Key Takeaways

  • H.R. 10102 would tax electricity used by data centers exceeding 1 MW at 1 cent per kilowatt-hour, including power generated onsite.
  • The proposed tax could raise about $1.76 billion annually, divided equally among five federal funds.
  • Critics argue the tax would generate public revenue but would not resolve how utilities assign data center-related grid costs.

Rep. Andrea Salinas, D-Ore., has introduced legislation that would place a federal excise tax on electricity consumed by a broad range of data centers, from enterprise facilities to the largest AI computing campuses.

The Data Center Community Reinvestment Act of 2026, or H.R. 10102, would amend the Internal Revenue Code to charge qualifying facilities 1 cent for every kilowatt-hour they consume. A data center would qualify if its maximum rated power capacity or total peak power load exceeds 1 MW.

Crucially, the tax would cover electricity generated onsite as well as power purchased from utilities or retail suppliers. Behind-the-meter generation, therefore, would not provide an escape route. Amazon Web Services, Microsoft Azure, and Google Cloud operate numerous facilities above the threshold, but the measure would also reach smaller colocation, institutional, and corporate data centers.

Salinas said communities should receive a larger share of the benefits associated with expanding digital infrastructure. “Every community deserves to drive a hard bargain with data center developers, and at the end of the day, every community deserves to share in the benefits from data center growth,” she said in a press statement.

The financial effect becomes substantial at hyperscale. Because the proposed tax applies a flat rate to total electricity consumption, facilities with hundreds of megawatts of capacity operating continuously could face millions of dollars in annual liabilities.

Electricity is already one of the largest recurring expenses in data center operations. A 1-cent-per-kilowatt-hour federal charge would meaningfully increase the overall energy component depending on the underlying local electricity rates.

That added cost could influence site selection when two markets offer similar economics. Still, industry analysts note that access to deliverable power, transmission capacity, generation supply, interconnection timing, and development certainty may outweigh the tax for major AI projects. A slightly cheaper tax environment has limited value if a campus cannot secure hundreds of megawatts on schedule.

The estimated $1.76 billion in annual proceeds would be divided evenly, providing roughly $352 million apiece to the Housing Trust Fund, Land and Water Conservation Fund, Hazardous Substance Superfund, Highway Trust Fund, and a newly created Energy Technology Trust Fund. The new energy fund would support federal loan guarantees for clean energy, advanced nuclear, and grid infrastructure projects.

The policy arrives amid a steep rise in computing demand. Lawrence Berkeley National Laboratory has documented U.S. data center electricity consumption of roughly 176 to 183 TWh during 2023 and 2024, equal to about 4% to 4.5% of national electricity use. Globally, the International Energy Agency reports that data centers consumed about 415 TWh in 2024, with demand projected to reach roughly 945 to more than 1,000 TWh by 2030, largely driven by AI workloads.

But does a flat consumption charge accurately reflect the burden a particular facility creates? Industry analysts argue it does not. Two 500 MW campuses may use identical amounts of electricity while creating very different transmission, substation, generation, and capacity requirements. Their effects also depend on location, peak-hour usage, operational flexibility, and willingness to accept curtailment.

Consumer advocates similarly note that the proposal would raise money for public programs without directly fixing how data center expansion affects existing electricity customers. Pointing to PJM capacity-market and transmission-planning rules, they have called for stronger large-load tariffs, clearer assignment of transmission costs, and options requiring data centers to supply generation or accept curtailment.

Operational efficiency could soften the tax exposure. ISO 50001 energy-management practices and ASHRAE thermal guidelines can help operators track consumption, improve cooling, and reduce avoidable load. Yet efficiency alone does not answer the cost-allocation question, as megawatt requirements and grid impacts remain highly variable based on local infrastructure limits.

H.R. 10102 remains at the introduced stage following its referral to the House Ways and Means Committee, Energy and Commerce Committee, and Science, Space, and Technology Committee. If enacted, its amendments would apply to electricity used and taxes received after enactment.