This Is The Problem With AI's Most Bullish Chart
In the wake of leadership turmoil at Nvidia and the Anthropic IPO, Liquid Compute chief product officer Stanley Lee and SK Ventures managing partner Paul Kedrosky unpack the hidden fragilities in the AI ecosystem. They explain why rising GPU rental indices reflect tightening credit conditions and venture-funded capital costs rather than genuine end-user demand. The conversation also explores the unsustainable project finance math behind multi-trillion-dollar CapEx forecasts, Nvidia's role as a fractional-reserve lender, and Anthropic's pivot into vertical enterprise software.
- Mid-market neoclouds, often former crypto miners, bought GPU fleets using 70% to 80% loan-to-value debt packages while committing zero cash equity of their own. Read →
- Stanley Lee points out that Nvidia actively finances both sides of the market, injecting equity into AI labs while offering balance sheet backstops to high-yield credit lenders. Read →
- Software development is the only sector displaying genuine labor substitution from large language models today. Read →
- GPU rental indices keep climbing, but that price rise reflects tightening credit conditions rather than surging software demand. Read →