FT: OpenAI's Revenue Falls About $20 Billion Short of What Investors Expected, and AI Stocks Wobble

The Financial Times reported on Thursday that OpenAI is earning roughly $20 billion less on an annualized basis than investors were led to believe. For a company whose valuation and spending plans rest on the expectation of explosive revenue growth, this is not a cosmetic correction. The reaction was immediate: AI stocks slid, and the debate over the sustainability of the business around large models moved back to the center of the agenda.

What we know so far

The initial summary of the report contains few facts, and it is worth stating that plainly. The size of the gap is known: on the order of $20 billion in annualized revenue. The source is known: an FT investigation. The framing is known: investors were "steered" toward higher expectations. What is not known is which documents and figures the calculation rests on, which projected figure the actual number was compared against, and which periods were used as the basis. The available text gives no OpenAI position to cite, so it is too early to assess the company's comments.

It is also unclear whether the report points to failed contracts or to revenue that grew more slowly than promised in presentations. The difference is fundamental. In the first case, the problem is demand; in the second, it is how the company managed market expectations.

Why "annualized revenue" is easy to misread

Annualized revenue is calculated simply: take income over a short period, usually a month or a week, and extrapolate it to a full year. The metric is convenient for fast-growing companies, but it is sensitive to assumptions. A single peak month, one-off large contracts, discounts that have not yet shown up in payments, or users on trial plans can all inflate the figure. Different market participants often calculate it differently, and a gap of tens of percent is possible here without anyone acting in bad faith.

So the question about the report goes beyond "how much." We need to know which figure was used as the reference point and who put it forward: the company itself, intermediaries in funding rounds, or analysts.

Why the market is reacting so nervously

The reason is asymmetry. Together with its partners, OpenAI has signed commitments worth hundreds of billions of dollars for compute and chips, and those commitments assume a scenario of very rapid revenue growth. If revenue lags, the pressure spreads along the whole chain: GPU makers, cloud providers, data center builders, energy suppliers. Their valuations rely heavily on the assumption that one of the largest buyers of capacity will be able to pay its bills.

Key parameters of the story:

• the gap: about $20 billion on an annualized basis;
• the metric: annualized revenue, meaning an extrapolation rather than reported revenue for the year;
• the source: the Financial Times, published on Thursday;
• the market reaction: pressure on stocks of AI-sector companies;
• undisclosed: the calculation method, the baseline forecast, OpenAI's response.

What to watch next

First: OpenAI's official response and, if one comes, any clarification of its metrics. If the company disputes the calculation, there will be numbers that can be checked. If it does not, the market will read the silence as confirmation.

Second: earnings from partners and suppliers. Quarterly results from cloud companies and chipmakers will show how much actual payment stands behind the announced contracts and whether purchasing plans are changing.

Third: for teams building products on the OpenAI API, the practical takeaway is more modest but more tangible. Build the ability to switch between providers into your architecture, check your contract terms for changes in pricing and limits, and do not tie your roadmap to a single model. This is routine hygiene, but amid news about the finances of a key supplier, it stops being theoretical.

Investors would do well to separate two things: the quality of the technology and the quality of the financial model. The models can be strong while the revenue forecasts are inflated. The FT report speaks to the latter, and so far there is no sign that it calls the former into question.