Environmental AI Governance: Divergent Paths to Green AI Systems in the U.S. and China
As artificial intelligence rapidly expands, the United States and China face similar sustainability challenges; however, their responses differ starkly, shaping AI’s long-term environmental footprint.
Current Sustainability Challenges
In Part 1 of the series on environmental AI governance, it was noted that energy and water resources are at risk due to data center overusage, which could lead to competition with local communities for electricity and water. According to polling by the AP-NORC Center for Public Affairs Research, 4 in 10 Americans are “extremely” or “very” concerned about the environmental impacts of AI.
U.S. Policy Response
The U.S. approach to addressing the environmental impacts of data centers has largely been piecemeal. Currently, there are no federal regulations specifically capping energy or water use in data centers; instead, market forces and local state authorities are key players in this arena.
Local states have employed tax exemption incentives to attract large tech corporations, bringing high-tech jobs to their cities. A report by CNBC found that as many as 42 states provide full or partial sales tax exemptions for data center projects or have no state sales tax at all. However, environmental concerns gained prominence only when air pollution and water reserves for citizens were threatened.
Some tech companies are proactively securing cleaner energy sources. For instance, Meta signed a landmark 20-year purchase agreement for the entire output of a 1.1 GW nuclear power plant in Illinois. Conversely, Meta’s expansion in the southern U.S., powered by fossil fuels, presents a dual challenge. Louisiana’s Entergy Corp. proposed building three new gas-fired power plants exclusively for Meta’s data centers, a project now under legal scrutiny for its potential climate impacts, expected to consume three times as much electricity as New Orleans.
Local Grievances and State Responses
Currently, most data center policies are centered around local grievances regarding energy and water overusage. Grassroots community organizers are voicing their concerns regarding corporate negligence. In response, several state policymakers have begun forming frameworks for data center development. States like Texas, California, Michigan, and Minnesota are drafting regulations to limit local environmental impacts.
Some U.S. states are stepping up with incentives for greener data centers. For example, Virginia’s 2024 proposal for Data Center Efficiency stipulates Power Usage Effectiveness (PUE) efficiency targets for tax breaks, while Oregon mandated water reporting for large cooling users in June 2025. Local governments are also employing third-party verification systems like LEED to incentivize sustainable data development.
Challenges Ahead
In conclusion, the U.S. regulatory approach to AI data centers remains a patchwork driven by corporate deals and local politics rather than a cohesive national strategy. Without a unified approach, the AI gold rush risks overwhelming both the power grid and environmental safeguards for sustainable urban development.
China’s Centralized Strategy
In contrast, China is adopting a more centralized, policy-driven approach to manage AI’s environmental footprint. Chinese analysts project that data centers will consume 400 TWh annually, about 3.2% of China’s total electricity supply, with expectations to quadruple in the next decade. Recognizing this surge, Beijing has launched initiatives to dictate the operations of data centers.
A key policy is the “East-West Computing Resources Transmission” (EWCRT), unveiled in 2022, directing the construction of supercomputing and data centers in China’s western and northern regions. These areas, including Sichuan, Inner Mongolia, Gansu, and Ningxia, benefit from cooler climates and abundant renewable resources, facilitating greener energy development.
Environmental Impact and Future Goals
By relocating AI data centers closer to renewable energy sources, China aims to reduce water usage for cooling and improve the carbon profile of its AI infrastructure. Research suggests that the East-West Data Project could reduce 1,500 metric tons of CO2 emissions between 2020 and 2050, equivalent to cutting 30,000 barrels of oil.
Moreover, Chinese authorities have mandated the use of clean energy for data centers, further integrating its energy policy with AI industrial strategies. In March 2025, the National Development and Reform Commission issued guidelines urging big data hubs to increase their renewable electricity share, with operators now facing regulations requiring them to purchase a specified percentage of green power credits.
Provincial governments are also rolling out complementary policies, such as Inner Mongolia offering incentives for data centers to partner with local wind and solar farms. Additionally, the Shanghai government and China Telecom collaborated to build the world’s first underwater data center, the Lin-gang Special Area, which combines renewable energy with deep-sea cooling for enhanced efficiency.
Comparative Analysis
While the U.S. relies on market-led innovation and decentralized governance, China pursues state-led coordination and strategic regional planning. The effectiveness of China’s initiatives in offsetting emissions growth remains uncertain, especially as its grid is still approximately 60% coal-fired. The surge in demand for data centers could lead to increased CO2 emissions unless renewable energy scales rapidly.
Balancing AI ambitions with environmental constraints poses a significant challenge. The differing models—bottom-up adaptation in the U.S. versus top-down orchestration in China—highlight their respective strengths: American flexibility fosters technological experimentation, while Chinese centralization enables systemic efficiency.
Future Collaborations
Both nations should ultimately focus on reconciling AI’s rapid growth with environmental stewardship. In August of this year, China’s Premier proposed the establishment of an international organization to foster strategic cooperation in artificial intelligence. This interest in partnership must be reciprocal.
The next article will introduce a collaborative policy framework that could merge U.S. and China’s respective advantages and strengths to address shared resource challenges. By establishing a bilateral “Green Compute Accord,” the U.S. and China could transform competition into co-innovation, creating globally aligned pathways to decarbonize the digital backbone of AI while ensuring resource and energy security for both economies.
