AI's Power Crisis: Is Nuclear the Answer for Data Centers?

DELTA, UT – December 12, 2025 – The artificial intelligence revolution is running on an explosive mixture of algorithms, data, and immense electrical power. As businesses race to deploy generative AI, a colossal challenge is emerging from the shadows: the technology's insatiable energy consumption is straining power grids and threatening to derail corporate sustainability goals. In a landmark move that directly confronts this crisis, Creekstone Energy has announced a partnership with nuclear services expert EnergySolutions to evaluate adding at least two gigawatts of next-generation nuclear capacity to its massive data infrastructure campus in central Utah.

This strategic pivot signals a potential paradigm shift in how the digital economy will be powered. The non-binding Memorandum of Understanding (MOU) establishes a framework to explore integrating advanced nuclear reactors at the Creekstone Gigasite, a campus purpose-built to serve the voracious power and cooling demands of AI and hyperscale data centers. It’s a bold bet that the future of AI may be inextricably linked to the future of nuclear energy.

The AI Energy Dilemma Hits the Grid

The scale of AI's energy demand is staggering. According to the International Energy Agency (IEA), global electricity consumption from data centers could more than double by 2026 to over 1,000 terawatt-hours—roughly equivalent to the entire electricity consumption of Japan. In the United States, the Electric Power Research Institute (EPRI) projects that data centers could account for as much as 9.1% of all electricity generated by 2030, a dramatic increase from just 1.5% in 2023, with AI being the primary driver.

This unprecedented load growth presents a trilemma for the tech industry and utility providers. First, aging grid infrastructure is ill-equipped to handle such concentrated, rapid increases in demand, leading to bottlenecks and reliability concerns. Second, tech giants with ambitious carbon-neutrality pledges are finding it increasingly difficult to source enough renewable energy to power their operations, as the sheer scale of their needs outstrips available wind and solar capacity. Finally, the search for suitable data center sites has become a global scramble for locations with three critical resources: abundant power, robust fiber connectivity, and sufficient cooling—often water.

Utah's Gigasite: A New Model for Power and Compute

Creekstone Energy is tackling this challenge head-on with its Delta Gigasite in Millard County, Utah. The project is not merely a data center park; it is a next-generation, multi-source power and digital infrastructure campus designed from the ground up to solve the AI energy problem. Spanning nearly 1,200 acres rezoned for heavy industrial use, the campus is planned to host 20 million square feet of data centers.

The project’s core innovation lies in co-locating massive power generation directly with compute. Creekstone’s initial strategy involves developing a portfolio of approximately 10 gigawatts of non-nuclear generation, including solar, natural gas, and offtake from the nearby Intermountain Power Project. The development is already well underway, having secured crucial zoning approvals in May 2025 and a Series B funding round in November 2025 to accelerate construction. An initial agreement is already in place with BluSky AI Inc. to bring modular data centers online as early as 2026.

By integrating power production and consumption, the Gigasite aims to bypass traditional grid constraints, provide power at a scale AI requires, and create a more resilient and efficient ecosystem. The decision to now formally evaluate nuclear power is a logical, if audacious, next step in this strategy.

A Nuclear Partnership for the AI Age

The collaboration with EnergySolutions brings deep nuclear-sector expertise to the table. While primarily known for its comprehensive decommissioning and waste management services, EnergySolutions is providing the critical technical and regulatory knowledge needed to navigate the complexities of a new nuclear build. The MOU outlines a structured "Phase 1" evaluation to develop a roadmap for deploying at least two gigawatts of nuclear power, with a target commercial operation date between 2030 and 2035.

“Evaluating the role that next-generation nuclear could play in our broader energy portfolio is an important step in refining the long-term strategy for the Gigasite,” said Ray Conley, CEO of Creekstone Energy. “Nuclear has the potential to complement our multi-source approach and support the growth of large-scale AI and digital-infrastructure development. EnergySolutions’ technical and regulatory expertise gives us confidence that this evaluation will be thorough, professional, and grounded in real-world conditions.”

For AI workloads that demand uninterrupted, 24/7 power, nuclear energy offers a carbon-free baseload source that intermittent renewables cannot provide alone. This makes it an ideal complement to the Gigasite's planned solar capacity, ensuring a firm, reliable power supply regardless of weather conditions.

Pierre Oneid, Executive Vice President at EnergySolutions, emphasized the methodical approach. “Creekstone is approaching this the right way, grounding every step in thorough analysis, clear options, and a disciplined process,” he stated. “Our role is to help them fully understand the nuclear pathways available and what it would take to make any of those options viable for a site of this scale.”

The Path Forward: Hurdles and Next-Gen Technology

Bringing two gigawatts of new nuclear power online within a decade is an ambitious goal fraught with challenges. The evaluation will assess a wide range of reactor technologies, from proven large-scale designs to advanced Small Modular Reactors (SMRs). SMRs, from vendors like NuScale Power, GE Hitachi, and TerraPower, are seen as particularly promising due to their potential for modular construction, enhanced safety features, and smaller footprint.

NuScale’s design is the first SMR to be fully certified by the U.S. Nuclear Regulatory Commission (NRC), but the path from certification to operation remains a complex and capital-intensive journey. The Creekstone-EnergySolutions evaluation will scrutinize not only the technological readiness of these options but also the formidable regulatory pathways, site integration requirements, and commercial delivery models.

While Utah's state government has shown a favorable stance toward advanced nuclear technology, any proposal will face intense public and regulatory scrutiny regarding safety, cost, and long-term waste disposal. However, the sheer necessity driven by AI's power demands is fundamentally altering the risk-reward calculation for nuclear energy. This evaluation in the Utah desert is no longer a theoretical exercise; it is a pragmatic response to a pressing business imperative. How Creekstone navigates this complex path could well define the blueprint for powering the next generation of American innovation.