Ed Zitron's 'AI Is Slowing Down' Argues the Math Doesn't Close — 610 Points on HN

Where's Your Ed At / Hacker News·high signal

In a June 8 essay (610 points, 664 comments on HN), Ed Zitron argues AI needs >$2T in annual revenue by 2030 to justify 190GW of planned data centers costing $9.5–15T, while Anthropic and OpenAI together project only ~$358B revenue by 2029. He cites Anthropic's $375B compute commitment, OpenAI's projected $852B burn through 2030, a -122% Q1 2026 non-GAAP operating margin, and enterprises (Uber, T-Mobile, Brex) capping token spend. The piece is the highest-engagement AI-skeptic argument of the week.