OpenAI tells Trump to build more power plants or China wins the AI arms race

Steven

OpenAI tells Trump to build more power plants or China wins the AI arms race

Summary

OpenAI has urged the US administration to commit to building 100 gigawatts of new electricity capacity per year to avoid falling behind China in the global AI race. The firm describes electricity as a “strategic asset” — “electrons are the new oil” — and warns of an “electron gap” after China added 429 GW of capacity in 2024 versus 51 GW in the US.

OpenAI says it is working with Wisconsin utilities to expand capacity for its new Stargate datacentre campus and claims its infrastructure plans will help stabilise grids by feeding energy back or cutting draw to protect consumers. The company also forecasts a need for roughly 20% more skilled trade workers over five years to support data-centre and energy builds.

Separately, Google has signed a 25-year power purchase agreement with NextEra to restart Iowa’s Duane Arnold nuclear plant (615 MW), with operations expected by early 2029 — a move positioned as carbon-free supply for cloud and AI workloads and as grid reliability support.

Key Points

  • OpenAI calls for a national project to build 100 GW of new energy capacity per year to secure US AI leadership.
  • China expanded generation by 429 GW in 2024; the US added 51 GW, creating what OpenAI terms an “electron gap.”
  • OpenAI frames electricity as a strategic asset for AI infrastructure — “electrons are the new oil.”
  • The company is coordinating with Wisconsin utilities for its Stargate datacentre campus and says it will both support the grid and limit consumer cost impacts.
  • OpenAI warns of a labour shortfall: about 20% more skilled tradespeople will be needed over five years to build and maintain data and energy facilities.
  • Google and NextEra signed a 25-year PPA to restart Duane Arnold nuclear plant (615 MW), targeting early 2029 start and increased grid reliability.

Context and Relevance

This is where energy policy, national economic strategy and AI infrastructure converge. The demand for large-scale compute is translating directly into national-level grid planning. If you work in cloud, datacentre operations, energy, public policy or industrial strategy, this signals potential regulatory moves, long-term procurement deals, and a need for workforce planning. It also shows big tech pushing for nationwide infrastructure changes to secure competitive advantage.

Author note

Punchy: This isn’t just corporate chest-thumping — it’s a direct ask for national-scale infrastructure investment. OpenAI’s framing turns power capacity into a geopolitical asset, and the comparison with China’s build-rate is a wake-up call for policymakers and industry alike.

Why should I read this?

Short version: if you care about where AI capability lives (and who controls it), read this. It explains why datacentres are now a national infrastructure issue, how that affects bills and jobs, and what big players are already doing about it — including deals to bring nuclear power back into the mix. It’s quick, relevant and a useful heads-up for anyone in tech, energy or government.

Source

Source: https://go.theregister.com/feed/www.theregister.com/2025/10/28/openai_100gw_power_demand/