Microsoft has disclosed for the first time just how much of its artificial intelligence revenue comes from a single customer: OpenAI. According to a filing reported by Bloomberg and The Business Times, the software giant recorded US$24.1 billion in sales from OpenAI during its fiscal year ending in June 2026 — a figure that suggests the ChatGPT maker likely accounted for around 70 percent of Microsoft's actual AI sales.
The rare disclosure, which surfaced as Microsoft reported its fiscal fourth-quarter earnings, gives investors and analysts their clearest view yet of a dependency that has long been suspected. For daily updates on the deals shaping the industry, AI Buzz Wire tracks the latest across cloud, models and infrastructure.
The Numbers Behind the Dependency
Microsoft chief executive Satya Nadella told investors earlier this year that the company was on pace at the end of the March quarter to record roughly US$37 billion in AI revenue over a full year. The company did not update that total when it reported fiscal Q4 earnings, but Bloomberg's analysis assumed the AI business continued growing at the rapid 123 percent rate Microsoft reported for March.
At that pace, Microsoft's AI business would have tallied about US$34 billion in the fiscal year ending in June — which can be directly compared with the new disclosure that OpenAI alone provided US$24.1 billion of revenue in that same period. That puts OpenAI's share at well over half, and plausibly near 70 percent, of Microsoft's total AI sales.
Under the agreement between the two companies, OpenAI pays Microsoft for computing power, the costs associated with building AI models, and a share of its own revenue. A Microsoft spokesperson confirmed the disclosed figure includes all sales and revenue share from OpenAI.
A Small Slice of a Much Bigger Pie
Context matters. Against Microsoft's total revenue, OpenAI's contribution looks far smaller — less than 10 percent. The company added about US$51 billion in commercial bookings in its most recent quarter, driven largely by customers other than AI startups. Even so, OpenAI made up most of the company's annual bookings growth, underscoring how much of Microsoft's forward momentum in AI still rides on a single partner.
Until this filing, Microsoft had never clearly released its full revenue from OpenAI. The company had previously disclosed AI run-rate milestones only twice: once for the quarter ending in December 2024, when it said the unit was on pace to generate more than US$13 billion over a year, and again in the quarter ending in March, when it pointed to more than US$37 billion.
Why Reveal It Now?
The timing of the disclosure has drawn attention from accounting researchers. Olga Usvyatsky, founder of data analytics firm Nonlinear Analytics, wrote in a note that the release may be tied to OpenAI's plans for an initial public offering. As OpenAI prepares to go public — a move Microsoft's massive 2025 investment helped set the stage for — greater financial transparency about the relationship becomes harder to avoid.
Microsoft has, in parallel, tried to reduce its reliance on OpenAI. The company has backed rival lab Anthropic, invested in building its own in-house models, and courted a broad base of enterprise AI customers through Azure. But the new numbers make clear that diversification has not yet meaningfully loosened the OpenAI link.
A Partnership Reshaped, but Not Severed
The financial entanglement persists even as the two companies have restructured their relationship. Last year, Microsoft and OpenAI renegotiated the terms of their landmark alliance, with Microsoft retaining rights to OpenAI's revenue through the end of the decade and a preferential position on compute and storage while giving OpenAI more freedom to build its own infrastructure and strike cloud deals elsewhere. Amazon's subsequent multibillion-dollar investment in OpenAI — and its hosting agreement — added a second hyperscaler to the picture.
Yet the new disclosure shows that, at least for now, the dollars still flow overwhelmingly through Redmond. OpenAI's enormous compute needs — training frontier models requires clusters of tens of thousands of GPUs running for months — translate directly into Azure consumption, and Microsoft books that spending as AI revenue. The $24.1 billion figure is, in effect, a measure of how much OpenAI's model ambitions cost to run on Microsoft's cloud.
What It Means for the AI Ecosystem
The disclosure lands at a moment of intense scrutiny over the economics of frontier AI. Amazon completed a roughly US$8 billion investment in OpenAI earlier this year, and the company's revenue has reportedly surpassed giants like Starbucks and McDonalds on a Fortune 100 basis. Microsoft's filing adds a new data point: the cloud infrastructure underpinning the boom is still overwhelmingly concentrated in a handful of hyperscaler-to-lab relationships.
For Microsoft investors, the figure is double-edged. A US$24.1 billion annual contribution validates the strategic bet on OpenAI, but it also quantifies the concentration risk. If OpenAI's spending on Azure compute were to slow — whether because of an IPO, a renegotiated deal, or a shift toward its own infrastructure — the impact on Microsoft's AI growth narrative would be immediate and visible.
Stay Ahead of AI
The business of AI is moving billions every quarter. Keep up with the deals, earnings and power shifts at AI Buzz Wire, your source for source-checked industry reporting.
Read more AI news →
