Coreweave (CRWV), SoftBank Group (9984.T), Broadcom (AVGO), Advanced Micro Devices (AMD), Nvidia (NVDA), and Oracle (ORCL) fell because traders are no longer treating OpenAI’s spending plans like free money.
A report said OpenAI has not hit some of its own growth and sales goals, and that was enough to hit the whole AI infrastructure trade on Tuesday.
The damage was not small. Oracle dropped 4%, even with its $300 billion five-year compute partnership with OpenAI still in place. Broadcom lost 4%. AMD fell 3%.
Nvidia slipped more than 1%. Qualcomm (QCOM) went down 0.2%, though it finished above its weakest level after getting some help Monday from reports that it is working with OpenAI on smartphone chips.
Coreweave, the debt-heavy neocloud stock tied closely to AI compute demand, fell more than 5%. SoftBank, one of OpenAI’s largest investors, sank about 10% in Asia.
OpenAI misses growth targets and investors sell the companies tied to its compute demand
The report said OpenAI has recently missed its own targets for user growth and revenue. That matters because OpenAI has signed massive deals for data centers and long-term computing power.
OpeAI’s finance chief Sarah Friar warned colleagues that slower sales could make it harder for OpenAI to fund future compute deals, which landed hard because OpenAI has become one of the biggest demand engines for the AI supply chain.
OpenAI fought back against the criticisms though. Sam Altman and Sarah said, “We are totally aligned on buying as much compute as we can and working hard on it together every day.” They also said any claim that they are split or stepping back from buying computing resources is “ridiculous.”
Oracle also stood by the partnership. A company spokesperson said, “We’re incredibly excited about our partnership with OpenAI and remain focused on building and delivering the capacity they need to support rapidly growing demand.” The spokesperson added, “OpenAI’s new 5.5 model is a significant step forward, and we expect continued momentum as access to their technology expands across cloud providers.”
For years, Sam has tried to secure as much data-center capacity as OpenAI can get. His view has been that not having enough computing power is the biggest limit on OpenAI’s growth. That thinking led to a huge run of deals last year and left the company tied to about $600 billion in future spending promises.
ChatGPT slows, Gemini gains users, and OpenAI faces a three-year cash burn test
OpenAI’s “buy everything” compute strategy had support from Sarah and the board while ChatGPT looked almost unstoppable. Then growth slowed near the end of last year, and the mood inside the company became less relaxed.
OpenAI had set an internal goal of reaching one billion weekly active ChatGPT users by the end of last year. It has not announced that number. That has made some investors uneasy because the AI boom is already priced like growth will keep coming fast.
See also Pakistan sets an audacious goal for AI in education system
The company also missed its yearly ChatGPT revenue target after Google (GOOGL) Gemini grew strongly late last year and took share from OpenAI. Subscriber cancellations have also been an issue. Earlier this year, OpenAI missed several monthly revenue targets after Anthropic gained ground in coding and enterprise products.
OpenAI recently raised $122 billion, the largest funding round Silicon Valley has seen. That gave the company more cash, but the spending load is still huge. With all the computing power OpenAI has signed up for, the company expects to use that money within three years, even if it hits aggressive sales goals. Some of the funding also depends on partner agreements, so not every dollar is fully locked in with no strings attached.
There are still areas growing inside OpenAI. Codex, its coding tool, is gaining popularity. The company is also cutting costs by scaling back projects such as Sora, its video-generation app. OpenAI has released GPT-5.5, a model that beat several industry benchmarks.
But the stock reaction showed that traders are now watching cash, targets, and compute bills more closely than hype. For Coreweave, Oracle, SoftBank, Broadcom, AMD, and Nvidia, that is the problem.
