There is a special kind of panic that takes hold when the most expensive private company in history starts missing its own targets. On Tuesday, the Wall Street Journal reported that OpenAI fell short on monthly revenue targets through the first quarter of 2026 and missed an internal milestone of one billion weekly active ChatGPT users by year-end 2025. By the close of trading, Oracle had given back 7.7%. CoreWeave was down 7.4%. SoftBank, which has committed roughly $60 billion to OpenAI across multiple rounds, dropped almost 10% in Tokyo overnight.
That is what a $660 billion AI capex thesis looks like when it stubs its toe.
The Friar-Altman Fracture
The most consequential detail in the WSJ story was not the revenue miss. It was the boardroom argument it triggered. CFO Sarah Friar, recruited from Nextdoor in 2024 specifically to bring public-company financial discipline to OpenAI, has reportedly warned colleagues and board members that the company’s spending pace is out of step with its revenue growth, that compute commitments to Oracle and CoreWeave may strain liquidity in 2027, and that OpenAI is not financially ready to IPO on Sam Altman’s preferred timeline.
Altman, the report says, disagrees. The two issued a joint statement Tuesday night calling the report “ridiculous” and insisting OpenAI is “firing on all cylinders.” That is a tell. CFOs and CEOs do not jointly issue defensive statements when they agree on the underlying numbers. They issue them when they need to project unity to the market.
Friar’s position is the more conservative one and, for now, the more defensible one. She knows what every public-company CFO in software knows: revenue forecasts are easy to sign off on when growth is exponential and brutal to defend when growth normalizes. OpenAI has spent the last two years in the exponential phase. The early-2026 numbers suggest the curve is bending.
Where The Growth Went
The Journal’s reporting points to two soft spots. Consumer ChatGPT user growth has stalled as Google’s Gemini integration into Search and Workspace has become genuinely competitive, particularly for free-tier users who never paid OpenAI a dime but counted toward the monetization funnel. Enterprise growth has been pinched by Anthropic’s Claude, which has built a dominant position in coding and increasingly in financial-services workloads, exactly the high-value vertical OpenAI needed to convert.
In other words, OpenAI is being squeezed from two directions: a free-tier hyperscaler with infinite distribution and a premium competitor with a better enterprise reputation. Neither problem is fatal. Both are real.
The Contract Math
The reason the market reacted as violently as it did is the contract structure. OpenAI signed a roughly $300 billion, five-year compute commitment with Oracle in 2025. It has a separate $11.9 billion contract with CoreWeave. SoftBank has committed up to $60 billion in financing across multiple tranches. Amazon Web Services has its own multi-year deal. These are obligations, not options. If revenue does not grow into the cost base, OpenAI either renegotiates with its compute partners on weaker terms, raises another massive private round on punitive valuations, or pulls forward an IPO into a market that may not be ready to absorb it.
That is why Oracle traded down hard. Investors who bought ORCL on the OpenAI thesis are now repricing the durability of those revenue streams. CoreWeave is in the same boat with thinner equity. SoftBank is in the worst position because Masayoshi Son’s exposure is concentrated and his balance sheet is already stretched.
The Bull Case
Veteran tech analyst Mark Mahaney took to Fortune yesterday to argue the market is overreacting, pointing out that OpenAI is still on track for $20 billion or more in 2026 revenue and that internal monthly targets are aspirational by design. He may be right. Internal forecasts inside hyper-growth companies are often set above plan to motivate teams. Missing them by single-digit percentages is normal. Treating the miss as evidence of structural weakness rather than execution noise would be a mistake.
There is also a simpler defense: the underlying compute demand for AI workloads is, by every credible measure, still accelerating. The question is who captures the value, not whether the value exists. OpenAI may end up with a smaller piece of a larger pie. That is a different problem from a shrinking pie.
What To Watch Next
Three things will determine whether yesterday’s selloff was a buying opportunity or the first inning of a real correction.
First, OpenAI’s next pricing move. If the company raises ChatGPT Plus pricing or rolls out a more expensive enterprise tier in May, that is a bullish signal of pricing power. If it cuts prices to defend market share against Gemini, that is bearish.
Second, the Oracle and CoreWeave Q1 calls. Both companies need to address the OpenAI relationship directly and disclose what fraction of guided revenue depends on it. Vague answers will get punished.
Third, an actual financial disclosure from OpenAI. The company has resisted publishing detailed financials because it is private. That stance is incompatible with an IPO. If OpenAI files an S-1 in the second half of the year, the documents will settle the Friar-Altman argument one way or the other. Coverage of the original WSJ reporting and the market fallout is collected at Yahoo Finance.
The Bigger Picture
For two years, AI has been the trade nobody could afford to be underweight. That is changing. OpenAI’s miss is not the end of the cycle, but it is the first material crack in the narrative that AI revenue growth would automatically validate any level of infrastructure spend. CFOs now have permission to ask harder questions. Investors now have permission to demand actual margin disclosures. Enterprise buyers now have permission to negotiate harder.
A more disciplined AI market is not a worse AI market. It is a more sustainable one. But the path from euphoria to discipline runs through stocks like Oracle, CoreWeave, SoftBank, Nvidia, and a dozen others that will spend the next quarter explaining themselves.
The Friar-Altman argument is, in a sense, the argument the entire industry is having out loud. Whoever wins it determines whether AI becomes the most durable platform shift of the decade or the next chapter in the long history of capital cycles. Tonight’s Microsoft, Meta, Alphabet, and Amazon earnings will provide the next data point. Until then, OpenAI’s investors are watching their own company’s revenue figures more carefully than they ever expected to.