Summary

Zitron argues that rising 10-year Treasury yields (now above 5.3%) are dramatically increasing the cost of all debt, with AI data center borrowers facing a compounding crisis of higher base rates plus deteriorating sentiment. The core claim is that OpenAI and Anthropic, with deeply negative EBITDA margins (OpenAI at -160%), cannot qualify for investment-grade credit ratings and would therefore need to raise $50B+ annually as junk-rated issuers — a structurally near-impossible task. He details how Oracle's $18B data center debt is already trading at distressed yields, lenders like SMBC and MUFG are becoming 'more selective,' and the entire AI demand stack is circular: OpenAI and Anthropic represent 80%+ of compute demand, funded by debt-laden hyperscalers, whose customers are mostly unprofitable AI startups. Zitron concludes that a burst of this bubble implies catastrophic earnings misses at Google, Microsoft, and Amazon, NVIDIA revenue drops of up to 90%, and the collapse of every neocloud — with global financial stakes that nobody in the industry is seriously modeling.

Key Insight

OpenAI and Anthropic's deeply negative EBITDA and lack of hard assets make raising the $50B+/year they need post-IPO structurally impossible at anything above junk rates, and because they constitute 80%+ of AI compute demand, their inability to meet $1.3 trillion in compute commitments would trigger cascading failures across every hyperscaler, Oracle, Broadcom, and NVIDIA simultaneously.

Spicy Quotes (click to share)

• 9

  Anthropic and OpenAI do not resemble the financial condition of any company I could find in the history of the stock market outside of WeWork.

• 7

  You can jingle the keys of 'annualized run rates' and 'adjusted operating income' in the faces of venture capitalists and journalists as much as you want, but investors require actual cashflows, even within a frothy market.

• 6

  The problem Oracle also faces is that things don't have to collapse for a collapse to occur.

• 7

  I only have to be half right for things to be extremely bad.

• 8

  Hyperscalers have no other hypergrowth ideas left — no new Google Search, Microsoft 365, or Facebook — to sell investors, and in pursuit of AI have become the most asset-burdened companies on the Fortune 500, except instead of oil and diverse stocks they have GPUs that only retain value during a hype cycle.

• 9

  The fact we can't cleanly answer this question as hundreds of billions of dollars get sunk into AI data centers may be the most glaring miss in the history of finance.

• 9

  You can truly, madly love LLMs, you can name your dog Dario and your guinea pig Sam, I don't care, but please, I'm begging you, stop making assumptions based on the best-case scenario, and start taking this seriously, because the consequences of me being right have global stakes.

• 7

  None of these statements actually answers a single question about 'how the fuck does all of this actually work?'

Tone

alarmed, combative, analytically rigorous