Key Takeaways

  • Amazon is financing milestone-based development work, not construction, for a proposed 960 MW nuclear facility in Sullivan County.
  • AEP and X-energy are evaluating 12 Xe-100 small modular reactors at the retired Breed Generating Station.
  • Project cost, ownership, licensing path, construction schedule, and commercial-operation date remain undecided.

Amazon, AEP, and X-energy have entered an agreement to evaluate one of the largest proposed small modular reactor deployments in the United States, linking rising data-center electricity demand with plans to repurpose a retired Indiana coal site.

The distinction between development and construction matters. Amazon is funding milestone-based feasibility and development work for the Sullivan County proposal, while a final investment decision has not been made. The funding amount also remains confidential. In other words, the agreement moves the concept forward, but it does not yet authorize crews to begin building reactors.

Under the proposal, X-energy would supply 12 Xe-100 reactors, each rated at 80 MW. Together, the units would provide 960 MW to the regional grid. AEP said Amazon intends to purchase electricity from the project through a long-term power agreement, creating the kind of committed demand that can help support financing for capital-intensive generation.

That structure reflects a broader change in how large technology buyers approach energy procurement. Conventional renewable contracts can help offset annual electricity consumption, but data centers operate around the clock and place substantial demands on transmission systems. Nuclear generation offers steady output that is not dependent on wind or sunlight, although advanced reactors bring their own commercial, regulatory, and supply-chain risks.

Committing to 960 MW is a substantial step for technology that has not yet reached commercial deployment. X-energy's Xe-100 is an advanced, high-temperature gas reactor that uses TRISO fuel. Its cooling configuration can use air, water, or a hybrid arrangement, giving developers more flexibility in regions where water availability is a concern. That potential benefit still has to be demonstrated at commercial scale.

The proposed location could help. The former Breed Generating Station in Fairbanks already has a history as a power-generation site, potentially offering access to transmission infrastructure, industrial land, and a workforce familiar with energy projects. Reusing former coal properties can also preserve the economic role of generation communities as the power system changes.

Still, a brownfield location does not remove the harder parts. The project would be subject to U.S. Nuclear Regulatory Commission review, and the partners have not announced which licensing pathway they expect to pursue in Indiana. Fuel availability, component manufacturing, engineering validation, local approvals, construction execution, and cost allocation could all influence whether the plan advances.

POWER Magazine reported that the partners have not disclosed a project cost, construction timeline, commercial-operation target, or final ownership structure. Those omissions are understandable at the feasibility stage, but they are central to the business case. Who carries construction risk? How would overruns be treated? And how much of the output would ultimately serve Amazon's demand rather than other regional customers?

Amazon says the facility could support 1,000 construction jobs and 100 permanent positions. Those figures remain projections because construction has not been approved. Indiana also has a cautionary nuclear precedent in the Marble Hill Nuclear Power Plant, which was abandoned half-finished in 1983 after $2.5 billion had been spent. The technologies and commercial structures differ, but the history helps explain why cost controls and staged commitments will receive close attention.

The Indiana proposal is the second deployment contemplated under the Amazon and X-energy collaboration, which targets more than 5 GW of new U.S. nuclear capacity by 2039. Amazon previously led X-energy's 2024 Series C-1 financing, which ultimately closed at $700 million to support reactor design, licensing, and the first phase of a TRISO-X fuel facility.

For Amazon, the arrangement provides a path toward firm electricity without immediately assuming the full risk of plant construction. For AEP, it offers a major prospective customer alongside utility and nuclear-operating capabilities. X-energy gains another potential deployment site and demand anchor. The near-term test is less dramatic than reactor construction, but just as consequential: whether the partners can turn feasibility work into a financeable, licensable project with a credible schedule.