Carbon capture and storage could become a key tool for limiting the climate impact of rapidly expanding data centers in the United States, according to a new study published in Energy & Fuels.
Projected growth in power demand
The analysis predicts that the total power capacity needed for U.S. data centers will rise from roughly 40 gigawatts in 2025 to 169 gigawatts by 2030. This more than fourfold increase reflects the surge in artificial intelligence workloads that require massive computing resources.
If emissions are not curbed, the carbon dioxide released by fossil‑fuel generators supplying electricity to these facilities could climb from about 90 million metric tons per year in 2025 to over 404 million metric tons by 2030.
How carbon capture could fit the picture
Researchers Hon Chung Lau, an adjunct professor at Rice University, and Steve C. Tsai, an energy transition consultant, examined publicly announced data‑center projects, estimating power needs, local electricity mixes, and the feasibility of storing captured CO₂ in underground saline aquifers.
Related: Nature’s Sights and Sounds Boost Mental Health
The study highlights several states—Texas, Virginia, Pennsylvania, Ohio, Arizona, Colorado, Utah and Illinois—as hotspots for data‑center expansion. Texas alone may require an additional 25 gigawatts of capacity by 2030.
Because data centers need reliable, round‑the‑clock electricity, the authors argue that natural‑gas combined‑cycle plants equipped with carbon‑capture technology could provide a practical near‑term solution. Natural gas is abundant, burns cleaner than coal, and many of the identified growth regions sit above saline aquifers suitable for long‑term CO₂ storage.
According to the report, 34 states possess enough saline‑aquifer capacity to hold more than a century’s worth of projected data‑center emissions beyond 2030. In 2025 those aquifers could accommodate an estimated 59 million metric tons of CO₂—about 66 % of the sector’s emissions that year. By 2030 the storage potential grows to roughly 299 million metric tons, covering about 74 % of anticipated emissions.
When out‑of‑state storage options are factored in, the researchers estimate that over 90 % of data‑center‑related CO₂ could be mitigated through capture and storage technologies.
Most of the emissions could be captured.
Related: Edwards Sees Boost in Heart Valve Sales
While the numbers are encouraging, the analysis rests on several assumptions. The study only counted facilities with publicly disclosed power requirements, and it treated unspecified power sources as draws from the state grid, assuming each state’s energy mix will stay constant through 2030. Those simplifications likely make the projections conservative.
It is worth noting that the geological capacity identified does not guarantee immediate deployment. Building the necessary capture infrastructure, securing financing, and managing regulatory approvals could take years, and market conditions may shift as renewable energy costs continue to fall.
Lau emphasizes that carbon capture is not the sole answer but points to a tangible option where the geology aligns with data‑center growth. “This does not mean carbon capture is the only solution,” he said, “but it does show that the geology exists to make a meaningful impact, especially in states where data center growth is strongest.”
The authors suggest their state‑by‑state framework can help policymakers and industry leaders balance digital expansion with climate objectives. By pinpointing locations where emissions are likely to rise and where underground storage is viable, the study offers a roadmap for integrating low‑carbon power sources into the evolving AI economy.