A bipartisan coalition of US senators has formally requested that the Energy Information Administration (EIA) publicly disclose electricity consumption data from data centres, citing mounting concerns over AI infrastructure’s impact on national power grids. The letter, sent this week, represents the first coordinated legislative pressure on federal agencies to quantify AI’s energy footprint.
Senators Elizabeth Warren and Josh Hawley led the effort, joined by colleagues from both parties in demanding transparency from the statistical arm of the Department of Energy. The lawmakers argue that the EIA’s current practice of collecting but not publishing data centre energy consumption figures leaves policymakers, utilities, and businesses operating without critical infrastructure planning information.
The timing reflects growing alarm over AI compute demands. Data centres now account for an estimated 4% of total US electricity consumption, according to industry analyses, with projections suggesting this could double by 2030 as generative AI workloads proliferate. Major technology firms have announced plans to invest over $100 billion in new data centre capacity over the next two years, raising questions about grid reliability in regions already facing supply constraints.
The senators’ letter specifically challenges the EIA’s decision to classify detailed consumption data as commercially sensitive, arguing that aggregate regional figures would not compromise individual company operations whilst providing essential transparency for infrastructure investment decisions. The agency has collected monthly electricity usage data from large data centre operators since 2023 but has not made this information publicly available.
This legislative intervention follows mounting pressure from state regulators and utility operators struggling to accommodate unprecedented power demands. Virginia’s Dominion Energy reported that data centre load applications in its service territory have exceeded 10 gigawatts—equivalent to powering 7.5 million homes—with the majority attributed to AI training and inference workloads.
The business implications extend across multiple sectors. Utilities face accelerated capital expenditure requirements to expand generation and transmission capacity, costs that will ultimately flow through to ratepayers. Technology companies may encounter stricter permitting requirements and longer development timelines for new facilities if energy consumption data reveals concentration risks in specific regions.
For data centre operators and their investors, public disclosure could reshape competitive dynamics. Firms demonstrating superior power efficiency metrics may gain advantages in securing permits and power purchase agreements, whilst those lagging on energy performance could face regulatory scrutiny. The real estate investment trusts and infrastructure funds backing data centre development would likely adjust valuations based on energy intensity metrics once disclosed.
Cloud service providers including Amazon Web Services, Microsoft Azure, and Google Cloud—which collectively control over 60% of global cloud infrastructure—have published voluntary sustainability reports, but these lack standardised methodologies and independent verification. Mandatory EIA disclosure would establish consistent measurement frameworks, potentially accelerating efficiency improvements across the sector.
The senators’ letter also references broader concerns about AI’s environmental impact coinciding with corporate net-zero commitments. Several technology majors have acknowledged that AI workloads have pushed their carbon emissions upward despite renewable energy investments, creating tension between growth strategies and climate pledges.
Energy market analysts suggest that transparent consumption data could catalyse innovation in cooling technologies, chip efficiency, and workload optimisation. It may also inform decisions about data centre placement, steering development toward regions with surplus renewable generation capacity rather than those dependent on fossil fuel peaker plants.
The EIA has not yet responded publicly to the senators’ request. The agency operates under Department of Energy oversight, meaning the decision may ultimately require approval from senior Biden administration officials weighing competing priorities around industrial policy, climate goals, and commercial confidentiality.
Market participants should monitor whether the EIA establishes a timeline for disclosure and what level of geographic and temporal granularity the data will contain. State-level energy regulators in Texas, California, and Virginia—the three largest data centre markets—are likely to use any published figures to inform their own permitting and rate-setting processes, creating immediate compliance implications for operators in those jurisdictions.
The bipartisan nature of this inquiry suggests energy transparency around AI infrastructure has emerged as a rare area of cross-party consensus, increasing the likelihood of sustained legislative attention regardless of future electoral outcomes.
