US Data Centers' Power Share Could Triple by 2030, Driving Rate Pressure
New LBNL projections prompt Columbia researchers to argue grid-enhancing technologies must be deployed faster to slow rate increases already running above inflation.
Columbia University researchers said Tuesday (2026-06-23) that grid-enhancing technologies and expanded demand response could ease electricity price pressures tied to data center growth, a finding that arrives as fresh government projections show demand rising faster than previously forecast.5
Lawrence Berkeley National Laboratory researchers said in a report released last week (week of 2026-06-15) that data centers could account for 9.5% to 15.3% of all US electricity consumption by 2030, up from 4.7% in 2024. The upper end of that range would represent a more than tripling of the current share inside six years.5
Those numbers carry direct implications for electricity bills. Investor-owned utilities sought $18 billion in rate increases last year, the most since the mid-1980s, the Columbia researchers said, citing the LBNL report. Regulators approved 66% of the dollar value of those requests. Electricity prices rose 6% in nominal terms in 2025, more than twice the rate of inflation.5
The LBNL projections represent a meaningful upward revision from the researchers' own earlier work. Two years ago, the same team put the 2028 data center electricity share at 6.7% to 12%. The new 2030 upper bound of 15.3% implies that aggregate infrastructure build-out is outpacing efficiency gains across hardware generations.5
The demand trajectory is visible in grid planning. Data centers now account for about half of US incremental electricity demand growth, according to the IEA's global energy assessment. Grid engineers, utility executives, and regulators describe a permitting and supply chain backlog that cannot match the pace of hyperscaler buildout, with roughly 5,000 miles of high-voltage transmission needed in the near term.3
In ten states, electricity regulators are elected by popular vote, adding political exposure to an approval process already under scrutiny. Rate increases are accumulating in dockets across the country, and the political pressure that comes with them is likely to affect how regulators balance utility capital returns against customer cost in future filings.4
The Columbia report identified a structural problem in utility regulation. Under traditional rate-of-return frameworks, utilities earn roughly 9% to 10% on capital investments. "This regulatory structure rewards capital deployment more than system optimization," the Columbia researchers said. Grid-enhancing technologies including advanced conductors, dynamic line ratings, and topology optimization could raise throughput on existing lines but do not automatically carry the same regulated return as new build. Analysis in the Proceedings of the National Academy of Sciences estimated that replacing existing transmission lines with advanced conductors could yield $180 billion in savings by 2050.5
The IEA adds a hardware dimension. A single advanced server rack could carry peak demand equivalent to 65 households by 2027 — concentrating load in ways that grid infrastructure was not built to absorb.2
Hyperscalers have not resolved the emissions math. Google's direct emissions jumped nearly 50% as it expanded AI infrastructure. Microsoft has acknowledged falling short of earlier clean-power commitments, with both companies citing construction pace as the binding constraint on decarbonisation timelines.1
The immediate test is rate cases. Utilities are seeking increases at the highest rate since the mid-1980s, and regulators approved roughly two-thirds of the dollar value requested last year. Whether demand response and grid-enhancing technology can slow the pace of future filings, or whether the structural capital build continues at scale, will determine how much of AI infrastructure's cost ends up on residential and industrial power bills.5