OpenAI’s Quarterly Revenue Tops $10 Billion for the First Time as Losses Widen to $10.9 Billion

OpenAI has crossed a line that no artificial intelligence company has crossed before. In the three months to the end of September 2026, the company behind ChatGPT brought in more than $10 billion of revenue in a single quarter, according to figures shared with investors and first reported on November 10 and 11. The number is roughly 50 percent higher than the quarter before it and close to three times the level of a year earlier, and it puts OpenAI’s annual revenue run rate at about $50 billion. Yet the same set of figures carries a second story that is just as striking: the company’s operating loss widened to $10.9 billion in the same quarter, meaning it still spent far more than it earned even as its sales exploded. The two numbers together capture the strange economics of the AI boom in a single frame — a business growing faster than almost any in corporate history, and burning cash at a pace to match. This is the story of OpenAI $10 billion quarterly revenue November 2026, what drove it, what it cost, and what it does and does not prove about the future of the world’s most watched AI company.

1. The Headline Numbers: $10 Billion in a Quarter

1.1 What the Figures Say

The core figures are simple enough to state and remarkable enough to need a second look. OpenAI’s revenue for the third quarter of 2026 topped $10 billion for the first time, according to reporting by Bloomberg, The Wall Street Journal and The Information, all drawing on information the company shared with investors. The quarter before, the second quarter of 2026, brought in about $6.7 billion. The quarter before that, the first quarter, brought in about $5.7 billion. Put side by side, the sequence tells its own story: $5.7 billion, then $6.7 billion, then more than $10 billion. The jump from the second quarter to the third was close to 50 percent in three months, a rate of growth that large companies almost never sustain, and the year-on-year comparison is steeper still, with quarterly revenue reported to have nearly tripled from the same period in 2025. On an annualised basis, the run rate — the pace of sales projected over a full year at the current level — now stands at roughly $50 billion. For a company that is only a decade old, and that was for most of its life a research organisation rather than a commercial giant, the scale is extraordinary. It is worth pausing on what a run rate of $50 billion means in plain terms: it places OpenAI’s current pace of sales alongside the annual revenues of some of the world’s largest established software and media companies, and it does so only a few years after ChatGPT turned the company into a household name. The figures were presented to investors on a call on Tuesday, November 10, 2026, and OpenAI’s chief financial officer, Sarah Friar, described the period as a blockbuster quarter. That description is hard to argue with on the revenue line alone. What makes the quarter genuinely unusual, though, is not the size of the revenue but the fact that it arrived alongside a loss that grew even larger, and it is that pairing which the rest of this article examines.

1.2 How the Quarter Compares

To understand how unusual this quarter was, it helps to look at the pace of growth rather than the absolute numbers. Between the first and second quarters of 2026, revenue growth was reported at about 18 percent — strong by any normal standard, but modest next to what followed. The step from the second quarter to the third, close to 50 percent, represents a sharp acceleration, and it broke a pattern in which the company’s growth had been impressive but not explosive. Earlier in the year, OpenAI had appeared to be coming out of a slower stretch, and some observers had begun to ask whether the first great wave of AI adoption was flattening. The third-quarter figures answer that question, at least for now, in the negative. The comparison with a year earlier is equally stark. A business that nearly triples its quarterly revenue in twelve months is doing something that very few companies of any size have ever done, and it is doing it at a scale where each additional percentage point of growth represents hundreds of millions of dollars. There is also a comparison to be made with the company’s own history. ChatGPT was released to the public in late 2022, and the commercial business built around it is only a few years old. In that short time it has moved from a research curiosity to a subscription business with hundreds of millions of users, and from there to an enterprise supplier whose products sit inside other companies’ workflows. The third-quarter numbers are the clearest single measure yet of how far that journey has gone. They also set a new baseline. Whatever happens next, the company will now be judged against a quarter in which it took in more than $10 billion in three months, and the question for investors, competitors and customers alike is whether that level can be held, let alone beaten.

2. Where the Growth Came From

2.1 ChatGPT Subscriptions and the Consumer Base

The growth was not the work of a single product or a single deal. According to the reporting that followed the investor call, it came from steady expansion in ChatGPT subscriptions, from progress in the enterprise business built around the company’s application programming interfaces, and from cloud computing partnerships. The consumer side remains the foundation. OpenAI says it has more than a billion users, a figure that covers the free and paid tiers of ChatGPT together, and the paid subscriber base has been the engine of the consumer business since subscriptions were introduced. Hundreds of millions of people now use the company’s products every day, and the daily inference load — the computing work required to answer each of those users in real time — is itself one of the largest computing operations in the world. That scale is both the source of the revenue and the source of the cost, a tension that runs through the whole quarter. The subscription business has also been broadening. ChatGPT is no longer only a chatbot that answers questions; it is a workspace in which people write, code, analyse documents and build things, and each of those uses gives the company another reason to charge, and another reason for users to stay. The reported figures do not break out exactly how much of the quarter’s revenue came from consumers and how much from businesses, and that split matters, because the two kinds of revenue behave differently. Consumer subscriptions tend to be steadier and more predictable, while enterprise deals can be larger and lumpier. What is clear from the reporting is that both sides grew, and that the consumer base provided the volume on which the rest of the business was built.

2.2 Enterprise, APIs and Codex

If the consumer business provided the volume, the enterprise side provided much of the acceleration. The company’s application programming interface business, which lets other companies build OpenAI’s models into their own products, has been growing steadily, and the cloud partnerships that carry those models to customers have become a major channel in their own right. But the most striking detail to emerge alongside the revenue figures concerns Codex, OpenAI’s coding tool. Greg Brockman, the company’s president, said at a public event on November 5 that Codex gained real market recognition in the third quarter, and that its sales surpassed the consumer business for the first time to become the largest single source of revenue. That is a significant shift. Coding tools were, until recently, a niche inside the AI industry; they are now one of its biggest commercial battlegrounds, with developers willing to pay for tools that write, review and debug code faster than a human team can. Codex competes directly with rival coding products, most notably Claude Code from Anthropic, and market research cited in the reporting suggests that competition is fierce. The fact that Codex has become OpenAI’s largest revenue source says something about where the money in AI is moving: away from general chat and toward specialised tools that do a specific, valuable job inside a working day. It also helps explain the size of the third-quarter jump, because enterprise and developer spending can scale much faster than consumer subscriptions. A single company deciding to roll out a coding tool to thousands of engineers is worth far more than the same number of individual subscribers, and those decisions, once made, tend to stick. The enterprise business, in other words, is where the growth curve steepens, and Codex appears to be its sharpest edge.

3. The Other Side of the Ledger: A $10.9 Billion Loss

3.1 What a Minus 109 Percent Margin Means

The loss figures are, in their own way, as remarkable as the revenue. OpenAI’s operating loss widened to $10.9 billion in the third quarter, up from about $10.1 billion in the second quarter, according to documents reviewed by The Wall Street Journal. The company’s adjusted operating margin — the loss expressed as a percentage of revenue — improved to minus 109 percent, from minus 150 percent in the previous quarter. Those percentages deserve a plain explanation, because they are the key to understanding the company’s finances. A margin of minus 109 percent means that for every dollar of revenue the company took in, it recorded about $1.09 of adjusted operating loss. In other words, it spent roughly $2.09 for every $1.00 it earned. That is an improvement on the previous quarter, when the equivalent figure was about $1.50 of loss per dollar of revenue, and it is a genuine step in the right direction. But it is still a very long way from profitability, and it means that the faster the company grows, the more cash it burns in absolute terms — at least for now. The $10.9 billion loss is larger than the entire quarterly revenue of most large companies, and it is larger than OpenAI’s own revenue was only a few quarters ago. This is the paradox at the heart of the AI business model as it currently stands: the products are selling faster than ever, and the costs of building and running them are rising just as fast, if not faster. The margin improvement shows that revenue is growing faster than costs, which is the direction any investor wants to see. The absolute loss shows how far there still is to go.

3.2 Why the Loss Grew Even as Revenue Soared

The reason the loss grew is not mysterious, and it is not a sign of waste in the ordinary sense. It is the cost of the computing power on which the whole business depends. Training frontier AI models requires vast clusters of graphics processors running for weeks or months, and serving hundreds of millions of users every day requires still more computing capacity, all of it rented or bought at enormous expense. The reporting on the quarter points to three main cost drivers: the cost of training new models, the cost of leasing the data-centre computing capacity to run them, and the cost of answering the daily requests of hundreds of millions of active users. Each of these grows with the business. A better model costs more to train. More users cost more to serve. More enterprise customers demand more capacity, more reliability and more support. The company has also been signing long-term deals for computing capacity that commit it to spending far into the future, and those commitments are part of why the losses look so large. There is a genuine debate about whether this spending is an investment that will pay off or a treadmill that speeds up as it goes. The optimists point to the margin improvement — from minus 150 percent to minus 109 percent — as evidence that the business is moving toward a point where each extra dollar of revenue costs less than a dollar to produce. The sceptics point to the absolute numbers and ask how long any company can lose more than $10 billion in a quarter, however fast it grows. Both readings are present in the coverage of these results, and neither can be settled by a single quarter’s figures. What the quarter does show is that the cost of staying at the frontier of AI is now measured in tens of billions of dollars a year, and that only a handful of companies in the world can afford to pay it.

4. The Investor Call That Revealed the Numbers

4.1 What Sarah Friar Told Investors

The figures did not come from a public earnings announcement, because OpenAI is a private company and does not publish audited financial statements in the way a listed company must. They came from an investor call on Tuesday, November 10, 2026, at which the company’s leadership shared the third-quarter numbers with existing and prospective investors. Chief financial officer Sarah Friar called it a blockbuster quarter, and the phrase was picked up widely in the coverage that followed, because it captured the tone of a company that believes its commercial momentum has turned a corner. The call was also the moment at which the scale of the business became concrete for people outside the company. Until then, OpenAI’s revenue had been a matter of estimates and leaks; now there was a number, shared by the company itself, that put quarterly revenue above $10 billion and the annual run rate at about $50 billion. The call matters for another reason, too. OpenAI has been preparing the ground for a possible public listing, and it filed confidentially for an initial public offering earlier in 2026, according to earlier reporting. Investor calls of this kind are part of how a private company builds the case for a listing: by showing institutional investors that the business is large, growing and, eventually, capable of making money. The third-quarter figures are plainly intended to be read in that light. They say, in effect, that the company has reached a scale at which it can be judged as a major business rather than a research project, and that its growth has accelerated rather than slowed. Whether investors accept that reading will depend on what happens next, and in particular on whether the losses begin to narrow in absolute terms rather than only as a percentage.

4.2 Why the Figures Are Reported, Not Confirmed

It is important to be precise about the status of these numbers, because precision is part of what makes them useful. The figures have been reported by Bloomberg, The Wall Street Journal and The Information, citing people familiar with the information shared with investors and documents seen by the Journal. They have not been confirmed by OpenAI in a public statement, and they have not been independently verified in audited financial statements. That does not make them unreliable — three separate outlets reporting the same core figures, with the Journal citing documents, is strong evidence — but it does mean they should be described as reported rather than as confirmed, and that is how this article describes them. The distinction matters because the details are sensitive. The exact split between consumer and enterprise revenue, the precise definition of the adjusted operating margin, and the treatment of partner sales all affect how the numbers should be read, and none of those details has been published in full. There is also the question of what counts as revenue when a sale is made through a cloud partner, a question that becomes important when comparing OpenAI with its rivals, as the next section explains. For now, the honest summary is this: the best available reporting, from multiple outlets with access to the investor materials, says OpenAI’s third-quarter revenue crossed $10 billion, its operating loss widened to $10.9 billion, and its adjusted operating margin improved to minus 109 percent. Those are the figures on which the rest of this article is based, and they are presented here as reported figures, not as the company’s own public disclosure.

5. OpenAI Against Anthropic: A Rivalry Measured in Billions

5.1 Anthropic’s $17 Billion Quarter

The most startling comparison in the coverage of OpenAI’s quarter is not with the company’s own past but with its chief rival. Anthropic, the company behind the Claude family of models, reportedly posted about $17 billion of revenue in the same third quarter, exceeding OpenAI’s figure by a significant margin, according to reporting that drew on Anthropic’s own prospectus materials. Anthropic’s annual run rate is put at about $81 billion after a reported $10 billion monthly increase, and its quarterly revenue is said to have risen roughly twelvefold year on year. Those numbers, if accurate, would make Anthropic the larger company by revenue, a remarkable reversal for a firm that was, until recently, widely seen as the smaller challenger. The prospectus figures also show a different profit picture: Anthropic is reported to have reached break-even in the second quarter, only for the third quarter to swing back to an operating loss of $567 million. That loss is far smaller than OpenAI’s $10.9 billion, but the swing shows that Anthropic, too, is spending heavily to keep up. The rivalry between the two companies has become the central contest in the AI industry. They compete for the same enterprise customers, the same developers, the same computing capacity and the same researchers, and each new set of numbers is read as a score in that contest. On the revenue measure, the third quarter belongs to Anthropic. On the scale of the consumer business, OpenAI still leads by a wide margin, with more than a billion users against a much smaller consumer base at Anthropic. The two companies are, in effect, winning different races, and the question of which race matters more is one that the market has not yet settled.

5.2 Why the Two Counts Are Not Like for Like

Before the comparison is taken too far, there is an important accounting caveat, and it is one that the reporting itself flags. The two companies do not count revenue in the same way, and the difference is large enough to change the picture. Anthropic counts sales made through its cloud partners in its revenue figure, while OpenAI counts only its share of some partner sales. That means a dollar of business done through a cloud partner can appear in full in one company’s total and only in part in the other’s, even when the underlying work is similar. The result is that Anthropic’s $17 billion and OpenAI’s $10 billion are not directly comparable, and anyone who treats them as a simple league table is being misled. There are other differences, too. Anthropic’s figures come from prospectus materials prepared for its own public listing, which are subject to a different set of rules and disclosures than the investor-call figures on which OpenAI’s numbers rest. The two companies also have different mixes of consumer and enterprise business, different pricing, and different deals with cloud providers, all of which affect what shows up as revenue. None of this means the comparison is useless. It does mean it should be handled with care, and that the headline gap between $17 billion and $10 billion overstates the difference between the two businesses. The honest conclusion is narrower but still significant: both companies are now operating at a scale of tens of billions of dollars a year, both are growing very fast, and both are spending heavily to do so. The rivalry is real, and it is expensive for both sides.

6. What the Numbers Mean for OpenAI’s Spending Plans

6.1 Chips, Data Centres and the TSMC Connection

The revenue figures land in the middle of a spending programme that is itself one of the largest in corporate history, and the two cannot be understood apart. OpenAI’s growth depends on access to computing capacity, and computing capacity depends on chips, and chips depend on the handful of companies that can make them. That chain is why the company’s finances are so closely watched by the chip industry, and why the chip industry’s news is so closely watched by OpenAI’s investors. Only days before these revenue figures emerged, the board of TSMC, the world’s largest chipmaker, approved up to $1 billion of investments in OpenAI and Anthropic — up to $500 million in OpenAI preferred stock and up to $500 million in Anthropic shares at its listing price — a move that was itself the subject of a separate report on this site. The connection is not a coincidence. TSMC makes the advanced chips on which AI training depends, and its decision to take stakes in its biggest AI customers is a sign of how intertwined the two industries have become. For OpenAI, the spending implied by its losses is spending on exactly this chain: on graphics processors, on data centres to house them, and on the energy to run them. The company has signed long-term agreements for computing capacity that run into the tens of billions of dollars, and those agreements are part of the reason its losses are so large. The question the third-quarter figures raise is whether the revenue growth justifies the scale of those commitments. A run rate of $50 billion makes the commitments look more affordable than they did a year ago. A loss of $10.9 billion in a quarter makes them look more demanding than ever. Both readings will be argued over in the months ahead, and the answer will depend on whether the revenue curve keeps bending upward faster than the cost curve.

6.2 The Path Toward an IPO

The other context in which these figures must be read is OpenAI’s path toward a public listing. The company filed confidentially for an initial public offering earlier in 2026, according to earlier reporting, and the third-quarter numbers are plainly part of the story it will tell public investors when the time comes. A public listing would be one of the largest in history, and it would require the company to explain, in public filings, how a business that loses more than $10 billion in a quarter can be worth the valuation it seeks. The revenue figures help with that explanation: a company growing at close to 50 percent a quarter, with a $50 billion run rate and a product used by more than a billion people, is a company that public investors will want to own. The loss figures make the explanation harder: a company that spends $2.09 for every $1.00 it earns will need to show a credible path to profitability, and it will need to show it soon. Anthropic’s own listing, which is further advanced, provides both a model and a warning. Its prospectus shows a company with larger reported revenue and a much smaller loss, but also a company whose break-even lasted only a single quarter before swinging back to a loss. The two listings, if both go ahead, will be judged against each other, and the market will decide which version of the AI business it believes in. For OpenAI, the third-quarter figures are a strong opening statement. They are not, by themselves, a complete case. The company will need more quarters like this one, and it will need the losses to begin narrowing in absolute terms, before the case is closed.

7. How Markets and the Industry Reacted

7.1 The AI Stock Rally

The figures arrived in a week in which AI stocks were already rallying, and they added fuel to that rally. On the Wednesday after the investor call, the S&P 500 rose about 1.4 percent and the Nasdaq about 1.7 percent, with AI-related shares leading the advance, according to market reporting. The rally was driven by more than OpenAI’s numbers alone — strong results from cloud computing firms and renewed enthusiasm for AI infrastructure played their part — but the OpenAI figures were read as confirmation that demand for AI products is still accelerating, and that the companies supplying the infrastructure behind them have a growing market to serve. The chipmakers were among the biggest beneficiaries, with shares in the companies that make AI processors rising sharply, and the cloud providers that host AI workloads also gained. The reaction is a reminder of how far OpenAI’s fortunes now reach beyond the company itself. Its spending is other companies’ revenue, and its growth is the reason investors are willing to pay high prices for shares in chipmakers, data-centre operators and cloud providers. That connection cuts both ways. If OpenAI’s growth were to slow, the companies that depend on its spending would feel it quickly. For now, though, the market has taken the third-quarter figures as good news, and the rally that followed is the clearest sign of that. It is also a sign of how much is riding on the AI boom continuing. The valuations of the companies in the AI supply chain assume that the growth OpenAI has just reported will continue, and that the spending behind it will continue with it.

7.2 What Analysts Are Watching Next

Analysts and commentators who have examined the figures have focused on a small number of questions, and those questions are likely to shape the coverage of the next quarter as well. The first is whether the margin improvement continues. A move from minus 150 percent to minus 109 percent is a step in the right direction, but the next step matters more: if the margin improves again, the path to profitability starts to look real; if it stalls or reverses, the sceptics will have their answer. The second is whether Codex can sustain its position as the largest revenue source. Coding tools are a fiercely competitive market, and Anthropic’s Claude Code is a formidable rival; a shift in that contest could move the revenue figures by billions. The third is the cost of computing capacity. The price of the chips and data-centre space on which the whole business depends is itself moving, and a rise in those costs would widen the losses even if revenue keeps growing. The fourth is the public listing. The timing of OpenAI’s IPO, and the valuation it seeks, will be read as a verdict on these figures, and the market’s reception of Anthropic’s listing will be watched as a guide. Finally, there is the question of the user base. More than a billion users is an extraordinary number, but the company will need to show that it can convert more of them into paying customers, and that the paying customers it has will keep paying as the novelty of AI wears off. These are the questions that will decide whether the third quarter of 2026 is remembered as a turning point or as a high-water mark.

8. What Comes Next

8.1 The Questions the Next Quarter Must Answer

The next set of figures, covering the final three months of 2026, will be watched more closely than any the company has yet produced, because they will show whether the third quarter was a step change or a one-off. The first thing to look for is the revenue number itself. If it holds above $10 billion and grows again, the acceleration will look like a trend; if it falls back, the third quarter will look like a spike driven by a small number of large deals. The second is the loss. The market will want to see the operating loss narrow in absolute terms, not merely as a percentage of a larger revenue base, because that is the point at which the business model starts to look sustainable. The third is the split between consumer and enterprise revenue, and in particular whether Codex continues to lead. A business that depends on a single fast-growing product is more vulnerable than one with several, and the company will be judged on the breadth of its growth as much as on its speed. The fourth is the IPO. If OpenAI files publicly or sets a date, the figures it discloses then will be the first audited, public account of its finances, and they will either confirm the reported numbers or revise them. And the fifth is the competitive picture. Anthropic’s listing, and the figures it discloses, will set a benchmark against which OpenAI will be measured, and the two companies’ next quarters will be compared line by line. None of these questions can be answered today. They are the reason the third-quarter figures, remarkable as they are, should be read as the beginning of a story rather than its end.

8.2 The Bottom Line

OpenAI’s third quarter of 2026 will be remembered as the quarter in which the company’s revenue passed $10 billion for the first time, rising close to 50 percent in three months to put its annual run rate at about $50 billion. It will also be remembered as the quarter in which its operating loss widened to $10.9 billion, and its adjusted operating margin, though improved to minus 109 percent, still meant it spent more than $2 for every $1 it earned. Both facts are true at the same time, and neither can be understood without the other. The revenue shows a business that has found a vast and growing market for its products, from ChatGPT subscriptions to enterprise APIs to Codex, now its largest single source of sales. The loss shows the cost of serving that market, measured in chips, data centres and computing power on a scale that only a handful of companies can afford. The comparison with Anthropic, which reported even larger revenue on a different accounting basis, shows that the contest at the top of the AI industry is now measured in tens of billions of dollars a quarter. The market’s reaction, a broad rally in AI shares, shows how much of the wider economy is now tied to that contest. What the quarter does not show is whether the business can be made profitable, how long the growth can continue, or what the company will be worth when it finally lists its shares in public. Those are the questions the coming days and quarters will provide the answers to, and they are the questions on which the future of OpenAI — and, to a surprising extent, the future of the AI industry — will turn. The coming days will provide the answers.

Comments

comments