Napkin math: even a full NY data center ban slows AI by less than a day
random_walker · x · 2026-08-24
The author does napkin math on what data center bans/moratoria actually do to AI progress. Key idea: inference efficiency at a given capability level keeps improving, so blocked capacity roughly translates into a time delay — how long it takes efficiency gains to catch up to that compute.
Main takeaways:
- A typical U.S. state enacting a 1-year moratorium slows AI efficiency progress by only about 5-10 hours;
- Accounting for highly substitutable data center siting (assuming 90% "leakage" as capacity gets rebuilt elsewhere), even a big state like NY banning all new construction slows AI by less than a day;
- The author admits some numbers may be slightly off (AI-assisted analysis with spot checks), but the order of magnitude holds: data center bans are not an effective way to slow AI.
Caveat: local environmental effects are a legitimate reason for such bans (not discussed here); but to the extent support stems from generalized anxiety about AI, blocking construction is an exceptionally ineffective channel for it.
More from AGI Musings
- Why data companies are printing money: labs pay millions for env-shaped nails — sarahookr · 2026-10-15
- Chollet backs Gakku: conversational AI as the biggest education leap of our generation — fchollet · 2026-10-15
- Subbarao Kambhampati to dissect the 'mythos of LLM thinking tokens' at UMN — rao2z · 2026-10-15
- Pure-play AI firms' combined revenue — $500B or $1T — will decide if AGI takes off this decade — Jsevillamol · 2026-10-15
- AI 2027 timeline gains believers: 2027 likely the most critical year for capability jumps — teortaxesTex · 2026-10-15
- Inside SAIF: a16z's $25M bet on AI rewriting the economics of cyber defense for small towns — a16z · 2026-10-15