Data centers’ electricity use could quadruple by 2035

Data centers are expected to use one-fifth of electricity generated in the U.S. by 2035, according to a BloombergNEF report cited by TechCrunch. That would be four times today’s level. The report also says new data centers built through 2033 could add global electricity demand nearly equal to India’s annual usage.

The immediate takeaway: data center growth is no longer just a cloud infrastructure story. It is becoming a power planning issue for utilities, grid operators, policymakers, and AI companies.

BloombergNEF predicts a surge in AI compute will push data center capacity to nearly 200 gigawatts over the next decade. Nearly half of that capacity is expected to be devoted to AI training and inference. Most of that AI-related power demand is projected to remain concentrated in the U.S.

By 2033, the U.S. is expected to host 64% of AI chips by power demand. That concentration means decisions about where AI infrastructure gets built could have outsized effects on local and regional grids.

BloombergNEF expects many new data centers to connect to electrical grids that are already strained. The PJM Interconnection, which spans Virginia to Illinois, is projected to see 34% of its electricity go to data centers, while ERCOT, covering most of Texas, would devote 22% of generating capacity.

TechCrunch notes that PJM has struggled with connection requests from large generators and large loads, pausing applications for new sources to connect to the grid for four years before reopening the queue in April. The article also reports that American Electric Power has threatened to pull out of the interconnection, electricity prices are up 76% over the past year, and data centers represented 38% of charges in PJM’s most recent capacity auction.

BloombergNEF’s 2035 electricity demand estimate is 83% higher than what the consultancy predicted in December, according to TechCrunch. Other forecasts have also risen: EPRI has more than doubled its 2024 estimate, while S&P’s forecast rose by more than a third between October and April.

The practical implication is clear: energy planning around AI and data centers is changing quickly. Grid operators and data center developers may need to treat power availability, connection queues, and local capacity constraints as core project risks rather than late-stage details.