Deutsch
English
Español
Français
Nederlands
Português
In 2026, the artificial intelligence revolution is reshaping not just computing but the global energy industry. With AI training and inference driving an unprecedented surge in electricity demand, major technology firms are directly investing in nuclear power — transforming from pure energy consumers into energy producers. This strategic pivot, marked by Microsoft's restart of Three Mile Island, Amazon's $20 billion conversion of the Susquehanna nuclear site, and X-Energy's $1 billion IPO for small modular reactors, signals a new era where tech giants become power companies to secure the baseload clean energy their AI ambitions require.
The Electricity Demand Crisis Driving the Nuclear Pivot
Global data center electricity consumption is projected to nearly double by 2030, according to Gartner, rising from 448 terawatt-hours in 2025 to 980 TWh by 2030. In the United States alone, data center electricity demand is expected to reach 6–12% of total national consumption in 2026. The International Energy Agency's Electricity 2026 report highlights that AI and data centers are among the fastest-growing sources of power demand globally, alongside the electrification of transport and industry.
The AI infrastructure energy challenge is acute because AI workloads are uniquely power-intensive. Training a single large language model can consume as much electricity as hundreds of homes use in a year, and inference — the process of running AI models in real time — adds continuous demand. Unlike traditional cloud computing, AI inference requires low-latency, always-on power that intermittent renewables like solar and wind cannot reliably provide without massive battery storage.
Major Tech-Nuclear Deals Reshaping the Energy Landscape
Microsoft and Three Mile Island: A Historic Restart
In one of the most symbolic deals of the nuclear renaissance, Microsoft signed a 20-year power purchase agreement with Constellation Energy to restart Three Mile Island Unit 1 in Pennsylvania. The reactor, which was shut down in 2019 for economic reasons, will provide 835 megawatts of carbon-free electricity exclusively for Microsoft's AI data centers. Constellation is spending approximately $1.6 billion to refurbish the plant, supported by a $1 billion federal loan approved in November 2025. The plant is expected to come back online in 2027, with its license extended to 2054.
The Three Mile Island restart controversy has drawn both praise and scrutiny. Supporters argue it preserves skilled nuclear jobs and provides reliable clean power; critics question the cost and the precedent of dedicating a major power plant to a single corporate customer.
Amazon's $20 Billion Susquehanna Investment
Amazon has committed over $20 billion to build two data center complexes in Pennsylvania, including one adjacent to the Susquehanna nuclear power plant. The nuclear-adjacent site can draw up to 960 megawatts — roughly 40% of the plant's output — via a 'behind the meter' connection that Amazon acquired from Talen Energy for $650 million in 2024. This arrangement, however, is under review by the Federal Energy Regulatory Commission (FERC), which has blocked the deal on procedural grounds, raising questions about grid fairness and market design.
Pennsylvania Governor Josh Shapiro called the investment "the largest capital investment in state history," expected to create 1,250 high-skilled jobs. Amazon has committed roughly $10 billion each to similar nuclear-powered data center projects in Mississippi, Indiana, Ohio, and North Carolina since 2024.
X-Energy's $1 Billion IPO and the SMR Revolution
On April 23, 2026, Amazon-backed nuclear startup X-Energy raised $1.02 billion in its Nasdaq IPO, pricing shares at $23 each — well above the initial $16–$19 range. The Maryland-based company specializes in Generation IV high-temperature gas-cooled small modular reactors (SMRs) and proprietary TRISO-X fuel. X-Energy's customer pipeline exceeds 11 gigawatts, with Amazon committed to purchasing up to 5 GW of nuclear power by 2039 and Dow Chemical partnering on industrial applications with up to $1.2 billion in Department of Energy support.
The small modular reactor investment boom reflects growing investor confidence that SMRs can deliver nuclear power at lower upfront costs and with greater siting flexibility than traditional large reactors. X-Energy aims to deliver its first commercial reactors by the early 2030s.
Strategic Implications: Tech Companies as Energy Producers
This nuclear pivot transforms the strategic calculus of the energy transition. Tech companies are bypassing strained electricity grids by co-locating data centers directly at nuclear plants, securing dedicated power supply without competing with residential or industrial customers. This 'behind the meter' model offers reliability and cost predictability but raises concerns about grid equity and the potential for inflated electricity costs for other users.
The private sector nuclear renaissance is also reshaping the economics of baseload clean power. Nuclear plants, which operate at capacity factors above 90%, provide the 24/7 carbon-free electricity that AI infrastructure demands — something renewables cannot yet match at scale. Goldman Sachs estimates that 85–90 GW of new nuclear capacity is needed by 2030 to meet projected data center demand, but well under 10% of that is currently under development.
Regulatory and Policy Context
The EU AI Act's full enforcement begins in August 2026, introducing binding requirements for high-risk AI systems including those used in critical infrastructure, education, employment, and law enforcement. Fines can reach €35 million or 7% of global annual revenue. This regulatory framework adds urgency for tech companies to secure compliant, sustainable energy sources for their AI operations.
In the United States, FERC's review of behind-the-meter nuclear deals will set important precedents for how dedicated corporate power purchases interact with wholesale electricity markets. The Biden administration and the Trump administration have both supported nuclear expansion through loan guarantees and tax credits, reflecting bipartisan recognition of nuclear's role in energy security and decarbonization.
Expert Perspectives
"The scale of AI electricity demand is unlike anything the grid has faced since the post-war industrial boom," says Dr. Maria Korsnick, president of the Nuclear Energy Institute. "Tech companies are not just buying power — they are investing billions to build new generation capacity. This is a fundamental shift in how energy markets work."
However, critics warn of unintended consequences. "If the largest corporate customers bypass the grid entirely, they avoid paying for grid maintenance and reliability that everyone else depends on," notes energy economist Dr. Severin Borenstein of UC Berkeley. "This could lead to higher costs for households and small businesses."
FAQ
Why are tech companies investing in nuclear power? AI data centers require massive, reliable, 24/7 electricity that intermittent renewables cannot provide. Nuclear offers carbon-free baseload power with high capacity factors, making it ideal for AI infrastructure.
What is the Three Mile Island restart? Microsoft and Constellation Energy are restarting Three Mile Island Unit 1 (not the reactor involved in the 1979 accident) to provide 835 MW of power exclusively for Microsoft's AI data centers, with the plant expected online in 2027.
How much are tech companies spending on nuclear? Amazon has committed over $20 billion to nuclear-powered data centers in Pennsylvania alone. Microsoft's Three Mile Island deal is valued at $16 billion over 20 years. X-Energy raised $1.02 billion in its 2026 IPO.
What are small modular reactors (SMRs)? SMRs are advanced nuclear reactors with lower upfront costs, factory fabrication, and flexible siting. X-Energy's Xe-100 is a Generation IV high-temperature gas-cooled reactor designed for electricity, industrial heat, and hydrogen production.
Will nuclear power raise electricity costs for consumers? Potentially. If large tech companies secure dedicated power through behind-the-meter deals, they may avoid grid costs that other customers must cover, potentially shifting costs to households and small businesses.
Conclusion: Can the Nuclear Renaissance Scale Fast Enough?
The AI-nuclear pivot of 2026 represents a defining strategic tension: tech giants are becoming energy companies to secure the power they need, but the question remains whether this private-sector nuclear renaissance can scale fast enough to meet AI's insatiable demand without crowding out other grid users or inflating electricity costs. With global data center electricity demand on track to double by 2030 and less than 10% of needed new nuclear capacity currently under development, the race between AI's growth and clean energy deployment will define the next decade of energy policy.
Follow Discussion