Microsoft’s Q1 Results Reveal $10bn From OpenAI — and a $7.2bn Receivable

Last updated: 30 October 2026. Figures below are drawn from Microsoft’s fiscal Q1 2027 earnings release and 10-Q (quarter ended 30 September 2026, reported 28 October 2026). Several figures are sourced via reporting that quotes those filings directly, as noted inline, because the filings themselves could not be fetched directly in the course of writing this piece — that limitation is set out in full in the “What we still don’t know” section below.

The 60-second version

• Microsoft’s fiscal Q1 2027 results, reported on 28 October 2026, disclosed Azure revenue in dollar terms for the first time: $33.2bn, up 48% year-on-year — a figure the company had historically given only as a growth percentage.
• The same filings show Microsoft recorded roughly $10bn in revenue from commercial arrangements with OpenAI in the quarter alone — about 11% of its $92.3bn total revenue, and more than its Xbox and Windows OEM businesses combined.
• The balance OpenAI has agreed to pay but hasn’t yet paid grew to roughly $7.2bn, up from $6bn three months earlier.
• Microsoft’s equity stake in OpenAI, now around 24%, produced a comparatively small gain this quarter (roughly $0.2bn after tax) against a $3.1bn loss in the same period last year.
• Days earlier, Anthropic’s own S-1 disclosed an $80.7bn paper loss alongside a $2 trillion valuation ask, and a widely circulated Apollo note argued AI has delivered “zero payoff” outside the tech sector itself. Microsoft’s numbers are the first hard data point either side of that argument can actually point to.
• The same week, Alphabet reported a near-$9bn mark-to-market loss on its own AI stakes — a reminder that “AI investment income” can run in either direction.

Key figures from the quarter

What Microsoft actually disclosed

Microsoft has reported Azure growth as a percentage for years, without a dollar figure attached. That changed this quarter: Azure revenue came in at $33.2bn, up 48% year-on-year and ahead of the roughly $32.3bn analysts had modelled, according to TechCrunch’s reporting on the release. Azure now sits inside a reorganised segment Microsoft calls “Agents and Infra,” which also folds in Microsoft 365’s cloud business and brought in $76.8bn for the quarter, up 25%.

Total revenue was $92.3bn, a 19% rise and the tenth consecutive record quarter, with net income of $36.6bn, up 32% on a GAAP basis, TechCrunch reported from the same filings. Remaining performance obligation — the backlog of contracted future revenue — rose 73% to $677bn, even as new bookings fell 12% against a tough comparison that CFO Amy Hood attributed to unusually large Azure commitments signed a year earlier.

The OpenAI dependency, sized properly

The more striking number sits inside those totals: roughly $10bn in revenue this quarter came from commercial arrangements with OpenAI, including revenue-sharing payments, which TechCrunch calculated at nearly 11% of total revenue — more than Microsoft’s Xbox and Windows OEM businesses generated between them. For the full 2026 fiscal year, OpenAI-linked revenue came to $24.1bn, a figure first disclosed in Microsoft’s FY2026 10-K, which this quarter’s run rate would roughly double if sustained.

What the $10bn is made of isn’t broken out. Microsoft doesn’t separate OpenAI’s own Azure consumption — compute it buys like any other customer — from the contractual revenue-sharing payment under the companies’ commercial agreement, which runs through 2030 subject to an undisclosed cap. That distinction matters: one is OpenAI paying for a service, the other is closer to a cut Microsoft takes of OpenAI’s business. Collapsing both into a single “OpenAI revenue” line, as most coverage of the quarter has done, flatters the headline without answering which kind of revenue is actually growing. It also sits oddly alongside Microsoft’s own push for distance from its partner: a 2.8-trillion-parameter in-house model it announced earlier this month, explicitly framed as a bet on AI independence from OpenAI — a commercial relationship and a competitive one, running in parallel.

The receivable, and why it’s not what it looks like

The number that drew the most second-guessing was the $7.2bn OpenAI owed Microsoft at quarter’s end, up from $6bn three months earlier — a jump some initial commentary treated as a $1.2bn shortfall between booked revenue and cash collected. That conflates two different things: $10bn is a flow measured over the whole quarter, while $7.2bn is a balance at a single moment, and an analysis of the disclosure made the point directly — the two aren’t meant to net against each other, and subtracting one from the other produces a misleading “gap” the accounting doesn’t support. A growing receivable on a growing revenue relationship is, alone, unremarkable; the real signal would be the receivable growing faster than the revenue behind it, which these two data points don’t establish either way.

What most coverage is getting wrong

Two framing errors recurred this week. The first is the receivable-subtraction error above. The second is treating the 11% figure as Microsoft’s own framing: it was reportedly worked out on the earnings call by a sell-side analyst, not volunteered by the company, and Satya Nadella is reported to have declined to adopt the ratio when it was put to him. A company confirming a number under questioning is a different kind of disclosure than one it chooses to lead with, and the two deserve different weight.

The third, broader error is treating this quarter’s numbers as settling the “AI bubble” argument either way. They don’t. Real cash changed hands — OpenAI’s $10bn was presumably real revenue for Microsoft — but Microsoft is also an investor in OpenAI, so some of the money OpenAI used to pay its Azure bill traces back, at least partly, to capital Microsoft itself put in. How much of the $10bn is genuine external demand versus Microsoft subsidising its own top line through its OpenAI stake isn’t something these disclosures resolve, and no one covering the release this week has shown otherwise.

Alphabet’s mirror image, the same week

The contrast with Alphabet, which reported in the same window, is instructive. Alphabet holds AI-lab stakes too — roughly 14% of Anthropic and a position in SpaceX — which supplied roughly $6 of Alphabet’s $9.11 per-share profit back in Q2 2026. This quarter the mechanism ran in reverse: a loss of close to $9bn on those holdings, which CFO Anat Ashkenazi attributed mainly to a roughly 12% drop in SpaceX’s valuation, cut Alphabet’s EPS to $2.12 from around $2.67, according to Yahoo Finance’s analysis of both companies’ results. Microsoft’s stock rose roughly 2% in after-hours trading; Alphabet’s fell nearly 10% after hours and then closed down roughly 13% the next trading session — its steepest one-day drop since going public in 2004, erasing close to $500bn in market value. Both companies sell compute to AI labs while holding equity in them — Google Cloud took in roughly $31bn this quarter, against the backdrop of Anthropic’s own reported $200bn, five-year spending commitment to Google’s cloud and chip capacity, disclosed in May 2026 — the exact structure that makes any single quarter’s investment line as much a bet on a handful of private valuations as a measure of the underlying business.

Practical takeaways for anyone building on this infrastructure

For teams that build, publish or run software on this infrastructure, three things from this quarter are worth acting on. First, concentration risk is now quantifiable: if roughly 11% of Microsoft’s revenue and an outsized share of Azure’s growth is tied to one customer, any renegotiation, outage, or dispute between Microsoft and OpenAI is a material event for Azure’s roadmap, not a footnote — teams making multi-year Azure commitments should ask directly how much of a region’s capacity planning assumes continued OpenAI demand. Second, read backlog alongside bookings, not instead of them: Microsoft’s remaining performance obligation grew 73% while new bookings fell 12%, so the growth story and the momentum story currently point in different directions. Third, when any AI vendor discloses revenue from another AI company, ask the composition question this piece asked about Microsoft and OpenAI — genuine external demand, or revenue-sharing and investment-linked payments — because OpenAI’s own compute bill tracking its last funding round almost exactly is the same pattern from the other side of the relationship.

What we still don’t know

• Neither Microsoft’s own investor-relations release nor its SEC filing could be fetched directly while researching this piece; the figures above rely on reporting (chiefly TechCrunch and Yahoo Finance) that quotes those documents, and on an independent analysis piece that cites the same filing. Readers who want the primary text should consult Microsoft’s own earnings release and its 10-Q filed with the SEC directly.
• The exact split of the ~$10bn OpenAI figure between Azure consumption and contractual revenue-sharing has not been disclosed, and Microsoft did not answer a direct question on the composition when it was put to the company by reporters this week.
• Why OpenAI’s receivable balance is growing — whether it reflects ordinary payment-cycle lag on a fast-growing contract, or something closer to strain — is not established either way by the two data points available so far.
• The level of the cap on OpenAI’s revenue-sharing obligation to Microsoft, which runs through 2030, has not been made public.
• Microsoft’s own explanation for choosing to disclose Azure’s dollar revenue for the first time this quarter was not given, and no outlet reviewed for this piece reported a stated reason.

FAQ

How much of Microsoft’s quarterly revenue comes from OpenAI?
Roughly $10bn of the $92.3bn Microsoft reported for the quarter ended 30 September 2026 — about 11% — came from commercial arrangements with OpenAI, including revenue-sharing payments.

Is the $10bn pure profit for Microsoft?
No. It’s revenue, not profit, and a portion of it flows from a contract structure in which Microsoft is also an equity investor in OpenAI — meaning some of the money is, in effect, moving between two entities that are financially linked to each other rather than representing a wholly independent transaction.

Does the $7.2bn OpenAI owes Microsoft mean OpenAI is behind on payments?
Not necessarily. It’s an accounts-receivable balance on a fast-growing commercial relationship, and a rising balance is an expected feature of a rising revenue line unless it’s growing meaningfully faster than that revenue — which hasn’t been clearly shown either way from the figures disclosed so far.

Does this settle whether AI spending is paying off?
No single quarter does. Microsoft’s numbers show real revenue changing hands; Alphabet’s same-week loss on its own AI stakes shows the same investment structure can just as easily subtract from earnings. Both are compatible with an AI build-out that is real, large, and not yet cleanly profitable to assess from outside.

Sources

• Microsoft FY2027 Q1 earnings release (primary; not independently fetched for this piece — see “What we still don’t know”)
• Microsoft Form 10-Q, quarter ended 30 September 2026 (SEC EDGAR) (primary; not independently fetched for this piece)
• TechCrunch: “Microsoft just shared Azure’s actual revenue in dollars: $33.2 billion”
• TechCrunch: “Microsoft recorded $10 billion in revenue from OpenAI”
• TechCrunch: “Microsoft hits 10 quarters of record-breaking revenue in a row”
• Yahoo Finance: “Alphabet and Microsoft are on both sides of the AI boom’s biggest bills”
• Techzine: “Microsoft earned 24 billion from OpenAI, but it’s unclear exactly how”
• Pebblous: analysis of the Microsoft–OpenAI revenue and receivable figures
• Computing: “Anthropic commits $200bn to Google Cloud,” citing The Information

Please follow and like us: