A 960-megawatt plan, with Amazon as first customer, depends on a reactor design that has yet to be built in the US.
Data centers need a lot of electricity, and the companies that run them are increasingly looking at nuclear power to supply it. On October 19, TechCrunch reported that Amazon will fund the development of a proposed 960-megawatt nuclear plant in Indiana and will be its first power customer.
The plan for 12 small nuclear reactors
- Where: the site of a demolished coal plant just south of Terre Haute.
- Who: the utility American Electric Power would operate the plant, and X-energy would supply 12 of its Xe-100 small modular reactors. At 960 megawatts across 12 units, that is about 80 megawatts each.
- How it connects: the plant would feed the public grid, and Amazon says the power would also serve its data centers.
- How big: this is Amazon and X-energy's second commercial agreement and roughly three times the size of the first, announced in 2024. Amazon also led a $700 million funding round in X-energy that October.
- The goal: the companies aim to build 5 gigawatts of new US nuclear capacity by 2039.
The risks, plainly
The schedule depends on X-energy first delivering a reactor plant at a Dow Chemical site in Seadrift, Texas. That project faces a lawsuit from the state of Texas over alleged regulatory non-compliance. X-energy has not broken ground there, and its construction permit is in the final stages of review by the Nuclear Regulatory Commission.
The industry's pitch is that building small reactors in factories, in large numbers, will tame the cost overruns and delays that have dogged recent US nuclear projects. TechCrunch notes that experts say those benefits could take about a decade to show up. X-energy went public in April 2026 at $23 a share, rose about 10 percent before the market opened on the news, then slipped, and its shares have fallen by nearly two-thirds from their first-day peak.
Why Amazon's nuclear plan matters
Converting a retired coal site into a power plant reuses existing grid connections, which is part of the appeal. But the real test is whether small modular reactors can be built on time and on budget. Until the first ones are running, announcements like this are best read as commitments to try, not proof that it works.
