The US needs energy to maintain AI leadership | Keryc
The United States leads the development of artificial intelligence, and that edge isn’t just technical: it depends on something much more prosaic and tangible that you might not think about every day: electricity. Can you imagine the engine that drives the next tech revolution running out of fuel? That’s the warning coming from OpenAI.
Why electricity is the new strategic asset
AI consumes energy at scale. It’s not just another utility: it’s a strategic asset to build the infrastructure that will sustain economic competitiveness and national security. According to an internal OpenAI analysis, the first $1 trillion invested in AI infrastructure could generate more than an additional 5% of GDP growth over a three-year period.
Sounds exaggerated? Think of electricity as the highway where AI models travel: without more lanes or better pavement, traffic grinds to a halt. And when we talk about traffic, we mean hundreds of millions of users accessing free and paid tools every week.
Electrons are the new oil: without more electrical capacity, the competitive advantage is at risk.
What OpenAI asks for and why it matters
OpenAI presented recommendations to the White House Office of Science and Technology Policy with concrete measures:
Strengthen the US industrial base.
Modernize regulations to unlock more energy.
Prepare workers with education and workforce development programs in AI.
Ensure frontier AI systems protect national security and expand their adoption across the federal government.
These proposals aren’t technocratic exercises: they’re policies so society gets clear benefits from AI — well‑paid jobs, local supply chains, and a more efficient use of resources.
The electron gap and the race with China
The gap is fast and huge. In 2024 China added 429 gigawatts of new electrical capacity — more than half of global growth. The United States added just 51 gigawatts. That gap — the 'electron gap' — can translate into a loss of technological leadership if nothing changes.
OpenAI proposes an ambitious project: build 100 gigawatts per year alongside the private sector. It sounds like a large-scale national strategy, comparable in ambition to historical projects like the interstate highway system or the space race.
What OpenAI is doing here and now
OpenAI isn’t stopping at recommendations. It’s building Stargate sites in Texas, New Mexico, Ohio, and Wisconsin, which will add nearly 7 GW of compute capacity and promise more than $400 billion in investment over the next three years. The announced goal is to fulfill a pledge of $500 billion and 10 GW by the end of 2025.
In Wisconsin, for example, OpenAI is working with the local utility to increase energy capacity in ways that also benefit the community: reducing strain on the grid, returning power, or avoiding higher costs for residents. That’s a concrete example of how these projects can affect your electric bill and local services.
Jobs and training: the other critical piece
Building data centers and transmission lines isn’t enough. You need people to install and operate them. OpenAI estimates that its US infrastructure plans will require around 20% of the current skilled trades workforce over the next five years.
That’s why they’re launching certification programs and a jobs platform to create pathways into well‑paid technical trades with transferable credentials. The idea is to bring economic opportunities to places beyond Silicon Valley and unblock labor bottlenecks that could slow expansion.
What can the United States do now?
The proposals are clear and practical: invest in energy, modernize rules to speed up projects, train and certify workers, and use public procurement to drive adoption that ensures both innovation and national security.
This isn’t science fiction. It’s industrial policy, investment, and large-scale training. If the United States thinks big — as it has in other eras — it has the capacity to close the gap and keep leadership in the age of intelligence.
We can debate whether 100 gigawatts per year is the exact number needed. What’s indisputable is the logic: without enough energy and skilled workers, AI remains a promise. With power and prepared people, AI can become a real lever of prosperity.
