OpenAI Misses Revenue, User Targets As CFO Fears $1.5 Trillion In Commitments Can't Be Paid
Earlier today, when previewing this week's earnings by the Mag 7 which account for over $10 trillion in market cap set to report Q1 results after the close on Wednesday, Goldman's Delta-One head Rich Privorotsky said that "Equities are being driven by one thing…AI spend", and warned that "it's hard not to respect the strength of the AI bid, but the velocity has been extreme. The upside surprise vs expectations has almost entirely come from AI spend…it’s the whole game."
Not only is the whole game, it is the one thing that has prevented the market from collapsing into the Iran war's stagflationary black hole, with "oil/product prices is sucking the oxygen out of the room...Europe underperforming, dispersion extreme."
But none of that matters as long as capex recipients, i.e., chip and semi stocks, keep surging on hopes and expectations that the LLMs and hyperscalers will keep pumping them full of cash day after day, for the unforeseeable future, which they have so far: recall that at the end of Q4, full-year capex estimates soared to a mindblowing $740 billion among just 6 hyperscalers (a number which is expected to rise to almost $1 trillion in 2027).
And at top of this trickle-down monetary waterfall is none other than Sam Altman's OpenAi, generously peeing money into the overeager mouths of hyperscalers around the globe, having built up staggering purchase commitments to the tune of $1.5 trillion because there will never be enough compute.
Maybe Sam's right: perhaps there truly is an insatiable need for compute (unless of course one uses Chinese LLMs and/or RAM chips, both of which have a fraction of the hardware demands of the latest and greatest US technology).
The problem arises when one asks if OpenAI will ever be enough revenue to satisfy these astronomic commitments.
For much of the past year, that has been the core thesis behind countless AI bear cases: now that even Michael Hartnett openly calls tech a "bubble", the question is not if but when, to which the bulls have calmly countered that as long as the drunken-sailor at the helm of OpenAi keeps spending at the rate he has been, the "when" isn't coming any time soon.
It now appears, however, that the "when" may have come much sooner than most thought.
According to the WSJ, OpenAI has recently missed its own targets for both new users and revenue, stumbles that have raised concern among some company leaders about whether it will be able to support its massive spending on data centers.
One of them is the company's finance chief: CFO Sarah Friar told other company leaders that she is worried the company might not be able to pay for future computing contracts if revenue doesn’t grow fast enough. In other words, that $1.5 trillion OpenAI had pledged to spend on various data centers, GPUs and memory chips... you can kiss all that goodbye.
Of course, none of this will come as a surprise to anyone familiar with Sam's mercurial style of capital allocation. As a reminder, when OpenAi made its $1.5 trillion flurry of deal announcements last fall, a few things were missing, among them how it plans to fund them, details of the bulk of the financial terms, and any mention of who was providing independent, clear-eyed advice on these complex mega transactions. The reason for that, as the FT reported at the time, is OpenAI still doesn’t know exactly how it will fund them, the terms mostly don’t exist, and advisers were overwhelmingly shunned.
In fact, we learned last October, Sam Altman came up with the “bold vision” himself and leaned heavily on a small number of lieutenants to flesh out the details and push the deals through with little involvement of bankers or lawyers.
One of the brilliant side quests completed by Altman during this period of epic obfuscation (and unprecedented wealth generation by Sam for himself from a "non-profit" thanks to nothing more than promises) was unleashing the AI circle jerk, pardon, circular financing concept, where one company would "invest" in its customer, only to see that money flow back to its through the income statement but not before lifting its PE by several turns; this process would be repeated countless of times lifting all AI valuations substantially even if no actual revenue or cash flow was created. Eventually, virtually every company in the AI sector was wrapped up in such circular structures that tied together suppliers, investors and customers (see "The Stunning Math Behind The AI Vendor Financing "Circle Jerk".")
Yet promises (and lies) can only go so far, and even the loftiest of grand schemes are eventually brought to the ground when the revenue fails to materialize. As it has for OpenAi.
As a result, the company's board of directors have started to closely examine the company’s data-center deals in recent months and questioned Sam Altman’s efforts to secure even more computing power despite the business slowdown, the WSJ reported.
The spending scrutiny is constraining Altman’s once-boundless ambitions ahead of a potential IPO that could take place by the end of the year (he desperately wants to go public before his former employee and arch nemesis, Dario Amodei takes Anthropic public).
Friar and other executives are now seeking to control costs and instill more discipline in the business, at times putting them at odds with their CEO; this may very well mean that the money spigot that has pumped hundreds of billions in capex promises is about to be shut as well, leaving the entire AI ecosystem in a Wile E Coyote moment, suspended in the air off the cliff, just before gravity kicks in.
In a desperate attempt to keep reality as far away as possible, the two heads of OpenAI had no choice but to deny there was any trouble in paradAIs: “We are totally aligned on buying as much compute as we can and working hard on it together every day,” Altman and Friar said in a joint statement. Any suggestion that the pair are divided or pulling back on securing new computing resources is “ridiculous,” they said.
Well, of course they would: the alternative would be an immediate collapse of OpenAI's valuation as revenue growth suddenly collapses, and takes the entire AI bubble with it.
Still, with OpenAI having difficulty to generate even 2% of its spending commitments in the form of revenue (ignoring that the company will likely never be profitable), denials may be all OpenAI has left.
For years, Altman has sought to lock up as much data-center capacity as possible, arguing that computing shortages were the biggest constraint to OpenAI’s growth. As noted above, Sam went on a "dealmaking" spree last year that put OpenAI on the hook for some $1.5 trillion in future spending commitments, and tied much of the tech sector’s success to OpenAI’s.
In other words, if OpenAI goes down, it will take the entire AI sector with it. And since AI is now 40% of the S&P500... you get the picture (if you don't, reread the comments above from Goldman's Delta One head).
Not that anyone can blame Sam for thinking he would get away with it: for a long time, he did. His “buy everything” computing strategy was buoyed by ChatGPT’s seemingly invincible success, and had the support of both Friar and the board. But the chatbot’s growth slowed toward the end of last year, especially as Claude starting stealing clients, sowing fresh doubt among company leaders about the approach.
What followed next was the first domino to fall: OpenAI missed an internal goal of reaching one billion weekly active users for ChatGPT by the end of last year, according to people familiar with the goals. The company still hasn’t announced that milestone, unnerving some investors the WSJ reports. It also missed its yearly revenue target for ChatGPT as well after Google’s Gemini saw massive growth late last year and ate into OpenAI’s market share. Worst of all, for the industry where there are still almost no switching costs, the company has also struggled with defection rates among subscribers, according to WSJ sources.
Things went from bad to worse in 2026 when OpenAI missed multiple monthly revenue targets earlier this year after losing ground to Anthropic in the coding and enterprise markets, people familiar with its finances said.
OpenAI recently raised $122 billion in what was the largest funding round in Silicon Valley history, putting it on more solid financial footing. But to get there, the company signed up for so much computing power that it expects to burn through that amount in the next three years, and that's assuming that it meets ambitious revenue targets. Some of the funding is also conditional and depends on specific agreements with partners (and may explain why Microsoft, which knows the company's business best of all, dramatically revised its agreement with OpenAI earlier today).
To streamline costs, OpenAi recently cut non-core projects such as its video-generation app Sora. OpenAI also recently released GPT-5.5, a powerful model that topped a number of industry benchmarks. Then again, in an industry where the frontier jumps every 2-3 months, the latest model will be obsolete by July.
Meanwhile, a blowback from within the user base is emerging: a number of AI companies including Anthropic have faced a capacity crunch for computing in recent weeks, leading to price increases for access to AI processors, outages and rationing. The challenges have rankled power users of AI products, especially coders who have grown frustrated when AI systems have been unable to finish tasks in a way they had come to expect from past use.
In a recent memo to investors, OpenAI said that it has been able to secure more computing capacity than Anthropic, giving it an advantage in reaching users. The memo, which was viewed by The Wall Street Journal, also addressed Anthropic CEO Dario Amodei’s veiled criticism of OpenAI at a recent business conference, when he said some companies had pulled “the risk dial too far” on data-center spending.
“In hindsight, that caution looks less like discipline and more like underestimating how fast demand would arrive,” the OpenAI memo said.
It would be extremely ironic is Anthropic's "caution" proves correct in the end, and OpenAI is forced to cancel its contracts as it simply does not have the money (but not before Masa Son implodes).
In recent months, Friar has also expressed reservations about OpenAI’s plans to go public by the end of this year, according to people familiar with the matter. She has emphasized to executives and board directors the need for OpenAI to improve its internal controls, cautioning that the company isn’t yet ready to meet the rigorous reporting standards required of a public company. Altman, who has favored a more aggressive timeline for an IPO.
OpenAI has to work through a slate of other issues ahead of a public listing. The company is currently experiencing a leadership vacuum after its second-in-command, Fidji Simo, unexpectedly took medical leave earlier this month.
But the knockout blow for OpenAi could, ironically, come from the person who funded the company in the first place back when it was still an "Open" non-profit. Court proceedings began today in a lawsuit by Elon Musk in which he is seeking to oust Altman and unwind OpenAI’s conversion into a for-profit company. Should Musk prevail, OpenAI may or may not survive, but Sam Altman will have no choice but to move on to his next scam.
Scam Altman has a incredible track record for being a con artist I don't think anyone has a "former ally turned enemy" list this big with directly with people he worked with
A massive new 18-month investigation dropped, revealing the full list of people who worked directly with… pic.twitter.com/1aOkUEsgkq