US backlash grows against AI data centers over power and inflation fears

This story captures an emerging political backlash against the physical footprint of the AI boom: power-hungry data centers, higher electricity prices, and supply-chain stress in memory chips.([m.cn.investing.com](https://m.cn.investing.com/news/stock-market-news/article-3147477?ampMode=1)) When figures as ideologically distant as Bernie Sanders and Ron DeSantis converge on skepticism about new AI data centers, it suggests compute expansion is bumping into hard social and infrastructure limits. For the AGI race, that’s a warning light. The field’s default assumption has been that you can always add more GPUs somewhere; this article shows that “somewhere” now has voters and utility regulators who may say no.

If moratoria or strict siting rules proliferate in key US grids like PJM, hyperscalers could face delays or higher costs in bringing new capacity online. That would slow the cadence of ever-larger training runs and force more aggressive efficiency work at the model, hardware, and algorithmic levels. It might also push more compute buildout overseas, further fragmenting the geography of AGI development. At the same time, political pressure could accelerate investment in grid upgrades and low-carbon generation specifically justified by AI demand, which in the medium term would actually expand the sustainable compute envelope.