Amazon’s massive Texas AI campus could become the US’s biggest CO₂ polluter

Amazon’s plan to build a huge AI data center campus in Texas is drawing intense scrutiny after new reporting showed its dedicated power plant could become the single largest source of greenhouse‑gas emissions in the United States. State regulators in Texas have approved air permits for a custom natural‑gas facility of around 7.65 GW with 35 turbines, authorized to emit up to 33 million tons of CO₂ per year—more than many coal plants.

The plant is designed to sit alongside a purpose‑built AI campus, supplying the enormous and increasingly steady electricity demand of training and running advanced generative models. Climate advocates say the project crystallizes a broader fear: that the AI boom will lock in decades of new fossil‑fuel infrastructure just as global emissions are supposed to be falling. Local communities and environmental groups argue the plan is fundamentally incompatible with US and corporate climate pledges, while some Texas officials emphasize jobs, tax revenue, and energy “reliability” as justification.

For policymakers worldwide, the Texas controversy is becoming a case study in AI infrastructure decarbonization. It highlights how siting choices, grid rules, renewable‑energy procurement, on‑site gas plants, and transparency about data‑center emissions will shape the true climate cost of AI. As other hyperscalers race to build similar campuses, pressure is likely to grow for binding standards on energy sourcing and lifecycle emissions in AI‑related data‑center projects.