OpenAI missed key internal revenue and user targets last year and into early 2026, according to a Wall Street Journal report that hit on April 28 and continued to ripple through markets on April 29, dragging down Oracle, Nvidia, Broadcom and AMD as investors questioned whether the AI infrastructure boom is outrunning its main customer.
What the WSJ reported
The Journal, citing people familiar with the matter, said OpenAI missed an internal goal of reaching one billion weekly active users for ChatGPT by year-end. The company has also missed multiple monthly revenue targets for ChatGPT this year, with the report citing rising subscriber defections and ground lost to Anthropic in coding and enterprise markets.
CFO Sarah Friar has reportedly warned other leaders that, if revenue does not accelerate, OpenAI could struggle to pay for the compute contracts it has been signing at a record clip. Board members are said to have begun pressing harder on the company's data-center commitments and on Sam Altman's appetite for more capacity even as growth slows.
Stock selloff in the AI complex
The report triggered an immediate repricing across companies most exposed to OpenAI's spending plans:
- Oracle fell roughly 4%, with attention on its multi-year, approximately $300 billion compute partnership announced last year.
- Broadcom dropped about 4%, AMD around 5% and Nvidia more than 3%.
The selloff bled into the broader semiconductor and data-center supply chain, with investors weighing whether OpenAI's revenue base can support the gigawatt-scale capacity it has reserved with Oracle, Microsoft, AWS, Google and CoreWeave.
OpenAI pushes back
In a joint statement to the Journal, Altman and Friar called the framing "ridiculous," adding that the company is "totally aligned on buying as much compute as we can and working hard on it together every day." Oracle separately defended OpenAI's trajectory, with a spokesperson saying it is "incredibly excited" about the partnership and "focused on building and delivering the capacity they need to support rapidly growing demand."
Why it matters
The story lands at a sensitive moment. OpenAI just closed a $122 billion funding round at an $852 billion post-money valuation and is preparing for an IPO, while its compute deals — with Oracle, AWS, CoreWeave and others — collectively run into the hundreds of billions of dollars. Each of those commitments is built on the assumption that ChatGPT and OpenAI's enterprise business will keep compounding.
For the four hyperscalers reporting earnings this week, the question now bites harder. Microsoft, Alphabet, Meta and Amazon are guiding to a combined ~$700 billion in AI-related capex this year on the bet that demand from frontier labs and their own customers will absorb every GPU they can buy. A confirmed slowdown at OpenAI — the single largest external workload in the AI economy — would be the first concrete data point suggesting that wall of spending may not all clear.
It is too early to call a top in AI infrastructure. But April 29 is the day the market started pricing in the possibility.
