According to Utility Dive, OpenAI warned the White House on Monday that China’s massive energy expansion could give it a decisive edge in the AI race. Christopher Lehane, OpenAI’s chief global affairs officer, told the White House Office of Science and Technology Policy that China added 429 GW of new power capacity last year compared to just 51 GW in the US, representing more than one-third of the entire US grid capacity. The company recommends the US prioritize building 100 GW annually of new energy capacity to close this “electron gap” and maintain competitiveness against China’s goal to lead global AI by 2030. Lehane’s letter cited International Energy Agency data showing China accounted for 63% of global renewable energy expansion in 2023 and is expected to maintain roughly 60% of that expansion through 2030. This warning comes as China’s AI models like DeepSeek and Alibaba’s Qwen demonstrate significant advances in performance benchmarks.
Table of Contents
Energy: The New AI Battleground
What OpenAI’s warning reveals is that the AI arms race has fundamentally shifted from being purely about algorithms and talent to being about raw energy capacity. Artificial intelligence models, particularly the large language models that companies like OpenAI develop, consume staggering amounts of electricity both during training and inference phases. The most advanced models today require energy equivalent to powering thousands of homes for extended periods. China’s strategic recognition that energy infrastructure equals AI competitiveness represents a fundamental shift in how nations should approach technological leadership. While the US has focused on chip restrictions and talent retention, China has been building the foundational infrastructure that will ultimately determine which nation can train and deploy the next generation of AI systems at scale.
Regulatory Bottlenecks Threaten US Competitiveness
The regulatory challenges facing US energy development cannot be overstated. While Lehane’s recommendations to streamline FERC Order Number 2023 interconnection timelines and create fast tracks for shovel-ready projects are sensible, they barely scratch the surface of America’s energy infrastructure problems. The US faces a perfect storm of local opposition, environmental review processes that can take years, and a fragmented regulatory landscape across states and regions. China’s centralized planning system allows it to approve and construct major energy projects in timelines that would be unimaginable in the US. This isn’t just about building more solar farms or wind turbines—it’s about transmission infrastructure, grid modernization, and interconnection capacity that can handle the massive, concentrated energy demands of AI data centers. The White House and Office of Science and Technology Policy face the difficult task of balancing environmental protections and community input against the urgent need for energy infrastructure that can support national competitiveness.
Nuclear Advantage Emerging
Perhaps the most concerning aspect of China’s energy expansion is its nuclear program. While the US currently leads in total installed nuclear capacity at 97 GW compared to China’s 55 GW, China’s growth trajectory tells a different story. With 28 reactors under construction adding 30 GW of capacity, China is positioned to surpass not just France but potentially the United States in nuclear generation within the next decade. Nuclear power provides the stable, high-density energy that AI data centers require—something intermittent renewables struggle to deliver consistently. The US nuclear industry has been largely stagnant for decades, with only a handful of new reactors coming online despite advanced nuclear technologies showing promise. If China successfully scales its nuclear program while maintaining its massive renewable expansion, it could achieve the energy diversity and reliability needed to support AI development that other nations simply cannot match.
Beyond the Numbers: Strategic Implications
The energy gap between the US and China represents more than just different infrastructure priorities—it reflects fundamentally different approaches to technological competition. China’s energy expansion is part of a coordinated industrial policy that aligns energy, technology, and economic objectives. The US approach remains fragmented across private sector initiatives, state-level policies, and federal regulations that often work at cross-purposes. The risk isn’t just that China might develop better AI models—it’s that they could achieve such overwhelming scale advantages that US companies become permanently disadvantaged. When one nation can train models with 100 times the computational resources because they have the energy infrastructure to support it, algorithmic advantages may become irrelevant. This is particularly concerning for open-source AI development, where China’s DeepSeek and other models are already showing strong performance despite US attempts to restrict chip exports.
Realistic Path Forward
Closing the “electron gap” that OpenAI identifies will require more than regulatory tweaks—it demands a fundamental rethinking of how the US approaches energy infrastructure for technological competitiveness. The recommendation to build 100 GW annually represents nearly double what the US achieved last year, and achieving this would require overcoming significant political, regulatory, and technical hurdles. More importantly, it requires recognizing that energy policy is now inseparable from technology and national security policy. The US must develop a coherent strategy that addresses not just generation capacity but transmission, grid resilience, and the specific energy demands of advanced computing. As detailed in OpenAI’s full submission, this will require coordinated action across multiple government agencies and private sector stakeholders. The window for action is closing rapidly, with China’s 2030 AI leadership target just five years away.
