Ed Zitron has long been one of the most vocal critics of the AI boom and has built a large following in the technology industry. This time, however, he takes his argument further than ever before, claiming that the real AI bubble is, in essence, an “OpenAI bubble.” If OpenAI ultimately fails, he argues, it could become the “Lehman Brothers” of the AI era, shattering not only the investment thesis behind AI but also triggering a broad repricing of data centers, AI infrastructure, and global technology stocks.

What did Ed Zitron say?

His argument begins with a simple premise: since ChatGPT launched at the end of 2022, OpenAI has effectively become the “credit anchor” of the entire generative AI era. Investors’ willingness to believe that AI will transform the world, that hyperscale data centers are worth building, that GPU demand will continue growing rapidly, and that large language model companies will eventually become profitable all depend on one core assumption—OpenAI’s continued rapid growth. OpenAI has not only defined the AI boom but also shaped the valuation framework for the entire AI ecosystem in capital markets.

In his own words: “The only reason the AI bubble has lasted this long is because OpenAI hasn’t collapsed yet.” Once that core assumption is broken, he believes the consequences will extend far beyond a single company. OpenAI has effectively become a “systemically important institution,” with its continued existence serving as the credit foundation of today’s AI investment cycle.

Ed Zitron’s warning is not without supporting evidence. He says he previously obtained and verified OpenAI’s audited financial statements:

Full-year 2025: Revenue of $13.07 billion, total expenses of $34 billion, operating losses of approximately $20.9 billion, and a net loss of $38.5 billion (the difference largely attributable to non-cash accounting items).

First quarter of 2026: Revenue reached $5.7 billion, while cash burn totaled $3.7 billion. Both figures nearly tripled from the same period a year earlier. Operating losses came to $9.3 billion. By the end of the quarter, OpenAI held more than $73 billion in cash and marketable securities.

Full-year forecast: OpenAI projects cash burn of $25 billion in 2026, rising further to $57 billion in 2027.

Off-balance-sheet commitments: As of the end of 2025, OpenAI’s commitments to cloud providers for computing capacity were estimated at $665 billion, with contracts extending through 2030. Most of these commitments do not appear on the balance sheet, but the payments must still be made regardless of future AI demand. Based on these figures, Zitron estimates that OpenAI could consume more than $852 billion in cumulative funding by the end of 2030, with spending on computing and infrastructure alone potentially exceeding $50 billion in 2026.

Based on this analysis, he concludes that OpenAI’s downfall would be driven by a combination of unprofitable subscription economics, limited advertising monetization potential, and ever-rising computing costs.

If OpenAI collapses, who falls with it?

Ed Zitron specifically singled out Oracle, arguing that it could be one of the companies most exposed to a potential OpenAI crisis.

Why Oracle? Because major technology companies and cloud providers have invested heavily in AI data centers, financing their expansion through debt, leases, and long-term infrastructure agreements. If OpenAI were to significantly reduce its demand for computing power, Oracle could be left with “billions of dollars in wasted capital expenditure, alongside endless debt and lease obligations.”

The chain reaction he envisions is as follows: OpenAI cuts computing demand → data centers become underutilized and capital spending is wasted → debt pressure on related companies increases → chipmakers, cloud computing firms, data center operators, power companies, and financial institutions holding related debt all come under pressure → investment across the AI ecosystem contracts → markets undergo a “violent and painful” repricing.

Ed Zitron predicts that OpenAI’s collapse would only occur after “AI data center debt and venture capital funding have been almost completely exhausted.” He also believes OpenAI will require at least three more large funding rounds over the next decade. Describing the potential scale of the fallout, Zitron writes: “It scares me enough that I almost hope I’m wrong.”