OpenAI’s growth miss just made its demand story harder to trust
OpenAI’s latest reported miss matters because the company has spent weeks telling investors, cloud partners, and customers that demand for its products is big enough to support one of the largest infrastructure bets in tech. Now Reuters, citing The Wall Street Journal, says OpenAI missed internal targets for ChatGPT users and revenue. That does not prove the AI boom is breaking. It does mean the growth story behind OpenAI’s expansion just got harder to take on faith.
The second pressure point is where the slowdown showed up. Reuters reported that OpenAI missed multiple monthly revenue targets earlier this year after losing ground to Anthropic in coding and enterprise markets, not just in broad consumer growth. Those are the parts of the business that matter most if OpenAI wants to show that AI demand is durable, paid, and defensible rather than just huge.
That tension is visible in OpenAI’s own numbers. In its April funding announcement, the company said it had closed a $122 billion funding round at an $852 billion post-money valuation, had more than 900 million weekly active users, over 50 million subscribers, and was generating $2 billion in revenue per month. In a January post about its business model, OpenAI said revenue grew from $2 billion in annual recurring revenue in 2023 to more than $20 billion in 2025, while compute use, the electricity-hungry data center capacity needed to run and train models, grew from 0.2 gigawatts to about 1.9 gigawatts over the same period.
That does not mean Reuters proved OpenAI’s compute buildout is suddenly in trouble. Giskard was right to bounce the earlier version on that point. The public record shows something narrower and still important: a company making enormous claims about demand has now been reported to have missed internal targets in the very segments investors treat as most valuable.
CNBC reported that chief financial officer Sarah Friar has been working with other executives to clamp down on costs while the board scrutinizes computing deals more closely. CNBC also reported that Oracle shares fell 4 percent, CoreWeave dropped more than 5 percent, and SoftBank fell about 10 percent after the Journal story. Those moves do not prove OpenAI has changed its plans. They do show how many other companies have tied part of their own story to OpenAI’s ability to keep converting AI excitement into paid usage.
OpenAI, for its part, is still arguing that the demand is real. The company said in January that capital is committed in tranches against real demand signals, meaning the full spending plan is supposed to scale with actual usage rather than blind optimism. It has also kept widening the set of products meant to absorb that demand. In a product post from five days ago, OpenAI said GPT-5.5 was rolling out to Plus, Pro, Business, and Enterprise users in ChatGPT and Codex, its coding product.
The cleanest skeptical read is also the simplest one. Internal targets at frontier AI companies are often aggressive, and the strongest negative claims here still reach the public secondhand through Reuters and CNBC rather than directly from the Journal. Altman and Friar have publicly denied reports of a rift over strategy. A missed internal target is not the same thing as a broken business.
But it is enough to change the burden of proof. OpenAI is no longer asking the market to believe only that ChatGPT is popular. It is asking partners and buyers to believe that demand in coding, enterprise, and consumer products will stay strong enough to justify extraordinary spending and extraordinary valuations. Reuters does not show that thesis has failed. It does show why investors, cloud partners, and customers will want fresher evidence before assuming the next leg of that story arrives on schedule.
What to watch next is whether OpenAI produces harder evidence that paid demand in coding and enterprise is keeping pace with the scale of the business it has already promised.