Nuclear power stocks are experiencing a dramatic surge in early 2025, driven by unprecedented demand from AI data centers and new government contracts signaling a broader industry transformation. Constellation Energy shares climbed 4% on Friday, building on Thursday’s 8% gain after announcing a landmark $1 billion deal to supply nuclear power to over a dozen US government agencies over the next decade—the largest energy purchase in US General Services Administration history.
Competitor Vistra Corp mirrored this momentum with an 8.5% jump Friday, adding to Thursday’s 8% gain. While the Constellation deal represents a relatively modest contract size, analysts view it as a bellwether for significantly larger agreements ahead, particularly with technology companies desperate to power their energy-intensive AI operations.
Morgan Stanley analyst David Arcaro noted that the pricing structure looks “very attractive” and could provide positive indicators for future deals. Notably, the contract included a surprisingly generous premium for a federal agency, suggesting that tech companies—facing less budgetary pressure and aggressive sustainability targets—would likely pay even higher rates for reliable nuclear power to fuel their data centers.
UBS analyst William Appicelli maintained his bullish stance, reiterating “buy” ratings on Constellation Energy, Vistra, and Talen Energy, with forecasts predicting an average 29% gain across these stocks in 2025. While impressive, this would represent a slight moderation from 2024’s extraordinary performance, when Vistra surged over 300%—outpacing even Nvidia’s celebrated rally. Talen Energy and Constellation Energy climbed over 200% and 100% respectively during the same period.
The nuclear power renaissance is fundamentally driven by AI data centers’ massive electricity requirements. These facilities consume enormous amounts of power to train and operate large language models and other AI systems, with demand already projected to exceed available supply in coming years. According to Apollo’s Torsten Sløk, the United States must more than double its power grid capacity by 2030 to meet this unprecedented demand, creating a critical infrastructure challenge that nuclear power is uniquely positioned to address with its reliable, carbon-free baseload generation.
Key Quotes
Pricing looks very attractive and could be a positive readthrough to future deals
Morgan Stanley analyst David Arcaro commented on the Constellation Energy government contract, suggesting the pricing structure indicates favorable terms for future agreements with private sector tech companies seeking power for AI data centers.
We think tech companies would be generally willing to pay a higher price given sustainability targets and less budgetary pressure, and there could be a broader customer set for these contracts than we were expecting
Arcaro explained why the nuclear power industry’s growth potential may exceed current expectations, as technology companies face fewer budget constraints than government agencies and are highly motivated to secure clean energy sources for their AI infrastructure.
Independent power companies should continue to benefit from strong demand and energy/capacity prices in 2025
UBS analyst William Appicelli provided his bullish outlook on the nuclear power sector, maintaining buy ratings and forecasting significant gains driven by AI-related electricity demand growth.
Our Take
The nuclear-AI nexus represents an unexpected but logical convergence that could define the next phase of technological development. What’s particularly striking is how AI’s infrastructure requirements are forcing a fundamental reassessment of energy policy faster than climate concerns ever did. Tech companies’ willingness to pay premium prices for reliable, clean power creates a compelling economic case that transcends political debates about nuclear energy. However, the 2030 timeline for doubling grid capacity appears extremely ambitious given regulatory hurdles and construction timelines for new nuclear facilities. This supply-demand imbalance could become AI’s next major bottleneck, potentially slowing innovation or forcing geographic concentration of AI development near existing power sources. The 29% projected gains for nuclear stocks may actually be conservative if power scarcity becomes acute.
Why This Matters
This development represents a critical inflection point where AI’s infrastructure demands are reshaping America’s energy landscape and creating massive investment opportunities. The nuclear power industry, long stagnant, is experiencing a renaissance driven entirely by artificial intelligence’s insatiable appetite for electricity. AI data centers require 24/7 reliable power that renewable sources like solar and wind cannot consistently provide, making nuclear energy the ideal solution for tech companies pursuing both operational reliability and sustainability commitments.
The implications extend far beyond stock prices. If the US must double its power grid capacity by 2030, we’re looking at hundreds of billions in infrastructure investment and potential energy bottlenecks that could constrain AI development itself. Tech companies are now competing for limited power resources, fundamentally altering their strategic planning and capital allocation. This also signals a broader acceptance of nuclear power after decades of public skepticism, potentially accelerating regulatory reforms and new reactor construction. The convergence of AI advancement and energy infrastructure represents one of the defining economic and technological challenges of this decade.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources: