
US data center power demand will more than double by 2030 as AI clusters strain grid capacity, with interconnection queues stretching past four years in some regions.
US data center power demand will more than double between 2026 and 2030, pushed by high-density AI workloads. The projection comes from utility load forecasts and industry capacity plans that grid operators have folded into their long-term planning documents.
The growth is not uniform. Regions with the deepest concentration of existing data center capacity – northern Virginia, parts of California, and the Pacific Northwest – are already seeing interconnection queues stretch past four years. In some PJM zones, the wait for a new grid connection now exceeds the expected construction timeline of the data center itself.
What makes the current wave different from earlier build-outs is power density. A single AI training cluster can draw 80 to 100 megawatts, roughly the same as 20,000 homes. Previous cloud expansions ran on racks with 4 to 8 kilowatts per rack. AI clusters push past 40 kilowatts per rack, which requires liquid cooling and substation-level upgrades that distribution utilities rarely have on hand.
Tech companies have responded by signing long-term power purchase agreements with wind and solar farms. Renewable generation adds its own timing mismatch. A solar farm takes two to three years to permit and build. A grid interconnection for that same solar farm can take five. The result is a growing gap between the date a data center lease is signed and the date the power actually arrives.
Some utilities are dusting off plans for new natural-gas generation that had been shelved during the renewable build-out of the last decade. State regulators have drawn mixed conclusions. In Georgia, regulators approved a gas plant expansion partly justified by data center load projections. In California, similar proposals have drawn opposition from clean-energy advocates who argue the load could be met with storage and demand response.
Data center operators, for their part, are investing in on-site backup generation. Most units are natural gas or diesel. Some sites add battery storage. Those backup units are meant for emergency use. In several constrained markets they have been dispatched during peak demand events, drawing scrutiny from air-quality regulators.
The doubling of power demand will require a corresponding expansion in generation and transmission capacity. That process, from planning to operation, typically takes five to ten years. The data center industry's ambitions are running ahead of that timeline.
Prepared with AlphaScala editorial tooling from the source reporting linked above. Indexable analysis may include a cited Alpha Score value. Publishing checks screen each story before release. Educational coverage, not personalized advice.