Capital stopped funding standalone products and started buying them at 2-4x revenue, in the same half-year that the continent's largest distribution network began building the platform those products will be offered space inside. The middle ground is what disappears.
A weekly intelligence brief from Base X Studio. Pan-African tech, global brand strategy.
African startups raised $1.44B in H1 2026, roughly level with last year, but across 146 disclosed deals instead of 252. M&A went the other way. Read the three numbers together and the story is selection, not scarcity: the same capital is going to fewer companies, and the ones it is not going to are being bought rather than wound down. Infrastructure and B2B multiples have compressed to 2-4x revenue, down from double digits, which prices the exit before anyone has decided to take one.
At 2-4x revenue you cannot sell your way out of a bad year, so stop treating acquisition as the fallback plan and price it as a live option now. Pick one of three stances this quarter and say it out loud internally: you are buying, you are building to be bought, or you are staying independent and funding that decision. The businesses that get taken out cheapest are the ones that never chose and got chosen for.
Voice and mobile carriage, the business the brand was built to describe, and the part of the company that is now shrinking.
Data as a share of group operating revenue. Fibre subscribers up 15.5%, BCX cybersecurity revenue up 36.6%, AI platforms handling 54.6% of prepaid service revenue.
Run the mix on your own P&L this week and put a number next to each revenue line. If more than half of it comes from something your positioning statement does not mention, your positioning is describing a business you used to run. Fix the claim before the next raise, because an investor who does the arithmetic will find the gap first and read it as a management problem.
MTN Group Fintech has brought in Ant International to rebuild MoMo as a super-app with a third-party mini-app layer. The mini-app layer is an invitation and should be read as one. The largest distribution network on the continent is building the container that single-product fintechs will be offered space inside, and its CEO has named the logic in his own words. Distribution, not product quality, is what is being consolidated.
If you sell financial or commercial services in Nigeria, the mini-app layer arrives this quarter and you have to decide before it does. Going in buys you distribution you cannot build and costs you the customer relationship, the data, and the ability to be chosen rather than found. Staying out means your acquisition cost has to beat a super-app's, which means your reason to exist has to be something MoMo cannot host.
The word doing the work is restructuring. This is not growth capital arriving at a healthy asset, it is development finance stabilising the 110,000km network across 25 countries that both the super-app play and the telco pivots depend on. Investors favouring asset-backed businesses over pure software is the same capital-scarcity logic as the M&A wave, seen from underneath.
Before you sign a multi-year connectivity or colocation contract this quarter, ask your provider how their balance sheet is financed and who holds the debt. If your unit economics assume bandwidth keeps getting cheaper, model a flat case as well. The network you depend on is being kept upright by development finance rather than by its own cash flows.
Three of these five are super-app or platform owners buying a capability rather than building it. The pattern is not distress, it is assembly, and the parts being assembled are companies that were, until recently, categories of their own.
Bina names an asymmetry running through the current economy: institutions, platforms and auditors operate masked while individuals are fully visible, and the gap between the two is where the money is now made. Her examples run from prediction markets profiting off other people's disasters to models trained on uncompensated creative work. The consequence that should worry any brand with an audience is the retreat into walled gardens, which means the channel you reach people on is closing from the inside. Her prescription, reciprocal transparency and finite bounded exchanges instead of infinite-engagement loops, is a product specification rather than a comms brief.
Your data practices are now positioning, so read your own consent flow this month as if you were the customer being asked. If your growth depends on an engagement loop with no natural end, you are on the wrong side of where your audience is moving, and the fix is to design a finish line into the exchange. Publish what you collect in plain language before somebody else describes it for you.
Six years and a couple of hundred technology companies later, Dunford has restructured Obviously Awesome from five components and ten steps to five and five, expanded the differentiated value step because it is where teams get stuck, and added new sections for multi-product companies. That last one is the admission worth reading. The standard positioning framework in the industry was built for a company selling one thing, and most businesses at the stage that need positioning are already selling three.
If you sell more than one thing, you are probably stretching one position across all of them and calling the thinness flexibility. Pick the offer that carries the company position this quarter and let the others position relative to it rather than competing for the same sentence. The tell that you have not done this is a homepage headline abstract enough to cover everything you sell, which means it describes none of it.
Bina argues this is not the AI era but the contract era. AI has made the symbols of care and effort free, so the implicit contracts that ran on those symbols are being renegotiated in the open. The mechanism matters more than the observation. Value is not moving to human-made things out of nostalgia, it is moving to things carrying irreversible cost, time, reputation, presence, risk, because cost is the only signal that cannot be generated.
Go and find the one place in your business where you replaced a costly signal with a generated one, and put the cost back this quarter. The AI-written performance review, the automated founder note, the onboarding call that became a sequence: each saved money and spent trust, and your customer noticed before you did. Keep the automation where the work is genuinely mechanical and pay the price where the point of the thing was that it cost you something.
The Dropbox and Slack era. A person signs up, gets value, and shares the product itself with someone else.
The current stage. Virality runs on people broadcasting what the AI made for them, not on sharing the tool that made it.
Growth loops that no longer route through a human interface at all. Bush's claim is that product-led growth is now table stakes, not an edge.
Bush's model rewards frictionless, costless, infinitely shareable output. Bina argues that exact output is discounted on arrival. Both describe 2026 accurately, so the tension is real and worth holding open.
Spend this quarter on the first five minutes of your product rather than the messaging above it, and make sure what a user shares from it carries your claim rather than just your logo. The trap sitting inside PLG 2.0 is generating output so cheap to make that sharing it signals nothing about you at all.
The 63 M&A deals are the bigger statistic but on their own they are an aggregate, and an aggregate does not tell you what businesses are being pushed toward. MTN and Ant International name the destination in plain words: platform consolidation over product sprawl, a mini-app layer, Nigeria first, 60 million customers already there. Put the two together and it is a pincer, not a trend. On the supply side, capital has stopped funding standalone products and started buying them at 2-4x revenue. On the demand side, the largest distribution network in Africa is building the place those products will live. For a scaling business the middle ground disappears in that squeeze. So decide this quarter which side of it you are on. If you are staying independent, name the thing you do that a super-app cannot host and put next year's budget behind that specifically. If you cannot name it inside an afternoon, you are not independent, you are unacquired, and those two words describe very different negotiating positions when somebody eventually calls.
When the continent's largest distribution network builds a platform designed to absorb single-product fintechs, staying independent stops being a neutral choice and becomes a bet with a closing window. The bet only pays if you can name something you carry that the platform cannot host, and Bina's cost-and-consent argument is the test for what that something has to be: a relationship expensive enough to have built that it could not be regenerated inside somebody else's mini-app layer.