OpenAI tops $10 billion in quarterly revenue for the first time

The number behind the announcement

OpenAI's quarterly revenue topped $10 billion for the first time in the third quarter of 2026, according to a Bloomberg report cited by TechCrunch, which cited anonymous sources. The jump represents growth of nearly 50% compared to the $6.7 billion reported in the second quarter, well above the roughly 18% rate recorded between the first and second quarters of the year.

The improvement also shows up in the company's adjusted operating margin, which improved from -150% in the second quarter to -109% in the third. In short: OpenAI is earning revenue much faster than before, but it still burns more than the value of its own revenue in operating expenses.

With that, the company reached an annualized run rate of $50 billion, according to Bloomberg. It's an impressive jump for a company that, according to the report itself, wasn't exactly struggling before: 18% growth on top of $5.7 billion in quarterly revenue was already considered large by any software industry standard.

Anthropic ahead in the enterprise market

The contrast that puts the number in perspective is Anthropic. The rival reported revenue of $17 billion in the third quarter, and said its annualized run rate jumped $10 billion in a single month, reaching $81 billion by the end of October. Anthropic is also seeking a $1.4 trillion valuation in an IPO that, according to TechCrunch, should happen soon.

The difference isn't just one of scale, it's about where each company makes its money. OpenAI says it has more than a billion users, a number that helped build the industry's most recognized brand, but most of that base consists of free users or subscribers to cheap plans. Anthropic, in turn, has consolidated its lead in the enterprise market, where average ticket size and recurrence are higher.

It's worth noting that the two companies calculate annualized revenue differently: Anthropic counts sales made through its cloud partners, while OpenAI reports only its share of part of the sales made through partners. This means comparing OpenAI's $50 billion to Anthropic's $81 billion line by line is misleading: these are metrics built on different criteria.

What explains OpenAI's turnaround

According to TechCrunch, the recent improvement combines a few concrete factors:

• More aggressive pricing and discounts for large accounts
• Better-performing models, which support pricier paid plans
• Greater focus on enterprise-oriented products and on AI agents capable of carrying out full tasks, not just answering questions

This combination is the kind of move that matters to whoever decides where to invest product budget. It's not just "the model got smarter": it's OpenAI trying, in practice, to close the gap Anthropic opened in the enterprise segment, which is precisely where contracts are larger and more predictable than the individual $20-a-month subscriber.

What this changes for those building AI products in Brazil

For a Brazilian founder who relies on a language model API as the core piece of their product, this result has a direct reading: the market's two largest providers are still operating in the red, even while growing fast. That means the price war and discounts that have benefited startups over the past few quarters aren't sustainable indefinitely under current terms, and API repricing remains a real roadmap risk, not a hypothetical event.

The battle between OpenAI and Anthropic for the enterprise market also signals where the big money is going: corporate contracts and agents capable of carrying out full tasks, not just chat. Anyone building a B2B tool with generative AI is, in practice, competing for the same slice of budget that these two companies are chasing directly, which makes product differentiation (not just access to the latest model) an increasingly central survival factor.

Market fragmentation is also a strategic data point. The TechCrunch piece notes that OpenAI and Anthropic also compete with models from Meta, Amazon, Microsoft, and Google, plus a crop of cheaper open-source models, mainly coming from China. For whoever decides product architecture, this reinforces the argument for not tying the business to a single model provider: the market is far from consolidated, and today's cheapest provider may not be the cheapest two quarters from now.

What remains open

OpenAI's numbers come from anonymous sources cited by Bloomberg, not from an audited balance sheet published by the company itself, which is the norm for a company that isn't yet public. Each company's run-rate calculation methodology remains different, which makes any direct ranking between OpenAI and Anthropic a simplification. And the question that remains, both for those who invest and those who build on top of these platforms, is how long the market will tolerate triple-digit negative operating margins before demanding higher prices or real profit.