OpenAI Q3 Revenue Tops $10 Billion, But Losses Hit $10.9 Billion

OpenAI generated $10 billion in revenue in the three months ended in September, its first quarter at that level, according to Bloomberg. Revenue rose about 50% from the second quarter and was nearly three times the year‑earlier level. The catch: the company's adjusted operating loss for the same period was $10.9 billion, up from $10.1 billion the prior quarter.

The figures were shared with investors on a call Tuesday, where President Greg Brockman and CFO Sarah Friar spoke, according to Quartz. OpenAI declined to comment.

Behind the headline loss, there's a real efficiency story. OpenAI lost $1.09 for every dollar of revenue in Q3, down from $1.50 three months earlier, Forbes reported. Its adjusted operating margin improved to -109% from -150% in Q2.

Most of the loss is the cost of buying compute ahead of the revenue it produces. OpenAI has signed massive infrastructure commitments, including a $300 billion five‑year agreement with Oracle, roughly $250 billion in Microsoft Azure purchase commitments, and a $38 billion seven‑year deal with Amazon Web Services, per.² Demand exceeds available compute, so added capacity converts to revenue quickly — the question is when the loss curve finally flips.

OpenAI's enterprise revenue doubled quarter over quarter and now makes up 50% of total revenue, making it the largest single source. CFO Sarah Friar attributed the growth to products aimed at builders and developers, telling ² that the driver is products like Coding and Cyber. Consumer sales rose 20% in the quarter, with subscriptions topping 100 million and ChatGPT reaching 1.2 billion weekly active users.

The contrast with Anthropic is stark. Anthropic posted roughly $17 billion in Q3 revenue — nearly $7 billion more than OpenAI — while its operating loss was just $567 million, a fraction of OpenAI's $10.9 billion, according to Forbes and Quartz.

Anthropic first surpassed OpenAI in quarterly revenue in Q2 2026, posting $11.6 billion against OpenAI's $6.7 billion, Quartz noted. One caveat: the two companies use different accounting methods. OpenAI's numbers come from investor documents, while Anthropic's are from its IPO filing, so the figures aren't perfectly comparable. OpenAI's ~$50 billion annualized run rate also excludes gross partner revenue that was once folded into a widely cited $68 billion figure.

OpenAI is finalizing a funding round that values it at $1.25 trillion — below the $1.4 trillion it initially sought. The company has closed an initial $11 billion tranche, Reuters reported, out of a round that was once targeted at up to $30 billion but is now expected to land closer to $20 billion, per.⁴

Meanwhile Anthropic is targeting a $1.4 trillion valuation in an IPO that's already oversubscribed. OpenAI confidentially filed for its own IPO in June 2026 but has told employees it won't go public until 2027. The valuation gap — and the efficiency gap — will shape how public markets price both companies.

For developers, the numbers confirm what the product roadmaps already hinted at: OpenAI's growth now depends on coding agents, security tooling, and AI assistants sold into the enterprise — not just consumer ChatGPT subscriptions. Friar's reference to Coding and Cyber products signals where OpenAI will keep investing, which matters if you're building on top of those APIs.

The compute math also matters. OpenAI is spending billions to lock up Oracle, Azure, and AWS capacity, which means API capacity and pricing will stay tied to how fast that compute comes online. If demand keeps outstripping supply, builders should expect continued pressure on pricing and rate limits even as the models get cheaper to run.