AI Pulse by Inblix

OpenAI: US needs 100 GW/year to beat China in AI

OpenAI Blog · Jul 12, 2026 · 2 min read · Read original article →

Curated by the Inblix editorial team


Featured image for article: OpenAI: US needs 100 GW/year to beat China in AI

OpenAI is framing electricity not as a utility but as a strategic weapon in the AI race, warning the White House that a growing ‘electron gap’ with China threatens American dominance. In a new submission to the Office of Science and Technology Policy, the company argues that US leadership hinges on a massive, urgent buildout of power generation. The numbers they cite are stark: China added 429 gigawatts of new power capacity in 2024—over half the global total—while the US managed just 51 GW, roughly one-eighth of that figure.

To close that gap, OpenAI is calling for a national project to add 100 gigawatts of new energy capacity annually. They’re putting real money behind the rhetoric. The company’s Stargate initiative, with sites in Texas, New Mexico, Ohio, and Wisconsin, is on track to deliver nearly 7 GW of compute capacity and over $400 billion in investment within three years. They claim they’re ahead of schedule to hit their full $500 billion, 10 GW target by the end of 2025.

Beyond the hardware, OpenAI is pushing for regulatory modernization and workforce development. An internal analysis suggests America will need roughly 20% of its current skilled trades workforce over the next five years just to build the necessary data centers and energy infrastructure. To address that, they’re rolling out certification programs and a jobs platform aimed at creating pathways into tech and skilled trades, particularly in communities far from Silicon Valley.

There’s also a softer sell here about grid benefits. OpenAI says its Wisconsin site is working with the local utility to strengthen the grid for residents, and that Stargate designs will contribute energy back to local systems or reduce draw during peak times to avoid hiking consumer costs. Whether that architecture works at scale remains to be seen, but the message is crafted to preempt the obvious criticism that massive data centers will cannibalize power and raise prices for everyone else. The subtext is clear: if the US doesn’t build this infrastructure, China will, and the economic upside—OpenAI claims a trillion dollars in AI infrastructure investment could boost GDP by over 5% in three years—will accrue elsewhere.

💡 Key Takeaways

  1. OpenAI warns of an 'electron gap,' noting China added 429 GW of power capacity in 2024 versus just 51 GW in the US, and is pushing for a 100 GW/year national energy target.
  2. The Stargate project is ahead of schedule to deliver 10 GW of compute capacity by late 2025, backed by over $400 billion in near-term investment across four states.
  3. OpenAI estimates the US will need 20% of its current skilled trades workforce over five years to build AI infrastructure, and is launching certification programs to help fill that gap.
  4. The company is designing its data centers to contribute power back to local grids, an attempt to head off criticism that AI infrastructure will drive up consumer electricity costs.

Keep reading: See related articles below for more coverage on this topic.

Get smarter about AI

The sharpest AI news, curated daily. Delivered free to your inbox.

Learn more

Glossary terms

← Back to all articles