OpenAI has just walked into an awkward financial situation. Its yearly revenue pace is $50 billion, not the $70 billion investors were hearing about in September.
That’s a $20 billion difference for a company currently valued at $852 billion and preparing to sell shares to the public.
Before anyone starts looking for missing money, though, there’s a catch. OpenAI hasn’t necessarily lost those sales. The question is how it has been doing business in competition with Anthropic, which appears to be making Wall Street unhappy.
This smaller number was drawn to attention by Financial Times, whereas Axios later revealed the reason behind why investors were given a far greater estimate before.
Presumably, the initial calculation was made to help compare the financial state of OpenAI to that of Anthropic. However, the companies do not treat some of their sales equally, thus complicating their revenues.
Here is how AI companies provide their products using cloud platforms.
When Anthropic sells software via cloud, it recognizes the total amount the client paid for the product as revenue, accounting for the platform’s cut as an expense of business.
In contrast, OpenAI does it in another way when making some deals.
Neither company is necessarily breaking accounting rules. Both follow GAAP, the standards governing financial reporting in the United States.
The important thing is who controls the transaction, who owns the customer relationship and is responsible for delivering the service.
OpenAI’s higher figure was essentially treating partner transactions as Anthropic did. The investors wanted to see similar numbers, and they got them, although it made their company appear much bigger through the headline.
Investors didn’t exactly welcome the accounting lesson.
In Thursday’s trading session, Nvidia (NVDA) dropped 3%, Oracle (ORCL) lost 6%, and CoreWeave (CRWV) tumbled 8%.
The decline was widespread in the semiconductor space. AMD (AMD) and Broadcom (AVGO) both fell 5%, while Intel (INTC) and Super Micro Computer (SMCI) fell 6% each.
The news could not come at a worse time.
OpenAI filed confidential paperwork for an IPO in June 2026, and its leadership team is weighing a stock market listing at some point in 2027.
Anthropic itself is preparing to go public, and its leaders have already contacted potential investors, aiming to achieve an outlandish valuation of $2 trillion.
Its reported annualized sales pace reached $65 billion by late July, according to figures shared with investors in August.
However, independent research firm New Constructs did not mince any words when giving its evaluation of the company’s proposed IPO – the company, the report says, warrants a valuation of no more than $150 billion.
Some financial information collected by Reuters gives us some insight into why.
According to reports, Anthropic earned revenues amounting to $4.6 billion in 2025 but had a net loss of $42 billion.
However, finance isn’t the only problem faced by these companies.
Experts have repeatedly expressed their concerns about potential AI catastrophes, while OpenAI has reported instances of its technology exceeding its boundaries.
Safety issues were the reason the company decided to cancel the launch of GPT-6.1 Astra because of the technology failing internal testing requirements.
Speaking about it in September, CEO Sam Altman acknowledged the challenges faced and claimed the safety debate made this an unsuitable period to launch OpenAI onto the public market.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.