The Wire
Dispatch
Latest Daily
OpenAI Just Told Wall Street the AI Machine Has Limits
Missed revenue targets, a worried CFO, and a trillion-dollar IPO on the line. The AI spending cycle just got its first real stress test.

What Happened
The Wall Street Journal dropped a bomb on Tuesday. OpenAI has been missing its own revenue and user growth targets. Not by a little. Enough that CFO Sarah Friar has been warning internally that the company might not be able to pay for its computing contracts if things don't turn around.
That's a wild sentence to read about a company preparing for an IPO at roughly a trillion-dollar valuation.
The numbers paint a clear picture. OpenAI hit around $25 billion in annualized revenue by early 2026. Sounds massive. But they had an internal target to reach 1 billion weekly active users and didn't get there. ChatGPT growth slowed in late 2025. Subscribers started canceling. And competitors ate into their lead. Anthropic grabbed enterprise and coding market share. Google's Gemini kept pushing. The moat that looked unbreakable a year ago is leaking.
The market reaction was immediate. Oracle dropped 7% in premarket. The company has a $300 billion five-year computing deal with OpenAI, making it the most exposed name in the entire AI infrastructure chain. CoreWeave tumbled. SoftBank fell almost 10% in Tokyo. Even Broadcom and AMD took 3 to 4% hits. This wasn't just an OpenAI story. It rippled across every company that bet on AI spending only going up.
And the spending is staggering. OpenAI projected $17 billion in cash burn for 2026. They don't expect to be cash-flow positive until 2030. They've committed to over $500 billion in cloud capacity across multiple providers. That kind of spending only works if revenue keeps compounding. When it doesn't, the whole thesis wobbles.
One analyst at Gabelli Funds framed it well. This isn't necessarily a crisis for AI spending broadly. It might just be OpenAI losing share to Anthropic and Gemini. The pie is still growing. But the slice OpenAI gets is shrinking faster than their cost base.
For investors, the read-through is simple. The AI capex cycle is not dead. But the idea that every dollar poured into AI infrastructure automatically turns into revenue just took a serious hit. If OpenAI, the company that started this whole wave, can't grow fast enough to cover its compute bills, what does that say about the dozens of smaller players making the same bet?
The IPO is still coming. The S-1 will clean this up nicely. But the market now knows something it was happy to ignore six months ago. Revenue growth in AI is not guaranteed. And when it slows, the infrastructure stocks feel it first.
Market analysis and trade ideas. Have a ticker in mind? Send us your request.










