Changpeng Zhao, the former Binance chief, said most artificial-intelligence companies are headed for bankruptcy — a stark warning that lands as some of the sector’s biggest names burn through billions with little to show on the productivity front.
Valuation boom, cash-flow bust
Zhao — known in crypto circles as CZ — didn’t name names, but the numbers behind the two most prominent AI labs tell the story. Anthropic just closed a $65 billion Series H round that valued the company at $965 billion, nearly tripling its February valuation of $380 billion. OpenAI sits at an $852 billion valuation after its March mega-funding and is preparing a confidential S-1 filing with Goldman Sachs and Morgan Stanley, targeting a public debut as soon as September.
But those eye-popping numbers mask deep red ink. OpenAI has guided to annual losses through at least 2028, including $74 billion in operating losses in 2028 alone, with $1.4 trillion in datacenter spending over eight years. Anthropic is on track for its first operating profit in the current quarter — a rare bright spot — but together the two companies underwrite more than half of the roughly $2 trillion in future cloud commitments held by Microsoft, Amazon, Google, and Oracle. That’s a lot of server time that needs to turn into revenue.
Uber’s four-month budget blowout
One customer is already questioning the return. Uber burned through its entire 2026 budget for Anthropic’s Claude Code and Cursor developer tool in just four months. That means the ride-hailing giant spent a full year’s AI allocation in a third of the time. Uber’s CTO disclosed the overrun, and the company’s COO asked publicly whether higher token usage actually improves consumer products. The question echoes a broader unease: if the biggest corporate adopters can’t get a clear answer on value, what about everyone else?
Productivity impact: still missing
The skepticism has data behind it. National Bureau of Economic Research data from February 2025 showed 90% of firms reported no measurable AI impact on workplace productivity. That’s a striking disconnect from the hype that has driven valuations into the trillions. CZ’s prediction that most AI companies will go bust isn’t just a contrarian bet — it’s a bet that the gap between spending and results will eventually snap closed.
The sector is growing, sure. But growth without profit, without productivity gains, and with customers slamming the brakes on spending is a recipe for a shakeout. Not every company will disappear — the survivors might be the ones that can show a dollar of output for every dollar of compute.
OpenAI’s S-1 filing — expected to be confidential for now — will force the company to disclose its financials more plainly than ever. Investors will get their first hard look at whether a $852 billion valuation can be supported by a business that doesn’t expect to turn a profit for years. That’s the real test, and it’s coming this fall.
