A new study from Rice University finds that carbon capture and storage (CCS) could significantly reduce emissions from the booming AI data center sector. Researchers estimate U.S. data center power capacity could jump from 40 GW in 2025 to 169 GW by 2030, with associated CO2 emissions rising from 90 million to over 404 million metric tons per year. The study suggests natural gas plants with CCS offer a practical near-term solution, especially in states like Texas, Virginia, and Ohio where growth is concentrated. The analysis shows that 34 states have enough saline aquifer storage capacity to hold more than 100 years of projected data center CO2 emissions beyond 2030. By 2030, these aquifers could capture up to 74% of data center emissions, or 299 million metric tons annually. When out-of-state storage is included, over 90% of emissions could be mitigated. The study provides a state-by-state framework for balancing AI infrastructure growth with decarbonization goals, though it notes that CCS is not the only option and that actual deployment depends on policy support and project economics.
