Microsoft's 27% OpenAI stake is about to make its earnings harder to read. Here's why operating income and free cash flow matter more than net income once OpenAI goes public.
Microsoft owns roughly 27% of OpenAI. That fact gets repeated constantly, usually as a kind of shorthand for “Microsoft is deeply exposed to the AI boom.” It is true, but it understates the situation. The relationship between these two companies is not a simple equity stake sitting quietly on a balance sheet. It is a tangle of investment, infrastructure contracts, and revenue-sharing arrangements — and once OpenAI goes public, that tangle is going to start showing up in Microsoft’s earnings in ways that will confuse a lot of investors who aren’t paying close attention.
This is not a doom piece about Microsoft. The company is performing well by almost every operational measure. But the OpenAI relationship is about to introduce a level of accounting noise into Microsoft’s results that most shareholders have never had to think about — and if you don’t know what to ignore, you risk reacting to numbers that don’t actually reflect how the business is doing.
Why OpenAI Has to Go Public — There Isn’t Really a Choice
Start with the scale of the numbers, because they explain everything that follows. OpenAI is generating roughly $25 billion in annualized revenue as of early 2026 — a genuinely enormous figure for a company that didn’t exist as a commercial entity a decade ago. But it is burning through cash at a similar or greater rate: estimates put 2026 cash burn at around $27 billion, rising toward $63 billion in 2027 as infrastructure spending accelerates.
Against that backdrop, OpenAI has committed to roughly $250 billion in Azure cloud spending with Microsoft alone, on top of separate multi-billion-dollar commitments to AWS, Oracle, and others. These are not aspirational numbers — they are contracted spending commitments stretching out toward 2030 and beyond. A company generating $25 billion a year and burning closer to $27 billion cannot fund that kind of infrastructure spend from operations. It needs external capital, and a lot of it.
Private funding rounds have been extraordinary — OpenAI’s valuation reportedly reached around $852 billion after a $122 billion round, up from roughly $110 billion just months earlier. But private capital at this scale eventually runs into its own limits: existing investors want liquidity, new investors want a path to exit, and the sheer size of future funding needs starts to exceed what private markets alone can comfortably absorb. An IPO — reportedly being targeted for Q4 2026 — isn’t really optional at this point. It’s closer to a structural necessity given the spending commitments already locked in. It’s a dynamic we’ve touched on before when looking at why so many software companies are still burning cash despite years of “path to profitability” promises — at some point, growth-at-all-costs has to be paid for by someone, and increasingly that someone is public market shareholders.
“OpenAI didn’t choose to go public because it wanted to. It’s going public because the spending commitments it has already signed leave it little alternative.”
What Changes for Microsoft Once OpenAI Is Public
Here is the part that matters for Microsoft shareholders specifically. Microsoft’s 27% stake in OpenAI is currently held as an investment in a private company — its value on Microsoft’s books is based on the most recent funding round valuation, updated periodically. Once OpenAI is publicly traded, that stake becomes marked to a daily market price.
This sounds like a technical detail. It isn’t. Under standard accounting treatment for equity investments of this size, fluctuations in the value of that stake flow through Microsoft’s income statement — specifically, through “other income and expense,” below the operating income line but still part of net income and GAAP earnings per share. A 27% stake in a company valued near $850 billion is worth roughly $230 billion. If OpenAI’s public market valuation swings 10% in a quarter — which, for a newly public AI company in a volatile sector, is not a dramatic assumption — that is a swing of roughly $23 billion flowing through Microsoft’s reported earnings, in a direction that has nothing to do with how Windows, Azure, or Microsoft 365 actually performed that quarter.
The Numbers That Will Actually Matter
If headline net income and GAAP EPS are going to become noisy and less informative — driven as much by OpenAI’s stock price as by Microsoft’s own operations — what should investors actually be watching? A few things, in rough order of importance:
- Operating income. This sits above the line where OpenAI equity fluctuations get recorded. It reflects Microsoft’s actual core business performance — Azure, Microsoft 365, LinkedIn, Gaming, and so on — without the noise of investment mark-to-market swings.
- Free cash flow. Cash generation from operations is unaffected by paper gains or losses on an equity stake. If Microsoft’s free cash flow keeps growing while net income swings around because of OpenAI’s share price, that’s a sign the underlying business is fine and the headline number is just noise.
- Azure revenue growth, specifically. This is the number that tells you whether the Microsoft-OpenAI relationship is actually paying off commercially — separate from the equity stake entirely. Azure and cloud revenue grew 29% year-over-year in the most recent quarter, with Azure itself guided at 39-40% growth for the following quarter. That’s the real signal.
- GAAP net income and EPS, in isolation. Once OpenAI is public, these figures will increasingly reflect a mix of Microsoft’s actual business results and the market’s ever-changing opinion of OpenAI’s value. A quarter where GAAP EPS falls sharply could simply mean OpenAI’s stock had a bad quarter — not that Microsoft’s business deteriorated.
- Quarter-over-quarter “swings” in reported profitability. If net income jumps 40% one quarter and falls 20% the next, the instinct is to look for an operational explanation. Increasingly, the explanation may simply be OpenAI’s share price — and reacting to that as if it were a Microsoft-specific signal would be a mistake.
The Double Exposure Most People Miss
There’s a second layer to this that gets less attention than the equity stake itself: Microsoft isn’t just an investor in OpenAI — it’s also OpenAI’s largest customer and infrastructure provider, via that $250 billion Azure commitment. This creates a relationship where Microsoft is exposed to OpenAI’s fortunes from two directions simultaneously.
If OpenAI thrives — strong revenue growth, successful IPO, expanding enterprise adoption — Microsoft benefits twice: its equity stake appreciates, and Azure consumption under the $250 billion commitment scales up as promised. This is the bull case, and it’s a genuinely strong one. Wedbush’s Daniel Ives has framed the restructured partnership as Microsoft trading some guaranteed near-term revenue for a more flexible, multi-model position — not as Microsoft losing leverage.
But if OpenAI stumbles — and a company burning $27+ billion a year with no path to profitability before 2030 is not without risk — Microsoft is exposed on both sides at once. The equity stake would lose value, and the commercial relationship that was supposed to drive a huge chunk of future Azure growth would underperform expectations. The recent restructuring, which removed Microsoft’s exclusivity over OpenAI’s models and capped the revenue share OpenAI pays Microsoft at $38 billion through 2030, can be read as Microsoft quietly hedging this exposure — diversifying its AI position rather than betting everything on one partner’s continued dominance. It’s a similar hedging instinct to what we’ve seen play out in the renewed AI browser war, where the big platforms are no longer content to rely on a single partner for their core AI capabilities.
What This Means in Practice
None of this means Microsoft is a bad investment, or that the OpenAI relationship is a mistake. Microsoft’s core business is performing well — double-digit revenue growth, expanding margins, an AI segment that’s crossed a $37 billion annualized run rate on its own. The concern here isn’t about Microsoft’s fundamentals. It’s about how those fundamentals will be reported, and how easy it will be for the headline numbers to obscure what’s actually happening.
For investors, the practical takeaway is straightforward: when OpenAI goes public and Microsoft’s quarterly results start to include large, OpenAI-driven swings in reported net income, don’t anchor on that number. Look at operating income. Look at free cash flow. Look at Azure’s growth rate specifically. These tell you how Microsoft’s actual business is doing. The headline net income figure, increasingly, will tell you as much about sentiment toward OpenAI’s stock as it does about Microsoft.
This is also, in a smaller way, a preview of a broader dynamic. As more AI companies go public and as more large tech companies hold significant equity stakes in them — Nvidia’s stakes in various AI labs, Amazon’s stake in OpenAI via its own investment, and so on — investors across the sector are going to need to get more comfortable separating “what the core business earned” from “what our AI investments happened to be worth this quarter.” Microsoft and OpenAI are simply the first and largest example of this pattern to play out at scale.
This article represents the opinion of the AllinAllSpace editorial team and does not constitute financial advice. Figures referenced are based on publicly available data and analyst estimates as of June 2026 and are subject to change. OpenAI’s IPO timeline, valuation, and structure remain subject to change pending final corporate decisions. Always conduct your own research before making investment decisions.