Tom Bilyeu reacts to comments by researcher Ed Zitron warning that parts of the AI investment boom may be masking concentrated financial risk. Zitron cites bank research he says shows a growing share of hyperscalers’ cloud revenue tied to OpenAI and Anthropic, arguing that markets may be treating AI demand as broad-based when it is heavily dependent on a small number of cash-burning customers. He also claims OpenAI materially contributed to Microsoft’s Intelligent Cloud growth in 2025.
Zitron describes OpenAI as “burning cash”, alleging a $20.9bn loss in 2025 and asserting that more than $800m of revenue came from SoftBank for a “Crystal Intelligence” programme he says shows little evidence of real-world activity. He further argues that the commercial structure between hyperscalers and the leading AI labs can involve circular incentives, likening the ‘smart people’ defence to the Enron-era “smartest guys in the room” dynamic.
Bilyeu adds his own interpretation of how high-risk AI-related lending could become systemic, outlining mechanisms he says move exposure from banks into markets via private credit and structured products, and repeatedly frames the core question as whether AI revenues arrive before debt burdens bite. The episode mixes financial literacy explainer segments with allegations about AI-company sustainability and the political risk of treating advanced AI as too strategically important to fail.
Ed Zitron claims that, based on his reporting on OpenAI’s finances, OpenAI accounted for 69% of the year-on-year growth in Microsoft’s Intelligent Cloud segment in 2025, and that excluding OpenAI the segment would have grown 8%. He frames this as evidence that parts of the AI revenue story are more dependent on a single counterparty than widely understood.
Read this moment →Ed Zitron, citing his prior financial reporting on OpenAI, claims the company lost $20.9bn in 2025. He also alleges that more than $800m of OpenAI revenue came from SoftBank for a “Crystal Intelligence” programme, and says he cannot find evidence the programme produced anything substantive.
Read this moment →Ed Zitron describes what he sees as circular commercial relationships between hyperscalers and leading AI labs, including Google selling and renting back infrastructure to Anthropic. He argues that the model looks attractive until the labs’ underlying economics fail, adding that the ‘smart people’ defence reminds him of the “smartest guys in the room” Enron narrative.
Read this moment →Researcher Ed Zitron says UBS estimates OpenAI and Anthropic will account for 27% of Google Cloud revenue this year and more than 48% next year. He argues investors in Microsoft, Google and Amazon should scrutinise whether AI-related cloud demand is as diversified as markets assume, or concentrated in a small number of cash-burning customers.
Read this moment →