US Representative Ro Khanna is preparing federal legislation that would place new responsibilities on artificial intelligence data centers as communities push back against rising electricity demand, new power plants and large infrastructure projects.
Bill Would Give Communities a Larger Role
Khanna represents a Silicon Valley district, but his proposal signals that support for AI innovation may increasingly depend on protections for the areas hosting its infrastructure.
CNBC linked the proposed legislation to an intensifying dispute over whether technology companies should build or finance dedicated power plants for their data centers instead of placing additional costs and pressure on existing electricity systems.
TechBuzz reported that the draft could require data centers to show that their electricity comes from sustainable sources, provide communities with greater input into where facilities are built and share more economic benefits with host regions.
The complete legislative language has not yet been released, leaving questions about how those protections would be enforced and which facilities would be covered.
Electricity Demand Drives Local Resistance
AI data centers contain thousands of high-performance processors used to train models and respond to users. Their electricity needs can equal those of entire communities, particularly when several facilities are concentrated in one region.
Khanna’s planned bill arrives as the rapid construction of AI infrastructure creates growing concern over household utility bills and the power plants needed to supply large computing campuses.
Microsoft, Google, Amazon and Meta have collectively announced more than $200 billion in data center investments through 2027, based on recent company earnings calls. The publication noted that local grids were not originally designed to absorb such large and rapidly arriving loads.
Residential electricity rates in Virginia’s Loudoun County, which hosts the world’s largest concentration of data centers, have increased by 18% since 2023. Local officials have begun rejecting some new permits despite the tax revenue generated by the industry, according to the report.
Michigan Primary Adds Political Pressure
The data center debate is also entering election campaigns.
CNBC placed Khanna’s proposal within a widening political fight over AI power infrastructure, as candidates respond to voters concerned about energy prices and nearby construction.
TechBuzz reported that Abdul El-Sayed endorsed state and local moratoriums on new data center construction after winning Michigan’s Democratic Senate primary on Tuesday. The publication described him as the first major political candidate to make opposition to new data centers an explicit campaign position.
The shift is notable because states previously competed to attract data centers through tax exemptions, infrastructure support and other incentives. Growing concern about electricity rates, water demand, noise and limited permanent employment is now changing that political calculation.
Tech Companies Search for New Power Sources
Technology companies have responded by investing in renewable power, grid upgrades and alternative energy systems.
The debate increasingly centers on whether AI developers should secure and pay for the additional generation required by their facilities.
Companies are examining dedicated solar farms, small modular nuclear reactors and other power arrangements as they compete for limited grid capacity. The report noted that OpenAI CEO Sam Altman has publicly discussed the need for major energy breakthroughs to sustain continued AI development.
Khanna’s proposal would not impose the nationwide construction moratorium sought by some AI critics. Instead, it appears intended to establish federal protections around who pays for the infrastructure, how communities participate in siting decisions and what responsibilities developers carry when their projects reshape local energy systems.
The bill’s prospects in Congress remain uncertain, but its introduction would mark an important change in Washington’s AI debate. The issue is no longer limited to how models use data or produce content. Lawmakers are now confronting whether companies can physically build the computing infrastructure they want without transferring its financial and environmental costs to the public.
