As AI workloads soar and hyperscale data centers multiply across the U.S., lawmakers and regulators are confronting a critical question: who should shoulder the rising cost of powering AI-scale infrastructure?

This week, three key developments underscore the growing political and economic tensions around energy consumption and emissions, in the age of AI.

U.S. Senators propose Clean Cloud Act to curb AI and crypto emissions

Senators Sheldon Whitehouse (D-RI) and John Fetterman (D-PA) introduced the “Clean Cloud Act of 2025,” a federal proposal to amend the Clean Air Act and bring crypto miners and AI data centers under direct emissions oversight. The bill would mandate that any facility using more than 100 kilowatts of power annually report its energy mix and carbon emissions — and cut emissions by 11% per year, with a target of net-zero operations by 2035 powered entirely by renewables. Facilities that fail to comply would face penalties starting at $20 per kilowatt, a fee that could rise if targets aren’t met. The legislation would also require operators to submit annual transparency reports on power use and emission reduction strategies. Read more

Takeaway: This is the boldest federal attempt yet to align AI infrastructure with climate goals — but it faces stiff opposition in a Republican-led Senate.

Arizona opens formal review into data center energy fairness

Arizona’s Corporation Commission (ACC) has launched a formal investigation into whether data centers — particularly those serving AI and cloud computing — are paying their fair share of the state’s infrastructure costs. With 129 data centers now operating across the state, many benefiting from favorable utility rates and tax incentives, questions are growing over whether residential and small business customers are effectively subsidizing large-scale operators. The ACC will evaluate new pricing models, including time-of-use rates, “behind-the-meter” options, and infrastructure cost-sharing schemes. According to Chair Kevin Thompson, the goal is to ensure fairness in cost distribution without discouraging future investment in Arizona’s growing digital economy. Read more

Takeaway: Arizona joins a growing number of states rethinking how they price energy for digital infrastructure — especially as grid stress and public scrutiny increase.

West Virginia pushes bill to fast-track data centers, backs coal

West Virginia’s legislature passed HB 2014, a bill that fast-tracks the approval process for new data centers and microgrids by eliminating local oversight and green energy requirements. Under the bill, fossil fuels — including coal — can be used to power AI infrastructure without environmental or zoning reviews at the municipal level. The bill is part of Governor Patrick Morrisey’s broader strategy to make West Virginia a destination for AI and heavy industrial investment by emphasizing low-cost, reliable power. Read more

Takeaway: West Virginia’s bet on deregulation and fossil fuels shows how energy policy is emerging as a competitive tool in the race for AI infrastructure.

Big picture

Across the United States, the explosive energy demands of AI are triggering deep regulatory rifts. Some states are introducing emissions mandates, transparency laws and fair-pricing reviews to make AI development more sustainable and accountable. Meanwhile, others states are opting for deregulation and fossil fuel incentives with the aim of attracting investment. As data centers continue to proliferate and grid loads surge, the fundamental questions of who pays — and how clean that power should be — remain unresolved.