AI power shortage is overstated: upgrades and exports shrink the gap, glut by 2028
toptickcrypto · x · 2026-10-20
The market has started pricing a power shortage as a real threat to the AI trade, but the quoted analysis argues the US power gap is far smaller than consensus:
- Accelerator power needs are overstated: much of the install base through 2028 will be upgraded to newer accelerators rather than added on top, and a meaningful share ships outside the US. Normalizing for cost and power, far less new US power is required.
- Equipment glut likely: the commenter adds that power generation supply surges by 2030, and turbines/recips could see a glut by 2028 — power is not the bottleneck.
More from Infra
- Fused linear DPO loss merged into NVIDIA TileGym: 7.2x less memory, 1.27x faster — blelbach · 2026-10-20
- If GitHub goes down, how does GitHub push code to GitHub to bring GitHub back up? — VeryWellVersed · 2026-10-20
- Meta: Co-Design for AMD GPUs Delivers Up to 1.5x Training Throughput vs Direct NVIDIA Ports — burny_tech · 2026-10-20
- Dev rewrites opencode server in Rust: 100MB vs 2.4GB RAM, session search 30,000x faster — microbass · 2026-10-20
- Muon optimizer tweak with a polynomial grid saves ~0.5% of LLM training compute — burny_tech · 2026-10-20
- Tech Fund estimates $414B in annualized AI revenue by year-end, ex-China — FinanceYF5 · 2026-10-20