Don’t Look Up

Source: Where’s Your Ed At Date: 2026-08-11 URL: https://www.wheresyoured.at/dont-look-up/

Summary

Ed Zitron argues the AI industry’s economics don’t close: OpenAI and Anthropic together lost tens of billions in 2026 while reportedly accounting for 70%+ of Microsoft’s, Google’s, and Amazon’s AI-attributed revenue, financed through circular VC-to-hyperscaler capital flows (citing Wells Fargo/Barclays/UBS analyst estimates and Nvidia’s uncommitted “$500B MOU”). He frames non-lab compute demand as still low-single-digit billions against $197B+ in 2027 hyperscaler cloud spend projections from the two labs alone.

Implications

Feeds the capability-clocks thread as a standing counter-narrative to release-cadence framing: if the compute spend underwriting weekly model releases isn’t revenue-backed, the “burst vs. exhale” reading of lab activity may track financing runway more than product-market pull. Worth revisiting against hyperscaler capex guidance in upcoming earnings.