Premium: AI Is Getting Way Too Expensive
Newsletter · Read
- Argues the AI industry debate over theoretical job displacement distracts from the real economic problem: infrastructure investment scaled far beyond any plausible profitable return.
- Cites industry-wide trailing-twelve-month revenues of roughly $110B against $122B raised by OpenAI alone in March 2026, and a $145B gap between Q1 2026 AI startup funding and total industry revenues.
- Frames the core bind: hyperscalers need to raise compute prices to reach profitability, but their main AI customers can’t absorb higher rates.
- Practical takeaway: the longer infrastructure spend outpaces revenue, the harder the eventual reckoning — a widening gap rather than a stabilizing one. (Premium post; only the preview was accessible.)