The Wall Street Journal reported Tuesday that OpenAI has missed its own projections for user growth and revenue, sending AI infrastructure stocks into a selloff. Oracle dropped more than 4%. Nvidia, Broadcom, and AMD each fell 3-4%. CoreWeave slid over 5%. SoftBank, one of OpenAI's biggest backers, plunged nearly 10% in Asian trading.
The core concern: OpenAI's massive financial commitments to computing capacity. CFO Sarah Friar reportedly warned colleagues that if revenue doesn't pick up, the company could struggle to fund future compute deals. These aren't optional expenses. OpenAI uses "take-or-pay" contracts that obligate it to pay for computing capacity whether it uses it or not. The company's $300 billion partnership with Oracle and its multi-billion commitments to CoreWeave are essentially fixed costs that don't care about revenue dips. OpenAI pushed back hard, telling CNBC the report was "ridiculous" and that they remain "totally aligned on buying as much compute as we can."
Some analysts think the market overreacted. Mizuho's Jordan Klein questioned how much fundamentals could have shifted in under 30 days, given OpenAI just closed a $122 billion funding round in late March at an $852 billion valuation. John Belton at Gabelli Funds saw the report as confirmation of what investors already suspected: OpenAI is losing enterprise market share to Anthropic and Google's Gemini. "This looks more like confirmation about OpenAI's recent market share trends," Belton said, not a sector-wide spending problem.
The competitive landscape is getting real. Anthropic keeps winning corporate customers. Google's Gemini models are gaining traction. Companies are spreading their AI spending across multiple providers rather than betting everything on OpenAI. Investors are fleeing OpenAI for Anthropic, which is probably healthy for the market. But it means OpenAI's path to justifying an $852 billion valuation just got harder, and the infrastructure companies that built their growth projections on OpenAI's spending are paying attention.