‘Lehman Brothers of AI’: Why One Critic Thinks OpenAI Will Crash the Entire AI Industry

By Danielle Liverance Published Jul 17, 6:17AM EDT

Quick Read

Oracle is down 35% this year carrying $638 billion in performance obligations chained to OpenAI, while NVIDIA holds $119 billion in supply commitments.
OpenAI posted $13 billion in revenue against $34 billion in costs in 2025, with a $40 billion SoftBank loan coming due March 2027.
Don't wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now.

‘Lehman Brothers of AI’: Why One Critic Thinks OpenAI Will Crash the Entire AI Industry — © Martin Barraud / Getty Images

Ed Zitron has a warning, and he is delivering it bluntly. In a post published July 15, 2026, the independent tech journalist argued that OpenAI is “the Lehman Brothers of AI,” and that its eventual failure would be “a watershed moment, the Lehman Brothers of the AI bubble, and an event that would define the end of one epoch, the start of another.” He goes further: “I believe that once OpenAI collapses it’ll have a violent, punishing effect on the entire stock market, a precursor to a much greater drawdown as everybody accepts that the AI bubble has burst.” This claim deserves weight because it is anchored to verified numbers.

Why Zitron’s View Carries Weight

Zitron and the Financial Times independently verified OpenAI’s audited 2025 financial statements. That grounding separates this from typical bubble commentary. When he calls AI infrastructure spending “the greatest capital misallocation in history,” a phrase also used by AI researcher Gary Marcus, he points at figures now on the record. His framing traces to a December 2024 essay titled “OpenAI is Lehman Brothers,” comparing AI funding behavior to pre-2008 mortgage-market dynamics, citing global venture funding of $28 billion in November 2024, more than half going to AI, and more than $200 billion in hyperscaler generative AI spend.

The Numbers

Per Zitron and the FT, OpenAI posted $13.07 billion in revenue in 2025 against $34 billion in total costs and expenses. That gap, roughly $21 billion, is the company’s operating loss, driven primarily by R&D and marketing. The headline figure investors encounter is a $38.53 billion net loss. That includes a large one-time, non-cash charge, reported somewhere between $30 billion and $41.6 billion depending on the source, tied to OpenAI’s 2025 conversion from nonprofit to for-profit structure. Conflating the two overstates physical cash burn. Even on the operating figure, the slope is steep: in 2024, OpenAI lost $5.09 billion on $3.7 billion in revenue.

The Contagion Mechanism

Zitron’s thesis is that OpenAI is load-bearing for the entire AI economy. If OpenAI stopped paying partners like CoreWeave and Oracle (NYSE:ORCL | ORCL Price Prediction), those companies could struggle to service their own debt. Oracle’s exposure is documented: Q4 FY2026 Remaining Performance Obligations reached $638 billion, up 363% YoY, with reported free cash flow of negative $23.69 billion against capital expenditures of $55.66 billion. Oracle plans to raise roughly $40 billion in FY2027 through debt and equity. The stock is down 35.29% year to date as markets reprice that concentration. Oracle’s Q4 FY26 8-K lays out the buildout in detail.

NVIDIA (NASDAQ:NVDA) sits adjacent as the chip supplier, with $119 billion in total supply-related commitments and a disclosed partnership to deploy at least 10 gigawatts of NVIDIA systems for OpenAI. NVIDIA has held up better, gaining 11.34% year to date.

The IPO Wrinkle

OpenAI confidentially filed IPO paperwork with the SEC in June 2026, following a $122 billion financing round in March 2026 that valued the company near $850 billion. Sam Altman has reportedly called any valuation below $1 trillion a “nonstarter,” with recent reporting pointing to a target between $1.75 trillion and $2 trillion. An IPO once eyed for late 2026 now appears more likely to slip to 2027. Altman said, “We have not decided on timing yet.” SoftBank has a $40 billion loan coming due in March 2027, per Motley Fool reporting.

The Other Side

This is contested opinion. Zitron argues rather than predicts timing, framing collapse as eventual. His December 2024 “Lehman Brothers” call has not come to pass more than a year and a half later. OpenAI’s revenue, while generating enormous losses, is still growing fast, from $3.7 billion to $13.07 billion in a single year. A bear early for eighteen months may eventually be right, or wrong about a company scaling into its costs.

The Investor Takeaway

The useful way to read Zitron is narrow. His argument is that a handful of companies carry concentrated financial exposure to OpenAI. Oracle and CoreWeave are clearest as infrastructure partners whose debt service leans on OpenAI’s payments. NVIDIA sits adjacent as chip supplier. SoftBank is exposed as investor. Whether or not you buy the Lehman comparison, knowing which balance sheets are chained together matters for a portfolio.

Contact [email protected] for any questions or corrections.

Follow 24/7 Wall St. on Google

About the Author Danielle Liverance →

I've spent more than 15 years inside enterprise software, working alongside the finance, sales operations, and HR leaders who run the revenue engines at some of the largest tech companies in the country.

My day job is helping enterprise executives make smarter decisions about retention, compensation, and growth. These are the same operational levers that show up in every earnings report investors actually read. That perspective shapes my writing for 24/7 Wall St.

The headline numbers are easy. The interesting stuff is underneath: how companies make money, what executives are worried about, and what any of it means for the person checking their 401(k) on a Sunday afternoon. I write about personal finance and business as someone who has spent her career inside the rooms where these decisions get made.