Tom’s Guide used a new interview with Balaji Tammabattula, described as an AI infrastructure engineering expert and chief operating officer of BaRupOn, to examine the widening gap between AI data center developers and the communities asked to host them. The article frames the tension around a familiar infrastructure tradeoff: large technology companies want to move fast on AI compute, while local residents and politicians are pressing for clearer answers on power demand, water use, environmental impact and strain on public services. Tom’s Guide names Microsoft, Meta and Amazon Web Services as major technology institutions investing in the continued national expansion of AI data centers. Tammabattula’s central argument is that the backlash was foreseeable. In his telling, the industry emphasized how quickly new compute capacity could be built, but gave less attention to the supporting systems that make those facilities acceptable to host communities. He told Tom’s Guide that when a large data center arrives needing hundreds of megawatts of electricity, water and other infrastructure, residents naturally ask what the project means for them. The article also points to growing political resistance, saying a coalition of states has placed moratoriums on AI data centers. The provided material does not identify those states or the terms of the moratoriums, so that point should be treated as part of Tom’s Guide’s framing rather than a fully detailed policy map from this cluster. Tammabattula says one misconception in the data center industry is that communities are simply opposed to development. His view is narrower: communities want to understand the tradeoff. According to the interview, the questions include whether electricity rates could be affected, whether the project will compete for local water, what it means for emergency services and roads, and whether children in the area will see actual opportunities from the investment. As an example of a different approach, Tammabattula pointed to LAMP in Texas. He said LAMP is being designed around dedicated behind-the-meter power generation, so the campus produces the power it needs rather than relying on existing community grid capacity. He also said the project is applying similar thinking to water and public infrastructure. On water, Tammabattula told Tom’s Guide that LAMP is planning large-scale rainwater harvesting and has identified the potential to capture more than 500 million gallons of rainwater based on the site’s characteristics. He also said the project is planning a fire station. The broader pitch is that AI infrastructure should not arrive in a community asking what resources it can take, but what infrastructure it can build and leave behind. For operators, the useful signal is that AI infrastructure is becoming less about shells, racks and chips alone, and more about the surrounding civic systems. In this single-source account, speed itself becomes a constraint: projects that move faster than their power, water and community plans may face resistance even when the demand for compute is clear. Who benefits: Developers that can show dedicated power, water planning and public-infrastructure commitments are better positioned in the framework Tammabattula describes. Communities may also benefit if projects are structured to add infrastructure rather than only consume local resources. Who's exposed: Projects that depend heavily on existing grid capacity or local water without clear mitigation plans are more exposed to pushback in this account. Developers that cannot explain effects on rates, roads, emergency services and local opportunity may face harder community negotiations.