TRASHFUTURE PREVIEW: I Feel Like Margin Call ft. Ed Zitron
Nov 14, 2025
Ed Zitron, a tech journalist focusing on big tech's business and politics, joins to explore the AI bubble. He dives into OpenAI's financial uncertainties and its quest for government support for data center funding. Zitron highlights the misleading nature of reported revenues versus high inference costs. The conversation also touches on the risks of fragile economic structures and the potential fallout if banks retreat from speculative AI investments. Plus, they kick off with new Epstein revelations, adding an unexpected twist!
AI Snips
Chapters
Transcript
Episode notes
Questionable Public Numbers
- OpenAI publicly presents very large growth numbers that may be exaggerated or misleading.
- Ed Zitron shows discrepancies between reported revenues and likely real figures based on inference costs and Microsoft payments.
Government Backstop Risk
- OpenAI sought government-style backstops to guarantee financing for data center builds.
- That shifts risk from private lenders to public institutions, echoing past financial crises.
Inference Costs Exceed Reported Revenues
- Ed Zitron calculated OpenAI spent $8.67 billion on inference Jan–Sep 2025 and likely earned far less revenue.
- He argues leaked investor documents and press reports have been misinterpreting or inflating their financials.