assignment deed registered anywhere. Translation: clean title on roughly 63% of what was pitched. Bankable catalog value at standard institutional discounts: $14–18M, not $35M. Deal restructured as a $12M facility against the clean subset, with an 18-month workstream to perfect title on the remainder. Cost of the title perfection workstream: around $280K in legal fees. Cost of not doing it before the pitch: $20M of value sitting in the room and unable to find its way into the facility.

This is the Test 1 failure. By far the most common.

The publisher whose Letters of Direction don’t stick. Pan-African publisher administering around 2,000 works. Bank attempting to finance against the publishing income stream. Letters of Direction drafted to the relevant distributors. Acknowledgements came back on 1,200 works. The remaining 800 were flagged because the underlying songwriter assignments to the publisher were either not registered with the relevant collection society, registered with conflicting splits, or registered to a different publisher entirely — residue from prior administration deals that were never properly terminated when the songwriters moved over. The Letter of Direction was technically valid. Commercially unenforceable on 40% of the catalog. Bankable revenue base: 60% of headline. The remaining 40% became a nine-month forensic clean-up before any of it could be financed.

This is the Test 2 failure. It catches publishers and administration companies most frequently, because they are accustomed to operating on commercial assumption rather than registered title.

The catalog with perfect documentation and no standing. Artist holding masters, publishing, sync, and neighbouring rights — all properly assigned, all documented, all enforceable. Pitched as solo borrower for a $4M facility against the catalog. Diligence identified that the holding vehicle was the artist’s personal-name sole proprietorship. Lender could not perfect security against an unincorporated person without exposing the artist personally to enforcement action across multiple jurisdictions. Restructure required: incorporate an IP HoldCo, assign all four rights to the HoldCo, perfect security against the HoldCo. Cost: around $15K and six weeks. Cost of not doing it before the pitch: roughly $200K in lender legal fees the artist was asked to underwrite, plus four months of additional delay.

This is the Test 3 failure. The smallest of the three categories, the cheapest to fix, but it kills deals at the eleventh hour with frustrating regularity because the holding-entity question is rarely raised in the commercial conversation that precedes the diligence.

The pattern across all three: the asset is real, the value is real, the cash flow is real. Bankability is gated by legal architecture work that has not been done.

The January essay’s 69% Year 1 return on allocated capital for direct payment lending assumes the lender deploys the capital. The capital does not deploy because the diligence wall stops most pitched catalogs from clearing the four tests. The constraint sits on the legal architecture side of the asset, not the credit side of the lender.

What This Phase Is For

The corpus has built a six-part argument over eighteen months: asset valuation, asset class framing, credit implementation, systemic capital design, development model, strategic doctrine. Every layer assumed the underlying assets were institutionally bankable in their current state. They are not, yet. The seventh layer, the one that makes the rest operational, is the legal architecture layer.

This is not a setback for the thesis. It is the thesis’s next phase.

The path from $80–150M of currently bankable African music IP to the $6–10B addressable catalog value runs through a workstream with three rough sequences. Years one and two: chain-of-title forensics and remediation across the top fifty institutionally-pitchable catalogs continent-wide. Estimated cost in legal architecture work: $8–12M. Resulting bankable AUM uplift: $400–600M. Years two through four: standardisation of new artist-label-publisher contract architecture so that catalogs created from 2027 onward are bankable by default — industry standard-form contract development, collection society interoperability work, holding-entity templates. Years three through five: securitisation infrastructure built on the cleaned base, with credit ratings methodology calibrated for African catalogs.

The $290M untapped market opens not when banks decide to lend, but when the legal architecture work is done on the asset side. The two are sequenced. Diligence remediation precedes credit deployment, not the other way around.

The label principals, publishers, and IP fund originators reading this who want to be first to the institutional capital that will follow have the same 18–24 month window the January essay flagged for the lenders. The work to do in that window is not credit modelling. It is title remediation.

What’s Next

The midweek companion to this essay will be a buyer-side field guide — how international institutional investors and strategic acquirers actually diligence African music IP, written from the experience of having run those mandates on the buy side. Useful if you are sitting on a catalog you intend to finance or sell, because the playbook the buyers will run on you is the same playbook you can run on yourself first.

For label principals, publishers, or IP fund originators who want this work done on a specific catalog before pitching institutional capital, this is what fractional general counsel engagements exist to do. A typical chain-of-title remediation runs six to twelve weeks and unlocks two to three times the bankable value of the asset. The work I have done on the buy side is the same work, applied earlier in the chain.

The asset class is real. The capital is real. The seventh layer of work to connect them is what this next phase is.

Saturday’s essay made the structural case: African Series B and C rounds are being priced against comp baskets that compressed 70% since 2021, and the same ghost numbers sit inside GP portfolio marks. The reset is mechanical. And the unwind is starting.

This piece is the practitioner version. Three pieces of work, in sequence — the segmented comp baskets that should replace the generic SaaS bundle, the discount stack methodology that converts public comps to defensible private valuations, and the six-question audit checklist for portfolio marks. Worked numerical examples throughout. Practical applications at the end.

The Three Archetype Problem

The most expensive mistake in current African fintech pricing decks is comp-basket conflation. A lending fintech, a payments processor, and an embedded finance B2B SaaS play are three structurally different businesses with three different comp universes. They are routinely priced against the same generic “fintech SaaS” basket.

Archetype 1: The Payments Processor

Payments processors sit on top of payment rails — card processing, mobile money rails, account-to-account transfers, bill payment, agent banking. Revenue is take-rate on transaction volume, typically 0.3% to 1.5% depending on rail and geography. Customer concentration risk is high (top 5 merchants often >40% of revenue). Capital intensity is moderate (settlement float, agent network buildout).

Public comps that work: StoneCo (Brazilian payments, currently 0.5x EV/Revenue, 6.2x EV/EBITDA), PagSeguro (Brazilian payments, similar range), Adyen (European payments premium, currently 8.1x EV/Revenue, 15.2x EV/EBITDA — note Adyen has compressed from 124x EBITDA in 2021 to 15.2x today). The honest comp range for an African payments processor at scale is 0.5x to 4x revenue depending on growth rate, margin profile, and geographic concentration. Above 4x revenue requires Adyen-level margin profile or genuine growth premium that survives diligence.

Important precision: the multiple comparison only works if both companies report on comparable bases. Many African payments fintechs book gross transaction volume as revenue; StoneCo also books gross. Net revenue / take-rate revenue must be compared like-for-like or the multiple becomes meaningless.

Pure software SaaS multiples do not work as comp here. A payments processor with thin take-rate margins is not a high-gross-margin SaaS business and should not be priced as one.

Archetype 2: The Lending Fintech

Lending fintechs underwrite credit — consumer loans, SME working capital, BNPL, salary advance. Revenue is interest income plus fees minus credit losses. Unit economics depend on net interest margin, loss rates, and cost of funds. Capital intensity is high (the loan book is the asset). The model is fundamentally a bank with a tech wrapper, not a software business.

Public comp that works as primary anchor: NuBank (LatAm neobank, currently 4.1x EV/Revenue, $73B market cap on $17B revenue with 29% revenue growth). Secondary comps include Bank Rakyat Indonesia and Tinkoff. Revolut as a private comp anchors at roughly 11.7x EV/Revenue but is not directly verifiable. The honest comp range for an African lending fintech is 2x to 5x revenue — with the upper end requiring NuBank-level customer acquisition economics, NIM discipline, and provable underwriting through a credit cycle.

Note that even NuBank, the global premium neobank by valuation, has compressed materially. Bank of America and UBS cut price targets in early 2026 citing valuation premiums in high-growth fintech contracting. The relevant directional point: the upper end of the lending fintech comp range has moved down, and African lending fintechs priced against pre-2024 NuBank multiples are anchored to comps that no longer exist.

Payments processors do not work as comp for lending fintechs (different unit economics). Pure SaaS does not work either (entirely different capital structure). The most common pricing failure in this archetype is using SaaS multiples on a lending business and ignoring the capital cost of the loan book.

Archetype 3: The Embedded Finance / B2B SaaS Play

Embedded finance and B2B SaaS plays sell software to financial institutions or to non-financial businesses adding financial services — KYC infrastructure, treasury management, payroll, embedded payments APIs, banking-as-a-service. Revenue is subscription plus usage. Margins are SaaS-typical (60-80% gross). Customer concentration is often material (top 10 customers >50% revenue is normal at this stage).

Public comps that work: enterprise SaaS at the relevant Rule of 40 cut. The honest baseline is the current public SaaS median (5.5x revenue Q1 2026) with adjustments for growth rate, retention, and customer concentration. The Rule of 40 filter is the standard discipline — companies below 40 (growth rate + EBITDA margin) trade at substantial discounts to median and should not be comped to median-tier names.

The honest comp range for African embedded finance / B2B SaaS is 3x to 7x revenue with the upper end requiring genuine Rule of 40 performance and customer diversification.

A Note on Hybrid Businesses

Many African scaleups span archetypes — Flutterwave is payments + lending + B2B SaaS; Wave is payments + remittance; Paystack-now-Stripe-Africa is payments + embedded finance. For hybrid businesses, the comp basket should be a weighted average across archetypes by revenue contribution. A 60% payments / 40% B2B SaaS business at $30M revenue should be comped at roughly (0.6 × 2.5x payments) + (0.4 × 5.5x SaaS) = 3.7x effective revenue multiple before discount stack. The weighted approach prevents the most common hybrid pricing failure, which is comping the entire business at the higher-multiple archetype.

The Discount Stack Methodology

Raw public comps are only the start. The discount stack converts public to private to African to FX-adjusted. Most decks apply discounts that are too small individually and stack them in the wrong order.

Read more

On 4 November, Optasia rang the bell at the Johannesburg Stock Exchange. The AI-driven credit-scoring fintech raised $345 million at a $1.4 billion market cap, with shares jumping 28% intraday before settling 19% above issue. Three weeks later, Cash Plus listed on the Casablanca Stock Exchange — $82.5 million raised at a $550 million valuation, first-day pop of 15%. Together, $427 million in primary capital across two African exchanges in a single month. The first homegrown tech IPOs on the continent since Jumia and Fawry in 2019.

The consensus read across the African VC market has been near-unanimous. Africa: The Big Deal called it the end of the drought. BusinessDay framed it as the IPO window opening. Trade press and investor newsletters have treated November as the inflection that puts African tech listings back on the table.

The PE Exit That Looked Like a Tech IPO

Both companies were already exit-ready before they listed. Optasia (founded 2012, rebranded from Channel VAS, Ethos Capital-backed, headquartered in Dubai, operating across 38 countries) reported $117 million in H1 2025 revenue and $54 million in H1 EBITDA. At its $1.4 billion market cap, the listing priced at roughly 6x revenue and 13x EBITDA on annualized H1 numbers. Those are payment-processor multiples. Not venture tech multiples. Optasia is a 13-year-old PE-backed multi-region fintech doing a PE exit on a public market.

Cash Plus, sitting on Mediterrania Capital Partners’ books, used the listing to deliver liquidity to existing shareholders alongside primary growth capital. Cell C also listed on the JSE in November and gets bundled into most “African tech IPO recovery” counts; Cell C is a distressed mobile operator emerging from multiple debt restructurings, structurally unrelated to anything happening in African tech.

Neither Optasia nor Cash Plus needed the public market. The public market was a preferred path for assets that already had multiple options. The rest of the African private capital stack prices off two reference points that have either compressed dramatically or never reliably existed.

Both reference points are foreign. Both are ghosts.

Share

Ghost Number One: The Comp Set

When an African Series B founder builds a pricing case for a round, the comp deck almost always includes a basket of US and EU public SaaS multiples. The implicit logic flows downward from a public anchor: public software valuations set the global ceiling, private deals trade at a discount, and African deals trade at a further discount.

The anchor moved. Then it moved again.

Public SaaS EV/Revenue multiples peaked at roughly 18.6x in late 2021. By the end of 2025, the public SaaS index sat at 6-7x. As of March 2026, the median had dropped further to approximately 5.5x — a 21% compression in a single quarter, the sharpest move since the 2022 rate-hike correction. The five-year drawdown from peak is now north of 70%.

Median private SaaS M&A multiples followed the same path with a lag, peaking around 6.4x in 2021, dropping to 2.9x in 2024 (SaaS Capital Index), and rebounding to a 3.8-5.3x range in Q1 2026 across multiple datasets. Aventis Advisors flagged something almost no one in the SaaS coverage universe used to say: EV/EBITDA is now displacing EV/Revenue as the primary metric, with the SaaS index trading at roughly 26.6x EBITDA — a profitability discipline that did not apply when revenue multiples ruled.

Now run the math on a typical African Series C deck being structured right now. Take a company at $30 million in annual revenue. Priced against a 2022-vintage public SaaS comp basket of 9x revenue, with a 25% private discount and a 15% Africa discount, the round lands at $172 million post-money. The same company priced against the current public median of 5.5x, with a defensible 30% private discount (Windsor Drake’s mid-point for the lower middle market range) and the same 15% Africa discount, lands at roughly $98 million post-money. The deck is pricing the round at nearly 1.8x the honest number — and that gap widens the further you anchor to 2021-2022 multiples.

Most decks are anchored worse than that. The comp tables built into advisor templates in 2021-2022 are still there. Nobody has refresh incentives — founders prefer the higher number and advisors get paid on round size, while existing investors prefer marks that hold without challenge. The only party with structural interest in the lower number is the new lead, and the new lead is increasingly noticing.

What happens next is straightforward arithmetic. A round priced against ghost comps lands at a post-money the next round cannot defend. The down round arrives, weighted-average anti-dilution adjustments fire, reserved-share refresh kicks in, and the cap table reconstruction work nobody planned for becomes Q3 board agenda.

For GPs, the same compression sits unseen inside portfolio marks. If a 2022 mark was based on a 2021-2022 comp set and has not been refreshed against current public multiples, the reported NAV is mechanically inflated. The DPI/NAV gap that everyone in African VC discusses as an exit problem is partly a marking problem. When realized exits price against current comps and reported marks did not move, the gap widens not because exits got worse but because marks were never rebased.

Ghost Number Two: The Listing Exit

The IPO is conventionally treated as the terminal liquidity event in a venture cap table. The conventional sequence runs from building scale to listing to distribution. The sequence assumes the listing actually delivers liquidity to insiders — that the public market provides float depth, an institutional buyer base, and post-IPO trading conditions that allow early investors to sell down meaningful positions over a reasonable horizon.

For African tech, that sequence has cleared three companies in seven years. Jumia listed on NYSE in 2019. Fawry listed in Cairo in 2019. Swvl went public via SPAC on Nasdaq in 2022 at a $1.5 billion valuation, then delisted to OTC pink sheets after losing more than 95% of its value from peak.

The TLP Advisory 2025 founder survey put concrete numbers on the founder side. 53% of Nigerian tech founders do not understand the NGX listing process. 46% prefer M&A exits. Only 21% would consider an IPO. The currency mismatch alone disqualifies most NGX paths for venture-backed companies — 77% of Nigerian founders raise in dollars and earn in naira, and dollar-denominated investors require dollar-denominated exits. NGX market cap of around $62 billion against NYSE’s $28.3 trillion tells you the float story without further analysis.

The November 2025 listings do not break the historical pattern. They confirm it. AVCA and PwC’s joint study on African PE-backed IPO exits found that listings have tracked below other markets and called for industry dialogue on how to drive activity. November 2025 is consistent with that finding rather than a break from it. Both Optasia and Cash Plus listed because they were already structured as PE-style assets with profitable unit economics, multiple potential buyers, and no acute capital pressure. They could afford the listing process because they did not need it.

Set against that, sponsor-to-sponsor secondary transactions hit a record 26% of African PE exits in 2025 (AVCA). Strategic M&A made up roughly 64% of the rest, with IPOs and other exits accounting for the remaining 10%. Sponsor-to-sponsor at record highs is a market signal: secondary participants are pricing off each other rather than off public market reference, which is rational only when the public market reference is broken or unavailable.

This is also why the LASPA structure mattered when Launch Africa pioneered it earlier this year. A structured secondary purchase mechanism that sets price through bilateral negotiation against fundamental analysis is exactly the workaround you build when the public market validation layer is not delivering. The market is constructing its own reference architecture because the imported one does not work.

What Honest Pricing Looks Like Now

The fix is mechanical.

Comp baskets need to be refreshed quarterly against live public multiples and segmented by business model rather than by geography. African fintech is not a single comp basket. A pure payments processor genuinely should comp closer to StoneCo or PagSeguro at 0.5-2x revenue. A lending fintech should comp against international neobanks. An embedded finance or B2B SaaS play should comp against US/EU SaaS minus discounts, with a Rule of 40 filter applied before any company qualifies as a comp at all.

For exit modelling, the IPO line should come out of most cap table forecasts. The realistic exit set for a Series B African scaleup is strategic M&A to a regional or global acquirer, sponsor-to-sponsor secondary to a later-stage Africa fund, or structured liquidity through instruments like LASPA. The terminal multiple in any forecast should reflect the acquirer’s listing-venue median minus a private discount minus an Africa discount minus an FX risk premium.

For LPs assessing GP marks, the single most informative question is: what comp basket was used in the most recent portfolio mark, and when was it last refreshed? A GP whose answer involves a 2022 template has marks that will reset under pressure. A GP whose answer involves quarterly refresh against current public medians has marks that will hold.

The ghost numbers do not vanish because the market notices them. They reprice through a sequence of down rounds, recapitalisations, quietly written-down marks, and the slow recognition that 2022 reference prices were never real for the African market. The reset is starting now.

The round you raise off honest comps is the round you do not have to defend at the next one. The mark you can defend at LPAC is the mark that does not require an apology in 2027.

Midweek paid piece works through the comp re-anchoring framework with specific baskets for fintech, B2B SaaS, and embedded finance, plus a portfolio mark audit checklist.

Saturday’s essay named the Founder Power Curve and ended with three tests: governance legibility, cap table cleanliness, and structural transferability. Pass all three and you are in the PE-addressable universe with negotiating leverage. Fail any one and the buyer either prices the friction into the deal or walks. The framework does the diagnostic work. What it does not do is tell you the clause-level architecture behind each test — the specific provisions, thresholds, and contract language that determine which side of the gate your company actually sits on.

The DPI wall is now in front of GPs holding 2019–2022 vintage Series A positions across the continent. Those funds need exits in the next 18 to 36 months. Portfolio companies in those funds will be pushed into sale conversations whether they are ready or not, because the GP’s survival depends on it. The companies that pass the three-test gate will transact at clean multiples. The companies that fail at least one test will transact at meaningful discount, transact through a longer-tail strategic acquirer pathway at lower price ceilings, or will not transact at all.

What follows is the clause-level architecture for each test — the specific provisions, thresholds, and contract language that determine which side of the gate your company actually sits on. The order matters. Governance legibility is the threshold gate. Cap table cleanliness determines whether the deal can close economically. Structural transferability determines whether the deal can close mechanically. Each test has its own failure modes, its own remediation costs, and its own window for being addressed.

Test 1 — Governance Legibility

The diagnostic question: can a third-party diligence team reconstruct your company’s decision-making history from documents alone, in 90 days?

If material decisions live in WhatsApp threads, founder memory, or undocumented side agreements with early investors, the answer is no. Every PE buyer’s investment committee requires documentary evidence of how the company has been run. The absence of that evidence is not something the buyer can fix after acquisition. It becomes governance reconstruction work that no PE buyer wants to fund, and the deal cannot proceed until it is done.

What compounds leverage at Series A:

Board composition. At least one independent director by Series A close. Formal nominating mechanics documented in the shareholders’ agreement. Board observer rights distinguished in writing from voting rights — a common failure mode is investors with observer seats who have accumulated informal decision-making influence that creates diligence ambiguity about who actually has authority.

Resolution discipline. All material decisions captured in written board resolutions, signed and dated. Material means capital raises, M&A activity, key hires above a defined compensation threshold (typically $150K or the local equivalent), related-party transactions of any size, and strategic pivots affecting the capital plan. Decisions below the threshold can live in management approvals; decisions above it need resolutions. The diligence test is whether a stranger reading the minutes understands why the company is where it is.

Audit infrastructure. Big Four or top-tier local equivalent audit engagement by Series A close. Clean opinion required. Qualified opinions are not automatic deal-killers at Series A, but they are at Series B and at exit. Audit firm changes mid-cycle without disclosed rationale create red flags that diligence teams will pursue — if you change auditors, document the reason contemporaneously.

Related-party transaction protocols. Written policy, pre-approval mechanics, separate signing authority for any transaction involving founder-affiliated entities, family members, or entities in which directors hold material interest. Commingled accounts between the operating entity and founder-owned vehicles are the single most common Series A governance failure, and they are almost impossible to clean up retroactively in a way that survives PE diligence.

Common failure modes and their consequences:

• Founder-controlled board with passive observer rights for investors. Cannot survive PE diligence. Remediation requires investor consent to restructure, which is difficult when the founder is negotiating from a weaker position than they held at the original round.

• Decisions documented retroactively when diligence demands it. Reconstructed governance never reads as cleanly as governance built in real time; diligence teams can tell the difference, and they price accordingly.

• Board minutes that capture decisions but not rationale. A diligence team reading minutes needs to understand not just what was decided but why — without rationale, the documentation fails the reconstruction test.

Remediation cost by stage:

• At Series A: $15,000–$40,000 in legal and audit scoping work. Most Series A governance issues are cleanable if addressed at the round itself.

• At Series B: $80,000–$150,000 and four to six months of structured remediation work. Requires investor consent for material changes.

• At exit: deal economics. Governance reconstruction during exclusivity gives the buyer leverage to extract price concessions or walk away entirely.

Read more

Building a camel is table stakes. But that alone does not make you sellable.

The same operating discipline that produces a sustainable African company can produce a company no PE buyer can transact on. Operating excellence is one problem. Transferability is another. Both are required for a founder outcome, and they sit in sequence — the second only matters once the first is solved. Most founders who have correctly internalised the camel model will still lose their exit at the transferability layer. They lose it because the legal scaffolding around the business makes the sale impossible, not because the business itself failed.

Twelve weeks ago, in The Camel Imperative, I argued that Africa’s next winners won’t be unicorns. They’ll be companies built on five principles: cash conversion before growth, positive unit economics at the transaction level, FX-hedged revenue architecture, local cost discipline, and exit optionality designed for the regional banks, telcos, and corporates buying at $30–80M. That argument has held. The buyer class that essay named is now the dominant exit pathway. AVCA’s 2025 data confirmed sponsor-to-sponsor transactions reached a record 26% of African PE volume — which means PE firms are increasingly buying from other PE firms, not from founders. The reason matters, and it gets to what the Camel essay didn’t address.

The Window Is Closing Faster Than Founders Realise

The DPI wall is now in front of GPs holding 2019–2022 vintage Series A positions across the continent. Those funds need exits in the next 18 to 36 months or they face fundraising collapse on their next vehicle. That timing pressure transmits directly to portfolio companies. Founders in those portfolios will be pushed toward sale conversations whether they’re ready or not, because the GP’s survival depends on it.

The push is happening into a buyer class with diligence requirements that strategic acquirers do not impose. Strategic buyers absorb operational mess because they’re buying capability, market position, or talent. They can integrate, restructure, and write down what doesn’t fit. PE buyers operate under a different constraint. They’re buying a financial asset that needs to operate, grow, and re-sell within a defined hold period. Every defect they inherit becomes their problem to solve before their own exit, which means they price defects in or walk away.

The decisions that determine whether you carry defects into a sale process are made at Series A. By Series B those decisions are largely locked. By the time exit-readiness becomes urgent, they cost equity, time, and sometimes the deal itself to remediate. The founder who learns this during diligence learns it too late.

The Founder Power Curve

Four inflection points map where founder leverage compounds or collapses. At each point, a specific decision category determines whether the company becomes more sellable or less.

Inflection Point 1 — First Close. The decisions made here look small and are not. Cap table architecture from day one matters more than founders typically appreciate: a single share class for founders, vesting schedules with cliff and acceleration provisions, founder employment agreements with IP assignment that survives termination. The opposite trajectory is recognisable in retrospect. Stacked convertible instruments — SAFEs and bridge notes layered over multiple closings, post-priced-round side-letters to specific investors, conversion mechanics never cleanly executed across jurisdictions — surface during diligence as a forensic exercise nobody wants to fund. Informal equity grants made to early collaborators without documentation. Intellectual property held in personal name, or worse, jointly with a co-founder who later left. PE diligence will find every uncleaned instrument. Each one becomes either a price discount or a deal-killer, depending on how aggressive the buyer is and how exposed the seller is to time pressure. Most First Close defects can be remediated cleanly at Series A if caught early. By Series B, remediation costs equity, time, and sometimes the deal.

Inflection Point 2 — Series A Governance. What compounds leverage here is building a real board. Independent directors where appropriate. Formal board resolutions for material decisions. Written delegation of authority. Audited financials from year one. Related-party transaction protocols documented and enforced. What collapses leverage is the founder-controlled board with passive observer rights for investors, decisions documented in WhatsApp threads, commingled accounts with founder-owned entities, and no audit trail for material changes in capitalisation or strategy. PE buyers cannot transact on a company whose decision-making history is undocumented. They cannot defend their investment committee approval without that documentation. Post-LOI becomes governance reconstruction work, and reconstructed governance never reads as cleanly as governance built in real time.

Inflection Point 3 — Series B Cap Table. Protective provisions calibrated to investor stage, drag-along rights with reasonable thresholds, ROFR and co-sale terms structured to permit secondary liquidity without blocking primary exits, anti-dilution that doesn’t trigger a death spiral in a down round. The collapse trajectory is more common than founders admit. Stacked liquidation preferences that mean founders earn nothing below a high exit threshold. Veto rights distributed across too many investors creating exit blocking. Protective provisions written in early-stage shorthand that don’t scale to growth-stage deals. The cap table determines whether the deal closes at all. A buyer can love the business and walk away when waterfall mechanics make the founder mathematically indifferent to closing — because a founder with no economic upside has no incentive to navigate the deal through to completion.

Inflection Point 4 — Exit-Readiness Threshold. This is where the prior three points either pay off or fail to. The compound-leverage decisions: 18 to 24 months of audited financials available on demand, customer contracts with clean assignment provisions, employment agreements with non-competes and IP assignments that survive termination, regulatory licences held in the operating entity rather than scattered across subsidiaries, a data room maintained continuously rather than built reactively. The collapse decisions are the mirror image. Financials reconstructed from bank statements during diligence. Customer contracts that terminate on change of control. Key employees on handshake deals with no formal employment terms. Licences held in the wrong entity or in personal capacity. Data room thrown together in 60 days under exclusivity. Every gap surfaced during diligence either reduces valuation or kills the deal. In distressed exit conditions — which is what most African Series B exits will be in the next 24 months — both outcomes get worse, because the seller has lost the leverage to push back on buyer demands.

The curve compounds. Each inflection point’s decisions either remediate prior gaps or stack on top of them. Founders who get Point 1 right have a meaningfully easier path through Points 2 and 3. Founders who get Point 1 wrong spend Series A and B capital on remediation work that should have been part of the original build — and that capital is not coming back, regardless of how the exit ultimately resolves.

Share

What the Numbers Look Like

Three pieces of evidence ground the framework.

The first is in the AVCA 2025 sponsor-to-sponsor data. PE firms buying from other PE firms now represent 26% of African PE deal volume, a record. The implication that doesn’t get discussed enough: PE buyers are increasingly buying from other PE sellers because those companies have been cleaned up by the selling fund’s portfolio operations team. Founder-led companies arriving at the PE buyer class directly face a meaningfully higher diligence standard than secondary PE assets, precisely because they haven’t been pre-cleaned by an institutional seller. That disadvantage isn’t visible to founders until they’re inside diligence, at which point it’s too late to address without conceding price.

The second is what the transferability layer is worth in valuation terms. Consider an illustrative scenario based on patterns observable in recent African fintech transactions. Two Series B companies, both at roughly $8M ARR, both growing 60% year-over-year, both EBITDA-positive. Company A has clean governance, audited financials going back three years, single jurisdiction of incorporation, IP cleanly held in the operating entity, and a standard four-class cap table. Company B has the same operating metrics, but commingled accounts that were partially separated last year, IP still held partly in the founder’s personal name, three jurisdictions across operating subsidiaries, and four classes of preferred with stacked liquidation preferences from successive down-rounds. Same business at the operating level. Company A transacts in the 4–5x revenue range commonly observed in clean African fintech sales. Company B transacts in the 2–3x range — if it transacts at all. The transferability layer is worth 40–50% of enterprise value at exit, and that spread shows up nowhere in the operating financials. It’s purely a function of how the company is wrapped.

The third piece of evidence is where most African PE deals actually die. Not in valuation negotiations or commercial terms, but in contract review. Customer contracts, employment agreements, and supplier arrangements containing change-of-control provisions written by local counsel without exit-readiness in mind. Customer contracts carry the most weight here, because they underpin the valuation itself — particularly key enterprise contracts that account for a meaningful share of total revenues. The buyer cannot assume the contracts as drafted. The seller renegotiates them under time pressure during exclusivity, often with counterparties who realise their leverage and extract concessions. I have seen this kill multiple transactions over the last few years, and in every case the underlying drafting failure was preventable at the point of original negotiation, sometimes years earlier. This is the practitioner judgment layer in action — contract drafting that anticipates exit at the point of negotiation, rather than discovering during diligence that the wording doesn’t support a sale.

The Three-Test Gate

Three tests determine whether you are in the PE-addressable universe at all.

The first is governance legibility. Can a third-party diligence team reconstruct your decision-making history from documents alone, in 90 days? If material decisions live in WhatsApp threads, founder memory, or undocumented agreements, the answer is no. The deal cannot proceed without governance reconstruction work that no PE buyer wants to fund.

The second is cap table cleanliness. Can a buyer model the waterfall in 30 minutes and confirm that founder, management, and key investors are all economically aligned to close at a target valuation? If the waterfall requires a specialist to model, or if the alignment math reveals that the founder earns nothing meaningful below a high threshold, the deal stalls before terms are even negotiated.

The third is structural transferability. Can the operating entity be sold without consent friction from contracts, regulators, jurisdictions, or minority holders? If the answer requires a renegotiation campaign with third parties before close, the buyer either prices in the friction or walks.

Companies that pass all three are the addressable universe for the emerging PE buyer class. Companies that fail any one are not in that universe. They may still reach a strategic acquirer exit, which has a different and longer-tail buyer pool. They may exit through a secondary sale (albeit likely at meaningful discount). Or they may not exit at all. The point isn’t that any of these is automatically wrong — a strategic exit at the right multiple is a real outcome. A secondary in which a founder sees liquidity is still liquidity. The point is that founders should know which exit pathway they are actually building toward (if any), rather than discovering it when the GP starts pushing for a sale process they aren’t structured to support.

This is not a product problem or a market problem. It’s simply a preparation problem. The decisions that determine whether you pass the three-test gate are inexpensive to get right at Series A — the marginal cost of doing the work properly versus doing it cheaply is small. By Series B you’re paying to remediate on top of new work. By exit-readiness you’re paying with deal economics, or the deal itself. Operating discipline keeps the company alive long enough for an exit to be possible. Without the transferability layer, that exit doesn’t happen.

What Comes Next

The three-test gate is the qualifying filter. The clause-level architecture that determines whether you pass each test is where most founders need practitioner judgment they don’t have in-house — because local counsel drafts for the round in front of you, not for the exit two rounds away. Midweek piece, The PE-Readiness Diagnostic, walks through the specific governance, cap table, and contract criteria PE acquirers apply to African founder-led companies, and where the most common failure modes sit. Paid subscribers get the full diagnostic. Free subscribers get the framework summary.

The window for getting these decisions right is the 12 to 24 months before your GP starts the conversation you didn’t know was coming. Most founders find out the hard way. The ones who don’t, build for it deliberately.

The exit conversation most African founders and GPs are having starts with the wrong question. “Are we ready to exit?” is not the question that determines the outcome. The question that matters is: ready for which buyer, on whose terms, at what price?

Trade sales are the dominant exit route for African private capital — and two buyer classes now define that market. Each requires a different company. Most cap tables are built for neither — or built for the one that pays less.

Work through both tracks below against your current position. Be precise. Vague answers produce vague diagnoses — and vague diagnoses are how founders end up in a thirty-day exclusivity window discovering problems that should have been fixed eighteen months earlier.

Track One — PE Readiness

Private equity acquires to sell again. That single fact determines everything about what they need from you at due diligence. They are not buying your story. They are underwriting a financial model that has to work twice — once when they buy, once when they sell. Every item below is something a PE fund will verify before issuing a term sheet. If you can’t verify it yourself first, they will find it — and price the uncertainty into the offer.

Financials

Are your last three years of accounts audited by a recognised firm — not a local bookkeeper, an actual audit? Are those accounts segmented by revenue line, geography, and product, or presented as a single consolidated figure? Can you produce a trailing twelve-month P&L within five working days of being asked?

I’ve seen data rooms where management accounts were presented as a substitute for audited financials — sometimes with genuine confidence that the distinction wouldn’t matter. It always matters. PE funds model entry on audited segment financials. The due diligence process will either stall at that point or produce a price adjustment that reflects the uncertainty introduced. In the transactions I’ve seen move cleanly, the financials were audit-ready before the first conversation, not after the LOI.

Read more