BNEF's 42 GW scenario gap leaves gas and storage markets without a demand anchor
BloombergNEF's two US data center power forecasts diverge by 42 GW for 2030, a spread wide enough to swing decisions on gas plants, storage and grid investment.
BloombergNEF analysts said on Wednesday (2026-07-22) that their two scenarios for U.S. data center electricity demand by 2030 differ by 42 gigawatts, a spread they described as highlighting mounting uncertainty about the buildout's scale even as construction accelerates. The gap is roughly equivalent to New York state's peak load, and it is widening because actual development keeps outrunning projections.8
Power markets have moved well past debating whether data centers reshape demand. The argument now is over pace and geography — and the 42 GW range between BNEF's cases makes positioning on either side uncomfortable. Electricity demand from data centers could reach 9% to 17% of U.S. supply by 2030, or up to 790 terawatt-hours, against roughly 4% today (2026-05-19), according to the Electric Power Research Institute.3
Almost all U.S. regions ended 2025 with more data center capacity than BNEF had anticipated. Texas showed the largest gap between forecast and actual buildout. The firm's analysts said they tracked about 100 GW of new project capacity added in the last year alone, BloombergNEF Senior Associate Nathalie Limandibhratha said on Wednesday (2026-07-22).8
The numbers get larger still from other quarters. Data centers could add roughly 125 GW of U.S. electric load over the period, BofA analysts said — nearly double what BloombergNEF forecast back in December, and more than BNEF's own analysts expect the grid to accommodate, Canary Media reported on Friday (2026-07-24).7,8
Battery storage companies are watching the buildout closely. Fluence is engaged in over 30 GWh of data center-related projects globally, with a meaningful portion in the U.S., CEO Julian Nebreda said.3 But the sector faces hard constraints: lengthy grid connection queues and a supply chain heavily dependent on China are hampering the industry's ability to scale, Reuters reported on Monday (2026-05-18).3
The underlying storage market is already moving fast. The U.S. added a record 57.6 GWh of new battery storage capacity in 2025, according to the Solar Energy Industries Association, bringing total deployed capacity to 166.1 GWh. The group projects annual deployments reaching 110 GWh by 2030, with a significant share driven by data center demand.3
Data centers now account for more than 1% of global electricity use, according to the IEA. In 2025, nearly 40% of all power demand from data centers globally came from the U.S., more than any other country including China, Canary Media reported.2,6 AI's share of U.S. data center energy consumption sits at 10-20% currently and will likely increase significantly, according to analyst Porter, cited by TIME.2
That trajectory carries implications for fossil fuels. BloombergNEF found the data center expansion required to support AI is expected to keep fossil fuels in use for longer, Dezeen reported.1 The buildout is also straining generation planning in ways that look unlikely to resolve quickly.
The pressure is not confined to the U.S. In Asia Pacific, hyperscalers are prioritising scalability and delivery certainty in the race for power, with pipeline projects frequently measured in the hundreds of megawatts and some approaching gigawatt scale, according to a report flagged on Tuesday (2026-05-26).4 In China, grid operators are resisting plans to boost renewable power for AI data centers, concerned that peak demand at those facilities is difficult to forecast and would raise risks for power firms, Oilprice reported on Monday (2026-06-22).5
Back in the U.S., the 42 GW gap between BNEF's two scenarios translates into the difference between roughly 25% and 40% growth in data center electricity demand by 2030. Gas plant construction decisions, storage procurement volumes, and power prices in Texas and Virginia — the two most active markets — all sit inside that range of outcomes.8
Texas, which has already beaten BNEF's forecasts by the largest margin of any U.S. region, is the clearest early signal. If its buildout continues to outpace projections through the remainder of 2026, BNEF's higher scenario starts to look like the base case. Grid interconnection queues remain the constraint most capable of forcing a slowdown, but every quarter of construction data published so far has pointed in the same direction.8