The Second Derivative: Why No One Understands the AI Boom
10 hours ago
- #Credit Risk
- #Economic Cycles
- #AI Finance
- The article compares the AI boom to the 2008 mortgage crisis, highlighting that both rely on acceleration (second derivative) rather than levels or growth rates.
- It argues that AI financing is structured like a credit-driven real estate cycle, not a technology cycle, with heavy reliance on take-or-pay contracts and debt.
- OpenAI is identified as a key 'subprime' borrower, dependent on continuous equity refinancing to service massive compute commitments without profitability.
- Hyperscalers' capital expenditure is accelerating, but the second derivative has turned negative, signaling potential regime change despite record spending.
- The risk lies in concentrated loan books exposed to cash-burning AI labs, with $2.1 trillion in backlog tied to counterparties like OpenAI and Anthropic.
- A market shift rewarding capex discipline over spending could trigger a reflexive collapse, similar to 2008, due to negative convexity in financing instruments.
- The article emphasizes that deceleration, not decline, is sufficient to break the structure, as seen in housing when price growth slowed.