AI Data Centers Spark Local Opposition in U.S. and Korea Amid Power Concerns

As the AI industry scales, electricity has become a critical resource, requiring massive GPU clusters that consume significant power and generate heat, necessitating cooling systems. In the U.S., Gallup found 70% of Americans oppose local AI data center construction, with 48% strongly against it, citing concerns over power usage, water consumption, environmental impact, noise, and rising utility costs. Unlike traditional factories, data centers create limited long-term employment—around 100-200 jobs per facility—with most work occurring during construction.

U.S. data centers consumed about 4.4% of national electricity in 2023, with projections reaching 9.5%-15.3% by 2030. The U.S. Department of Energy warns that grid congestion and connection delays are emerging challenges. Meanwhile, South Korea has passed a special law to streamline AI data center approvals, aiming to avoid reliance on foreign cloud services for domestic AI development.

Korean residents in Gyeonggi Province, Goyang, Incheon, and Seoul have protested proposed data centers over power usage, noise, cooling issues, and fire risks. Energy Economics Research Institute projects that by 2038, data center electricity demand will reach about 4.8% of national targets—low enough not to strain supply but highlighting grid congestion in Seoul as a key issue.

Ulsan is testing a model where AI data centers integrate with local manufacturing sectors like automotive and shipbuilding, aiming to create broader economic ecosystems rather than isolated facilities. The focus should shift from merely building centers to ensuring tangible community benefits: fair cost-sharing for infrastructure upgrades, localized job creation, and reuse of waste heat for district heating or industrial processes.