For the past two years, one story has dominated every conversation in finance. Artificial intelligence. Trillions of dollars invested. Nvidia becoming the most valuable company on earth. Every major tech company racing to spend billions on data centres. The narrative has been unstoppable.
Yesterday, the Wall Street Journal dropped a report that put the first serious crack in that story.
OpenAI, the company behind ChatGPT, the most recognisable AI product on the planet, valued at $852 billion has been quietly missing its own revenue and user targets. And its own CFO is worried the company might not be able to pay its bills.
Here’s everything you need to know.
What the Wall Street Journal Actually Reported
OpenAI recently missed its own targets for new users and revenue. CFO Sarah Friar told other company leaders she is worried the company might not be able to pay for future computing contracts if revenue doesn’t grow fast enough. Members of the board have taken a harder look at OpenAI’s data centre agreements and raised doubts about Sam Altman’s drive to acquire still more computing power even as growth slows.
Specifically, OpenAI missed its internal goal of one billion weekly active users and did not meet its target revenue for the year. Growth in ChatGPT slowed toward the end of last year.
OpenAI pushed back hard. Sam Altman and Sarah Friar issued a joint statement calling the report “ridiculous” and saying they are “totally aligned on buying as much compute as we can.” But notably, they didn’t dispute the numbers.
Where the Users Are Going
Here’s the part that explains everything. ChatGPT isn’t just growing slower. It’s losing ground to competitors with real data to back it up.
Twelve months ago ChatGPT held 87% of the AI chatbot market. By January 2026 that number had fallen to 68% – a 19 point decline that Similarweb analysts called the most significant market shift in generative AI history.
Google Gemini has been the biggest beneficiary, surging from 5.4% to 18.2% market share in the same period. Gemini hit 750 million monthly active users, benefiting from being built directly into Android phones, Google Search, and Google Workspace, meaning users don’t even have to choose it, it’s just already there.
Anthropic’s Claude has been the other big winner, particularly in enterprise. Claude more than doubled its daily average users from February to March alone. Website visits jumped from 3.6% to 6.6% of the market. Daily users went from 0.8% to 1.8%.
The enterprise numbers are even more striking. Among companies purchasing AI services for the first time, Claude now wins approximately 70% of head-to-head deals against OpenAI. A year ago, one in 25 businesses paid for Anthropic. By March 2026 that number jumped to nearly one in four.
The reason is product. Anthropic’s Claude Code generated over $2.5 billion in annualised revenue by February 2026, more than doubling since the start of the year. Enterprise users represent more than half of that revenue. Businesses doing serious technical work are choosing Claude for coding. Businesses doing research and document analysis are choosing Claude for its accuracy. And everyone with a Google account already has Gemini in their pocket.
ChatGPT still has 800 million weekly users and is by far the most recognised AI brand in the world. But it’s no longer pulling ahead. It’s defending ground on multiple fronts simultaneously.
Why This Matters So Much
OpenAI’s entire business model is built on one bet: spend enormous amounts of money on computing power now, grow fast enough to justify it later.
OpenAI has a $250 billion commitment on Microsoft’s Azure cloud through 2032. It has newly opened distribution on Amazon Web Services and Google Cloud. These are not small obligations. They’re multi-decade financial commitments that require the company to keep growing at an extraordinary pace just to service them.
If revenue growth slows, the maths gets very uncomfortable very quickly. You can’t keep spending $250 billion on computing power if the product isn’t generating enough revenue to pay for it.
The shortfall comes as OpenAI ramps up spending on data centres and talent to support its most advanced models, raising fresh questions about how quickly that investment can translate into sustainable growth as it approaches its expected market debut.
That market debut is an IPO targeting a $1 trillion valuation. The odds of OpenAI IPOing above $1 trillion in 2026, which started the morning at 59% on Polymarket, dropped dramatically to around 25% as the story spread.
The Market Reaction Was Immediate
When this report hit, it didn’t just affect OpenAI. It sent shockwaves through every company connected to the AI spending story.
Oracle, which has a $300 billion five-year partnership to supply computing power to OpenAI, dropped 4%. Chipmakers including Broadcom and Advanced Micro Devices declined 4% and 3% respectively. Nvidia fell more than 1%.
The logic is straightforward. If OpenAI is struggling to grow fast enough to justify its spending, every company that supplies OpenAI with chips, computing power, and infrastructure is at risk of seeing those contracts reduced or cancelled.
The entire AI investment thesis – ‘spend big on infrastructure now’, just got its biggest public challenge.
Is This Actually a Big Deal or Just Noise?
Here’s where it gets genuinely complicated.
The bulls have a point. OpenAI is still the most used AI product in the world. ChatGPT has hundreds of millions of users. The company just raised $122 billion from some of the most sophisticated investors alive. And some analysts pointed out that missing internal targets in a market moving this fast isn’t necessarily alarming, every company sets ambitious targets.
But the bears have a point too. The AI story has always required a leap of faith — that the enormous amounts being spent on infrastructure would eventually translate into sustainable revenue. That faith has never really been tested before. The competitors gaining ground aren’t small startups. They’re Google with $185 billion in AI capex planned this year and Anthropic growing revenue 10x annually for three consecutive years. These are serious challengers eating into OpenAI’s core markets.
The Musk Factor
There’s one more layer to this story that makes it even messier.
On the same day the WSJ report dropped, Elon Musk took the stand in Oakland at the opening of his trial against OpenAI. Musk testified and argued against what he described as converting a charity into a profit-seeking entity, asking for the unwinding of OpenAI’s for-profit conversion and more than $134 billion in damages.
So in the same week OpenAI is trying to go public at $1 trillion, its CFO is reportedly worried about paying the bills, its biggest competitors are taking significant market share, and its co-founder is suing it for $134 billion in open court.
Not a great week for Sam Altman.
What This Means For the Bigger AI Story
This report matters beyond OpenAI. It’s the first serious public evidence that the AI spending boom might be running ahead of actual revenue generation.
Every major tech company (Microsoft, Google, Amazon, Meta) is spending tens of billions on AI infrastructure. The assumption has always been that demand for AI products would grow fast enough to justify that spending. OpenAI missing its targets, while Gemini and Claude quietly take their market share, raises the question of whether that assumption holds across the whole sector.
The Magnificent Seven earnings this week. Investors will be looking not just at profits but at whether AI revenue is actually growing fast enough to justify the spending. If it isn’t, the entire AI infrastructure trade starts to look different.
The Bottom Line
OpenAI is still the most important AI company in the world. ChatGPT is still the most used AI product on the planet. The company just raised $122 billion from some of the most sophisticated investors alive.
But for the first time, there is credible public evidence that the gap between AI spending and AI revenue is wider than the market thought.
