Blog Post by Thomas Glawson, Junior Associate
Introduction
As Artificial Intelligence (AI) technology develops and expands exponentially, it has become the dominant point of contention in the geopolitical landscape, particularly between the United States (U.S.) and China. The pursuit of AI dominance is viewed by both nations as a matter of national security.[1] Winning the AI race is expected to usher in an industrial revolution, an information revolution, and a “renaissance” of intellectual achievement.[2] Likewise, AI is crucial for optimizing complex energy systems, which are increasingly becoming electrified and digitalized.[3] AI has begun to shape the contours of foreign and domestic policy for both nations; not for algorithmic dominance, but the material components that make AI possible: energy, microchips, and rare earth minerals.
Geopolitical Stances and the Race for Dominance
The high stakes of the AI arms race are constantly emphasized by industry leaders and government officials. Jensen Huang, CEO of Nvidia, has stressed the vital need for the U.S. to win the global AI race, asserting that China is only “nanoseconds behind America in AI”.[4] He urged the federal government to promote policies to secure America’s victory, while noting that Western nations are currently constrained by a “cynicism” regarding the technology.[5] Interior Secretary Doug Burgum emphasized that the US must win the AI arms race, linking energy security to national security.[6]
The competition is fueled by the vast potential of AI to drive efficiency. For example, AI applications in power plant operations and maintenance could yield potential cost savings of up to 110 billion USD annually by 2035.[7] In end-use sectors, energy savings of 8% could be achieved by 2035 in light industry.[8] AI can also accelerate innovation dramatically, such as achieving a 45,000-fold acceleration in the scientific rate of discovery of three-dimensional protein structures in medicine.[9]
Key Issues
The AI competition is fundamentally defined by critical bottlenecks in energy infrastructure, computing components, rare earth minerals and supply chains.
Energy Demands
AI model training and deployment primarily occur within data centers, where servers equipped with specialized accelerators (GPUs) account for around 60% of electricity demand.[10] According to the International Energy Association, global data center electricity consumption is projected to double to roughly 945 terawatts (TWh) by 2030, growing about 15% per year—more than four times faster than all other sectors combined.[11] The US and China account for nearly 80% of this global growth through 2030.[12]
In the US, data centers already account for about 6% of national electricity demand.[13] The US has the highest per-capita data center consumption in the world, projected to grow to over 1,200 kilowatts (kWh) per capita by the end of the decade.[14] However, a lack of new power plant constructions mean that eight of 13 regional grids are at or below critical spare capacity levels, which analysts warn could become a bottleneck for further data center developments in the US.[15] Natural gas is currently the largest source of electricity for US data centers, with over 40% of the share.[16]
To meet the surging electricity demands, technology companies are engaging with nuclear power. Google has unveiled a plan to restart the Duane Arnold Energy Center, a US nuclear facility in Iowa to power their AI infrastructure; the plant was shuttered in 2020, and is expected to return to service in 2029 as a part of 25-year agreement with NextEra Energy to purchase power from the facility.[17] Likewise, the power company Constellation announced it is in the process of restarting the Three Mile Island nuclear plant in Pennsylvania following a 20-year power purchase agreement with Microsoft.[18] Nuclear power is set to play a significant role in meeting US data center electricity demand, particularly after 2030 when the first Small Modular Reactors (SMRs) are expected to be commissioned.[19] Hyperscalers (operators of large data centers) are among the key corporate backers of SMR development, with technology companies planning to finance more than 20 gigawatts(GW) of SMRs to date in the US.[20] SMRs are expected to provide a source of baseload low-emission electricity to data centers operators after 2030 in both the US and China.[21]
China is well-positioned for meeting energy benchmarks, as the nation has been on a massive electricity construction spree involving renewables, coal-burning plants, and nuclear reactors following power shortages in 2021 and 2022.[22] By 2030, China is expected to have roughly 400GW of spare power capacity, which is triple the projected needs of the entire global data center fleet.[23] Coal currently dominates China’s data center electricity supply mix, with a nearly 70% share.[24] However, renewables (solar and wind power) and nuclear power are expected to increase significantly after 2030, reducing reliance on coal.[25]
Microchips and Restrictions
Microchips, especially Graphics Processing Units (GPUs) are fundamental to AI computing power.
Under the Biden Administration, the US CHIPS and Science Act of 2022 was passed, aimed at reestablishing advanced manufacturing for logic and memory in the US, as well as boosting and establishing other chipmaking activities.[26] After the law took effect, the federal government began negotiations and proposed deals for more than 30 projects by the end of October 2024.[27] After Donald Trump won the 2024 election, the Department of Commerce’s (DOC) CHIPS Office converted those proposed deals into awards, agreeing to more than $30 billion roughly two months before Trump took office.[28] The Trump administration has continued to utilize the groundwork laid by the CHIPS Act. On July 10th, 2025, The White House announced the US AI Action Plan, which aims to restore American semiconductor manufacturing.[29] The Action Plan lists a variety of recommendations. First, they sought to revamp the CHIPS Program Office as well as removing all extraneous policy requirements for CHIPS-funded semiconductor manufacturing projects.[30] Second, the DOC will review semiconductor grants and research programs to ensure that they accelerate the integration of advanced AI tools into semiconductor manufacturing.[31] Third, the DOC will streamline regulations that hinder semiconductor manufacturing efforts, especially permitting semiconductor manufacturing facilities.[32] Lastly, it recommends that the US government should prioritize investment in developing and scaling foundational manufacturing technologies—especially those related to AI—through various federal agencies.[33]Additionally, it seeks to strengthen AI compute(hardware) export control enforcements by leveraging location verification features on advanced AI compute to ensure chips are not diverted to “countries of concern.”[34]
In response, China has implemented aggressive measures, mandating that new state-funded data center projects must only use domestically-made AI chips.[35] Data centers that are less than 30% complete have been ordered to remove foreign chips or cancel purchase plans.[36] This action is intended to bolster local rivals like Huawei and has reportedly caused Nvidia’s market share in China to drop to zero.[37]
Rare Earth Minerals and Supply Chains
Critical minerals are the backbone of artificial intelligence production.[38] These minerals—essential for components like magnets and cooling systems—are necessary to sustain the massive data processing and energy requirements of AI systems.[39]
Since 2020, the United States Geological Survey (USGS) has maintained a Critical Minerals list, which serves as a necessary tool for the federal government’s critical minerals strategy, shaping federal investment, permitting decisions, and strategic planning for mineral development nationwide.[40] This list identifies minerals deemed most critical for America’s economic and national security.[41] For the draft 2025 list, USGS focused on America’s dependence on other countries for mineral commodities and the impact any disruptions would have on the US economy.[42] Minerals were included if a yearlong supply chain disruption was projected to decrease the US GDP by $2 million or more, resulting in a risk characterization of high, elevated, or moderate.[43] While not all the critical minerals are rare earth minerals, there is significant overlap, with many rare earth minerals appearing on the list due to high-risk assessments.[44]
China holds a strategic advantage, controlling 70% of the world supply of rare earth mining and 90% of processing.[45] As a result, China has a tight control over the essential rare earth supply chains, and has seen Chinese investors investing into the country’s domestic industries.[46] In response to growing concerns over lofty valuations in pure-play AI stocks (company-issued stocks that focus on only one line of business or industry), Chinese investors have pivoted their investment into utilities and metal producers.[47] It is estimated that about one-third of China’s total AI spending by 2030 will go toward building supporting facilities, with metals, power, and cooling systems absorbing the bulk of investments.[48] This massive buildout is driving demand for specific metals, notably copper and aluminum.[49] Data centers are becoming a key driver for China’s copper demand, expected to grow at an average annual rate of 20% through 2030.[50] Demand is also increasing for emergency backup equipment like energy storage systems.[51]
Despite production from the US’s only rare earth mine (MP Materials), approximately 80% of the quantity produced was exported for refinement in 2024, meaning domestic industry accounts for only 20% of US consumption.[52] Crucially, the US currently has little light earth processing capacity and no heavy rare earth processing capabilities, with most of this processing being done in China.[53] China has previously used their processing dominance as an economic weapon through export controls.[54]
To address and mitigate the risk of these gaps, the US is increasing international cooperation with allied countries. The US Interior Secretary Burgum announced the US will soon launch a global trading club for critical minerals; aimed at securing supply chains and ensuring allies do not buy from adversaries.[55] The US has identified Canada and Australia as key partners for filling in every point of the supply chain–including mining, refinement, components, and recycling.[56] Additionally, the US and Thailand recently signed a Memorandum of Understanding(MOU), which aims to develop domestic processing industries, encourage investments, and support a transition to open, efficient, secure, and transparent critical minerals and rare earth markets.[57]
Conclusion
The geopolitical competition for Artificial Intelligence supremacy between the United States and China is far more than a race about the AI algorithm quality but rather a multifaceted struggle relying on the physical foundations of energy infrastructure and secure supply chains for microchips and critical minerals. The nation that most effectively secures these resources–whether through international alliance like the US or through state-directed self-reliance like China–will unlock AI’s full potential. This struggle for technological supremacy will also define the future global balance of power.
[1] America’s AI Action Plan at 1, The White House (July 10, 2025), https://www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf.
[2] Id.
[3] International Energy Agency, AI for Energy Optimisation and Innovation, International Energy Agency (2025), https://www.iea.org/reports/energy-and-ai/ai-for-energy-optimisation-and-innovation#abstract.
[4] Anubhav Mukherjee, Jensen Huang U-turns on China ahead of US in AI Race Claim, Says Vital THat America Wins, Livemint (Nov. 6, 2025), https://www.livemint.com/companies/people/jensen-huang-u-turns-on-china-ahead-of-us-in-ai-race-claim-says-vital-that-america-wins-11762443976843.html.
[5] Id.
[6] Taylor Delandro, US Must Win AI Arms Race, Warns Interior Secretary Doug Burham, The Hill (Nov. 3, 2025), https://thehill.com/homenews/5586354-us-must-win-ai-arms-race-warns-interior-secretary-doug-burgum/.
[7] International Energy Agency, supra note 3.
[8] Id.
[9] Id.
[10] International Energy Agency, Energy Demand From AI, International Energy Agency (2025), https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai.
[11] Id.
[12] Id.
[13] Dan Murtaugh, Goldman Sees China Power Push Giving It Edge Over US in AI Race, Bloomberg News (Nov. 14, 2025), https://www.bloomberg.com/news/articles/2025-11-14/goldman-sees-china-power-push-giving-it-edge-over-us-in-ai-race.
[14] International Energy Agency, supra note 10.
[15] Murtaugh, supra note 13.
[16] International Energy Agency, Energy Supply For AI, International Energy Agency (2025), https://www.iea.org/reports/energy-and-ai/energy-supply-for-ai#abstract.
[17] Google Joins Microsoft In Plans to Restart US Nuclear Plants to Power AI Infrastructure, ABC News Aus. (Oct. 27, 2025) https://www.abc.net.au/news/2025-10-28/google-microsoft-restarting-nuclear-plants-for-ai-power/105941378 (last visited Nov. 15, 2025).
[18] Id.
[19] International Energy Agency, supra note 16.
[20] Id.
[21] Id.
[22] Murtaugh, supra note 13.
[23] Id.
[24] International Energy Agency, supra note 16.
[25] Id.
[26] Samuel Moore, What the CHIPS Act Looks Like Now – A Flurry of Contracts Went to a Range of Projects, IEEE Spectrum (July 28th, 2025).
[27] Id.
[28] Id.
[29] The White House, supra note 1, at 16.
[30] Id. at 16-17.
[31] Id.
[32] Id. at 14.
[33] Id. at 7.
[34] The White House, supra note 1, at 21.
[35] China Bans Foreign AI Chips from State-Funded Data Center, Sources Say, Yahoo Fin. (Nov. 5, 2025), https://finance.yahoo.com/news/exclusive-china-bans-foreign-ai-080808297.html.
[36] Id.
[37] Mukherjee, supra note 4.
[38] Delandro, supra note 6.
[39] Josey Walden, The New Cold War: Rare Earths, AI, and Strategic Competition with China, American Security Project (July 02, 2025), https://www.americansecurityproject.org/the-new-cold-war-rare-earths-ai-and-strategic-competition-with-china/.
[40] Zahava Urecki, Building a Winning Roster: USGS’s Draft Critical Minerals List, Bipartisan Policy Center, (Oct. 8, 2025)
[41] Id.
[42] Id.
[43] Id.
[44] Id.
[45] Natalie Sherman, US and Australia Sign Rare Earth Deal to Counter China’s Dominance, Brit. Broad. Co. (Oct. 20, 2025), https://www.bbc.com/news/articles/cly9kvrdk2xo.
[46] China’s AI Bet Pivots to Power, Metals as Tech Bubble Fears Grow, Bloomberg News (Nov. 14, 2025), https://www.bloomberg.com/news/articles/2025-11-15/china-s-ai-bets-pivot-to-power-metals-as-tech-bubble-fears-grow.
[47] Id.; The Investopedia Team, What is a Pure Play? How They Work, Advantages, and Risks, Investopedia (Aug. 31, 2024), https://www.investopedia.com/ask/answers/04/042904.asp.
[48] Bloomberg News, supra note 46.
[49] Id.
[50] Id.
[51] Id.
[52] Walden, supra note 39.
[53] Id.
[54] Id.
[55] Delandro, supra note 6.
[56] See Sherman, supra note 45; See Andrew Latham, Canada is Quietly Building The Trusted Minerals Base The Pentagon Needs, The Hill (Oct. 24, 2025), https://thehill.com/opinion/national-security/5569516-canada-critical-minerals-us-defense/.
[57] Memorandum of Understanding Between The Government of The United States of America and The Government of the Kingdom of Thailand Concerning Cooperation to Diversify Global Critical Minerals Supply Chains and Promote Investments, The White House (Oct. 26, 2025), https://www.whitehouse.gov/briefings-statements/2025/10/memorandum-of-understanding-between-the-government-of-the-united-states-of-america-and-the-government-of-the-kingdom-of-thailand-concerning-cooperation-to-diversify-global-critical-minerals-supply-cha/.