Microsoft Booked $10 Billion From OpenAI in One Quarter, About 11% of Its Revenue. Read It as a Concentration Risk Lesson for Your Own Customer List.
Microsoft reported $10 billion in revenue from commercial arrangements with OpenAI in the quarter ended September 30, according to TechCrunch's reading of its regulatory filing. Total revenue for the quarter was $92.3 billion, so OpenAI-related revenue is about 11% of the whole company. That share is TechCrunch's math, not a figure Microsoft states.
I care less about Microsoft's problem than about what the number does to your thinking. Most solo operators have a customer or a platform that is 11% of their business or far more, and nobody is writing a news story about it.
What the filing shows
The $10 billion includes revenue-sharing payments, but TechCrunch said Microsoft's filing does not split it between revenue share and Azure services, and it had asked for that detail. For fiscal 2026 as a whole, OpenAI-related revenue was $24.1 billion. TechCrunch extrapolates about $40 billion a year at the current pace, which is a simple four-times-the-quarter estimate, not a forecast.
Receivables from OpenAI, meaning money OpenAI owes Microsoft for past transactions, rose to $7.2 billion at the end of September from $6 billion at the end of June. Microsoft's ownership stake in OpenAI slipped from about 25% in June to 24% in September, and Microsoft did not say why.
A big customer that owes you more each quarter is a customer whose health is your health.
Why this matters at one-person scale
Microsoft can absorb a hit to an 11% revenue line. You probably can't absorb a hit to a 40% one. The pattern is the same, though: revenue concentrated in one counterparty, growing receivables, and a relationship where the other side also holds equity or leverage over you.
Solo operators drift into this without noticing. One agency client pays for half your hours. One marketplace sends most of your signups. One API provider sits under your whole product. Each looks like a good decision when it happens, and each becomes a single point of failure.
The rule I'd use
Count concentration three ways, once a quarter. Revenue share: what percent of income comes from your top customer. Distribution share: what percent of new users or leads come from your top channel. Dependency share: what percent of your product's function runs through your top vendor. Anything over 30% on any of the three gets a written plan, even a short one, for what you do if it vanishes.
The plan doesn't have to be elaborate. For a customer, it is a pipeline of two or three alternatives you talk to every month. For a channel, it is an owned asset like an email list. For a vendor, it is a second provider you have actually tested, which is the swap-layer argument I keep making about model APIs.
Also watch receivables. If your top client's invoices take longer to land each month, that is the early warning, and it shows up before anything else does.
The honest counter-take
The opposite view deserves a real hearing. Concentration is how small businesses get started. A single anchor client gives you cash flow, focus and a reference, and spreading yourself thin across ten small clients can mean you do nothing well. Microsoft's own bet on OpenAI looks smart so far, and the filing shows it paying off in revenue.
And a filing shows one side of a relationship. I do not know the contract terms, and a revenue line is not the same as risk. So take this as a prompt, not a verdict on either company. Know your numbers, pick your concentration on purpose, and write down the exit before you need it.
Author
Lukas
@lukcombinator