Amazon is making a major strategic pivot toward nuclear energy to power its rapidly expanding AI infrastructure, according to internal documents obtained by Business Insider. In June 2024, CEO Andy Jassy approved an ambitious plan from Amazon’s Global Energy Strategy team to invest heavily in nuclear power, which the company believes is “the most economical and credible path to scale carbon free energy” for the next decade’s capacity growth.
The driving force behind this nuclear push is generative AI. Running AI workloads in data centers requires significantly more electricity than traditional computing tasks, creating an unprecedented energy demand challenge. Amazon has responded with a multi-pronged nuclear strategy that includes both existing plants and cutting-edge technology.
Last week, Amazon led a $500 million investment in X-energy to develop small modular reactors (SMRs), following an earlier $650 million deal to purchase energy from a nuclear plant in Pennsylvania. The company has also been “actively conducting due diligence” on more than a dozen SMR opportunities and has developed relationships with utilities, nuclear construction companies, and the US Department of Energy.
Amazon’s nuclear ambitions extend beyond these announced deals. Internal documents reveal the company considered tapping into at least four additional existing nuclear power plants under a project code-named “Piano Man.” These potential deals span Maryland (0.88 gigawatts), Texas (1.2 gigawatts), and two additional Pennsylvania sites (0.96 and 0.9 gigawatts). For context, a 1-gigawatt plant can power approximately 876,000 households annually.
Amazon isn’t alone in this nuclear embrace. Google recently partnered with Kairos Power for SMRs, Microsoft struck a deal with the Three Mile Island nuclear plant, and OpenAI CEO Sam Altman has backed nuclear projects. This industry-wide shift reflects AI’s “insatiable hunger for power” and Big Tech’s struggle to meet sustainability commitments while scaling AI infrastructure.
The company’s energy strategy also explores geothermal and natural gas options, aiming to deliver power “in large amounts beyond 2030.” Amazon emphasized it will “prioritize opportunities with the earlier access to power” while targeting technologies “on the cusp of scaling and mass adoption.”
Key Quotes
We have strong conviction that nuclear power is the most economical and credible path to scale carbon free energy to meet the next decade’s rapid capacity growth.
Amazon’s Global Energy Strategy team wrote this in an internal third-quarter document, explaining the rationale they presented to CEO Andy Jassy in June 2024. This statement reveals Amazon’s strategic thinking that nuclear, despite alternatives, offers the best combination of reliability, cost-effectiveness, and carbon-free operation for AI-driven data center expansion.
Machine Learning, GenAl and commercial capacity demand continues to significantly outpace available power capacity globally and requires new solutions that are reliable, cost-effective and carbon free.
This quote from Amazon’s internal document explicitly connects the company’s nuclear strategy to AI workloads. It highlights the unprecedented challenge facing cloud providers: AI’s computational demands are growing faster than traditional energy infrastructure can support, necessitating radical new approaches to power generation.
There will be a learning curve because some of these technologies are not mature yet.
Jacopo Buongiorno, a nuclear-engineering professor at MIT, provided this cautionary note about small modular reactors. While acknowledging nuclear’s potential benefits in safety, reliability, and cost-efficiency, he emphasizes that SMRs represent relatively unproven technology with varying designs, suggesting Amazon and other tech giants are taking calculated risks with these investments.
While we target obtaining large amounts of power over time, we will prioritize opportunities with the earlier access to power.
This statement from Amazon’s internal document reveals the company’s strategic urgency. Despite planning for massive power needs beyond 2030, Amazon is prioritizing any solution that can deliver energy sooner, indicating the immediate pressure AI workloads are placing on existing infrastructure and the competitive advantage early power access provides.
Our Take
Amazon’s nuclear pivot represents a fascinating collision between AI ambition and physical reality. While the tech industry has spent years discussing AI’s transformative potential, this story reveals a less-discussed constraint: electricity is becoming the bottleneck for AI advancement. The fact that Amazon received CEO approval for a multi-billion-dollar, decades-long energy strategy specifically to support AI workloads demonstrates how fundamentally generative AI has changed cloud computing economics. What’s particularly striking is the timeline urgency—Amazon is willing to bet on unproven small modular reactor technology and revive controversial plants like Three Mile Island because traditional renewables simply cannot scale fast enough. This suggests we may be approaching an inflection point where energy availability, not algorithmic innovation, determines which companies can compete in AI. The industry-wide coordination among Amazon, Google, and Microsoft on nuclear also hints at a shared recognition that AI’s energy demands exceed any single company’s ability to solve independently.
Why This Matters
This story represents a watershed moment for both the AI and energy industries. Amazon’s nuclear commitment signals that generative AI’s computational demands are fundamentally reshaping corporate energy strategies and could revitalize the long-stagnant nuclear sector. The fact that Amazon’s CEO personally approved this strategy underscores how critical energy availability has become to AI competitiveness.
The broader implications are profound: AI’s energy requirements are forcing tech giants to make decades-long infrastructure bets that will shape the energy landscape through 2030 and beyond. This could accelerate small modular reactor development, potentially making nuclear more accessible and safer. However, it also highlights a concerning reality—AI’s growth is outpacing renewable energy solutions like wind and solar, forcing companies toward nuclear despite its controversial history.
For businesses and society, this trend suggests AI expansion may be constrained by energy availability rather than technological capability. Companies planning AI deployments must now factor in power infrastructure, while communities near potential nuclear sites face new economic opportunities and safety concerns. The success or failure of these nuclear bets will significantly influence AI’s trajectory and the feasibility of achieving carbon-free computing at scale.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- How AI Can Help Make Buildings More Energy Efficient
- Amazon to Invest Additional $4 Billion in AI Startup Anthropic
- EnCharge AI Secures $100M Series B to Revolutionize Energy-Efficient AI Chips
Source: https://www.businessinsider.com/amazon-nuclear-power-strategy-ai-2024-10