Virginia data-center demand highlights US grid constraints as AI loads grow
DataNews Editorial Team
Schneider Electric executive vice president Bin Lu said rising electricity demand from data centers and AI is becoming a constraint on US economic growth, with grid capacity increasingly concentrated in specific markets. Virginia illustrates the issue: data centers accounted for around 25% of statewide electricity demand in 2024, according to the TechInformed report.
- Virginia data centers accounted for around 25% of statewide electricity demand in 2024.
- EPRI projects that data centers could represent up to 57% of Virginia electricity demand by 2030.
- Virginia imported 36% of its total electricity supply in 2023.
- Schneider Electric’s Bin Lu said US grid constraints are encouraging interest in on-site generation and microgrids for data-center supply.
- Power deliverability and grid interconnection could become stronger differentiators in constrained data-center markets.
Bin Lu, executive vice president of Schneider Electric’s Power Products Division, said in an interview with TechInformed that electricity availability is becoming a limiting factor for US economic growth as electrification, data centers and AI increase demand.
The challenge is not solely the aggregate volume of data-center consumption, but its concentration in individual markets and the difficulty of adding generation and transmission capacity where demand is emerging. Lu said these constraints are pushing data-center operators toward on-site generation and microgrid configurations, although the report did not identify specific projects, capacities or deployment timelines.
Virginia is presented as the clearest example of concentrated demand. Data centers represented around 25% of statewide electricity demand in 2024. EPRI projects that their share could reach as much as 57% by 2030. The state also imported 36% of its total electricity supply in 2023 and was described as the largest net recipient of interstate electricity in the US.
The report states that data centers of all types, including AI facilities, account for less than 3% of energy consumption overall, although the geographic scope of that figure is not specified. The contrast with Virginia underscores how system-level averages can obscure acute local grid constraints.
For developers and operators, the implication is that power deliverability and interconnection availability could increasingly determine project competitiveness in constrained US markets. This may elevate the role of energy planning during site selection, alongside potential use of local generation and microgrid solutions. These are analytical implications rather than a forecast of specific deployments.
For AI data-center development, access to deliverable power and grid interconnection is becoming as consequential as capital availability and compute equipment. In constrained markets, this could increase the importance of local generation, microgrids and early-stage energy planning in site selection.
- — / September 24, 2026 · EnglishSource →
This article was prepared by the DataNews editorial team based on the sources listed above.