Imagine the scene inside OpenAI’s investor briefing on Tuesday. A slide lights up the room with a number so large it stops conversation: ten billion dollars in a single quarter. For a company that was a research lab not long ago, it is the kind of figure that sounds like the end of an argument. Then the next slide appears — and the argument starts all over again.
OpenAI Q3 revenue reached about $10 billion in the three months to September, according to investor documents reviewed by The Wall Street Journal and reported by Reuters. It was the first time the ChatGPT maker had crossed the $10 billion threshold in a single quarter, and the jump was dramatic: revenue rose by roughly 50 percent from the previous quarter.
The scale of the growth becomes clearer when you look backwards. A year earlier, in the third quarter of 2025, OpenAI posted revenue of $3.9 billion. In twelve months, quarterly sales have grown by more than two and a half times. Few companies in history have ever grown that fast at that size.
OpenAI Q3 Revenue: The Numbers Behind the Headline
The headline figure is only half the story. OpenAI’s operating loss widened to $10.9 billion in the same quarter, up from $10.1 billion in the second quarter. The company lost about $1.09 for every dollar of revenue it brought in — more money went out than came in, even at record scale.
There is, however, a silver lining buried inside those numbers. OpenAI’s adjusted operating margin improved to minus 109 percent, from minus 150 percent in the previous quarter. In plain terms, the loss grew by only about 8 percent while revenue grew by nearly 50 percent. Each dollar of sales now carries less loss than it did three months ago.
That improvement matters because it shows the direction of travel. OpenAI is not simply burning more money as it grows; it is growing into its costs, slowly. The question investors are really asking is not whether the company loses money today — everyone knows it does — but how long it takes before the lines cross.
The figures also come with a health warning. These are adjusted numbers shared with private investors, not audited public accounts. They exclude stock-based compensation, which the Journal reported at $2.6 billion for the quarter. OpenAI did not respond to a request for comment from Reuters.
Enterprise Is Now Half the Business
The most important change inside the numbers may be who is paying. Chief financial officer Sarah Friar told CNBC that enterprise revenue doubled quarter over quarter, and that the business is now split roughly equally between enterprise customers and consumers.
That is a remarkable shift. Back in April, OpenAI said enterprise customers generated more than 40 percent of revenue and were on track to reach parity with consumer revenue by the end of 2026. It appears to have got there early — a sign that businesses, not just individual ChatGPT users, are now driving the company’s growth.
Friar described the quarter as a blockbuster one and put OpenAI’s annual revenue run rate at roughly $50 billion. If that pace holds, the company is selling AI at a scale no rival can yet match — 1.2 billion people now use ChatGPT every week.
The Anthropic Contrast Investors Cannot Ignore
The timing of the disclosure was no accident. On the very same day, rival Anthropic took its initial public offering on the road, marketing shares to investors ahead of a Nasdaq listing under the ticker ANT, at a proposed price of $70 to $78 a share.
And Anthropic’s numbers tell a very different story. Reports citing its IPO filings put the company’s third-quarter revenue at $17 billion, with an operating loss of just $567 million. If those figures hold up, Anthropic is growing just as fast while losing far less — a comparison that will haunt OpenAI’s fundraising pitch.
Anthropic’s offering is expected to price around November 18, with trading to begin the following day. Its prospectus has turned the two private AI giants into a public test: can OpenAI’s scale justify its spending, or does Anthropic’s efficiency win the argument?
Investors have already voted with their wallets once this month. TSMC, the world’s biggest chipmaker, approved a $1 billion investment split between Anthropic shares and OpenAI preferred stock — a bet that both companies matter, whatever their different paths to profit. The chipmaker’s move is part of a much bigger pattern, as our report on TSMC’s $1 billion bet on Anthropic and OpenAI explains.
Where the Money Actually Goes
So where does $10.9 billion in losses actually go? The short answer is computing power. Training and running frontier AI models requires vast fleets of chips, data centres and energy, and the bill keeps rising as models get bigger and more people use them.
In May, OpenAI president Greg Brockman said the company expected to spend around $50 billion on computing power during 2026 alone. That is roughly the entire annual revenue run rate, spent on the machines that make the products work.
The company is also raising capital at a reported $1.25 trillion valuation, and it closed around $11 billion of funding earlier this month. Add a $38 billion compute deal with Amazon Web Services signed last year, and the picture is clear: OpenAI is buying its future in advance, and paying for it with other people’s money.
The bet is that the spending buys a lead that competitors cannot catch. Every dollar of compute today becomes a smarter model tomorrow, which attracts more users, which brings in more revenue. It is the classic Silicon Valley flywheel — just running at a scale the industry has never seen before.
The numbers have also sharpened a debate that has been building all year: is the AI industry in a bubble? Critics point to the losses and the endless fundraising as proof that the economics do not add up. Supporters answer that this is exactly what the early internet looked like — enormous spending first, enormous profits later. The truth, as usual, probably sits somewhere in between, and the quarterly figures are the only evidence either side has.
What Happens Next for OpenAI
The next test comes soon. Anthropic’s IPO will give the market its first chance to price an AI company of this scale in public, and OpenAI is widely expected to follow with its own listing, possibly as early as 2027.
Until then, the numbers will keep coming every quarter, and each one will be read the same way: is revenue still growing faster than losses? In the third quarter, it was — by a wide margin. If that pattern holds for a few more quarters, the losses start to look like an investment. If it breaks, they start to look like a problem.
For the 1.2 billion people who use ChatGPT each week, none of this changes what appears on their screens tomorrow. But behind the chat window, the most expensive race in technology history is still running — and the bill, for now, keeps getting bigger.
