Binance founder Changpeng Zhao, known as CZ, warned on Friday that most artificial intelligence companies will eventually fail, even as the sector sees unprecedented fundraising. His comments came during a period when two of the largest AI firms, Anthropic and OpenAI, are valued at nearly $1.8 trillion combined.
CZ posted on X that while AI technology itself will continue to grow exponentially, the current number of AI startups is too high. He predicted that even the surviving firms will face major price swings and new competition. “Most AI companies will go bust. There are just too many. Even survivors will see huge price fluctuations. There will be new survivor entrants too,” he wrote. CZ described this as a normal pattern in early-stage industries, where heavy investment typically results in only a few long-term winners.
The warning arrives at a time when capital is flowing heavily into a handful of leading AI labs. Anthropic announced a $65 billion Series H round on Thursday, giving it a post-money valuation of $965 billion, according to CNBC. That round, led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, nearly tripled Anthropic’s valuation from February. The company also reported an annualized revenue run rate of $47 billion, up from $30 billion earlier this year.
OpenAI, meanwhile, is valued at $852 billion after its March mega-funding round. The ChatGPT maker is preparing a confidential S-1 filing with Goldman Sachs and Morgan Stanley, aiming for a public debut as soon as September. Analysts expect its valuation to surpass $1 trillion at that point.
Corporate AI Spending Raises Questions
Despite the optimism around top AI firms, questions remain about profitability. Uber CEO Dara Khosrowshahi told analysts in May that the company was slowing hiring to absorb its AI investments, while struggling to show clear returns. Uber CTO Praveen Neppalli Naga disclosed in April that the firm had exhausted its entire 2026 budget for Anthropic’s Claude Code and Cursor developer tools in just four months. The COO also publicly questioned whether higher AI token usage improved consumer products, saying the connection “is not there yet.”
This pattern appears widespread. National Bureau of Economic Research data from February showed that 90% of firms reported no measurable AI impact on workplace productivity. OpenAI itself has guided to annual losses through at least 2028, including $74 billion in operating losses that year, even while committing $1.4 trillion to datacenter spending over eight years.
Revenue Growth vs. Cost Reality
Concerns have also intensified around hyperscaler revenue loops. Anthropic and OpenAI together underwrite more than half of the roughly $2 trillion in future cloud commitments held by Microsoft, Amazon, Google, and Oracle. Whether CZ’s prediction proves correct depends on how quickly AI revenue catches up with AI costs. Anthropic is on track for its first operating profit this quarter, but most of the sector still spends faster than it earns. The next real test will arrive when OpenAI’s S-1 reveals what a trillion-dollar AI company actually looks like on a balance sheet.
