The Physical Cost of AI: Communities Fight Back Against Data Centers

The rapid rise of artificial intelligence is changing the world, but it is also leaving a heavy mark on the physical environment. As reported in our previous coverage, major technology companies are seeing their carbon emissions rise as they build more data centers—large buildings filled with computers that store data and power AI services. While these facilities are essential for the AI boom, they are now facing a wave of opposition from local communities concerned about their impact on the planet and daily life.

According to The Verge, the fight against these projects is spreading across the United States. Residents are flooding town halls to express worries about rising energy costs, noise, light pollution, and the impact on local water quality. The scale of this opposition is significant. A study from Data Center Watch indicates that the number of active protest groups more than doubled between the end of 2025 and the first three months of 2026. During that same period, protesters successfully blocked or delayed at least 75 projects across the country.

This trend is not entirely new. Years before the current AI boom, Apple faced a lengthy legal battle in Ireland after announcing plans for a large data center in the town of Athenry. Although the company promised to use renewable energy and provide community benefits, local residents raised concerns about noise, traffic, and the impact on wildlife. After years of legal challenges and uncertainty, Apple eventually canceled the project in 2018. This early case set the stage for the challenges tech companies face today as they try to build massive infrastructure in local neighborhoods.

The Guardian reports that tech journalists are now spending more time in the physical world to investigate these claims. Reporters have been visiting sites in Scotland, London, and Wales to see if the promises made by tech companies match the reality on the ground. In one recent investigation, The Guardian found that an expensive AI complex in rural Scotland had misrepresented its plans to use only on-site renewable energy. This kind of "shoe-leather reporting"—a term for doing on-the-ground research—is becoming essential as the industry expands.

There are several reasons for this growing tension. First, the energy demand is massive. The US Energy Information Administration reports that commercial energy demand is expected to surpass residential demand this year, largely because of the AI data center buildout. This demand is predicted to double by 2027. Second, these facilities are some of the most complex structures ever built. As Blake Montgomery, the US tech editor for The Guardian, noted, these buildings are massive and change the landscape of the areas where they are placed. This physical presence creates a "sensory assault" for nearby residents, who must deal with constant noise and light.

Tech companies often claim that their data centers will be sustainable, but experts are questioning these commitments. The Guardian noted that it is not as simple as "waving a magic wand" to make a data center environmentally friendly. There are real constraints, such as the limited capacity of local electricity grids and the need for vast amounts of water to keep computer servers cool. When these physical limits are ignored, it can lead to problems for the surrounding community.

Looking ahead, the future of the AI industry may depend on how companies handle these physical reality checks. Tech companies are finding that they cannot simply build wherever they want without considering the needs of local people. As more communities organize, the ability of tech giants to complete these massive projects will likely face even more hurdles. Whether it is through legal action, petitions, or public protests, residents are making it clear that they expect transparency and accountability from the companies that want to build in their backyards. For now, the "AI boom" is being tested by the very real, physical limits of the world we live in.