The Nuclear Revival: A Faustian Bargain for AI’s Sake

The news that Microsoft has agreed to pay $16 billion to buy every watt of electricity from the revived Three Mile Island nuclear plant has sent shockwaves through the industry. This significant turning point could potentially revitalize stalled progress on new reactor construction, but it also raises profound questions about the true cost of America’s addiction to AI-driven innovation.

Microsoft is not alone in its investment; tech giants are pouring billions into nuclear energy as a means to power their data centers. According to reports, every major AI company – Microsoft, Google, Amazon, Meta – has signed at least one nuclear deal in 2026, committing to more than a dozen agreements worth nearly 10 gigawatts of capacity. These companies are not just buying electricity; they’re financing the construction of power plants directly, succeeding where governments and utilities have failed for decades: making nuclear power affordable to build again.

The shift away from intermittent energy sources like solar and wind is notable. As Rafael Mariano Grossi, head of the International Atomic Energy Agency, noted, nuclear power offers “round-the-clock reliability,” precisely what AI infrastructure needs. This reality challenges conventional wisdom about renewable energy sources.

Global electricity demand is projected to grow by more than 10,000 terawatt-hours by 2035, with data centers driving a significant portion of this growth. By decade’s end, AI processing in the U.S. is expected to use more power than the country’s aluminum, steel, cement, and chemical industries combined. The queue for projects waiting to connect to the U.S. grid has swollen past 2,600 gigawatts, with average wait times nearing five years.

Tech companies are under immense pressure to deliver seamless service for their AI infrastructure. Microsoft’s Three Mile Island restart is just one example; Meta has signed agreements worth up to 6.6 gigawatts of nuclear power, while Google has struck deals with two nuclear developers and a separate contract for 1,800 megawatts of new capacity. Amazon has invested $700 million in X-energy to build compact, factory-built units designed to be faster and cheaper than traditional plants.

However, there’s a fundamental flaw in this approach: it assumes that the manufacturing, licensing, and workforce required for these small modular reactors are proven technologies. Building a single new reactor costs billions and takes the better part of a decade under ideal conditions. The timelines are badly out of sync with AI’s growth.

The skeptics’ case is not without merit; it highlights the inherent contradictions in this Faustian bargain. However, they’re half right: while the new reactors may be years from producing a single watt, built for the 2030s, not today’s crunch, the power actually arriving in the next few years comes from restarts like Three Mile Island, not new construction.

This development raises profound questions about America’s priorities and values. Are we willing to sacrifice our environmental and health concerns for the sake of AI-driven innovation? Or will we find a way to balance these competing demands? The answer lies not in the deals themselves but in the broader context they reflect: a world where data centers are becoming the new industrial behemoths, driving energy demand like never before.

As we confront this uncharted territory, it’s essential to acknowledge that the nuclear revival is not a zero-sum game. It’s a Faustian bargain that requires us to confront the true costs of our addiction to AI and its attendant infrastructure.