S-5054-119
Read twice and referred to the Committee on Finance.
Sponsored by Mark Warner (D-VA)
What it does
This bill would do two things. First, it would remove AI data centers from eligibility for "bonus depreciation" — a tax provision that lets businesses deduct the full cost of equipment in the year it is purchased rather than over many years — unless the data center earns a LEED Gold or Platinum green building certification. Second, it would require operators of large data centers (those drawing 25 megawatts or more of power) to annually disclose detailed data on electricity use, water use, greenhouse gas emissions, and backup power to federal agencies or participating states, with civil penalties up to $100,000 per day for knowing violations.
Who benefits
Local communities near large data centers, who would gain access to information about water and electricity consumption affecting shared infrastructure. State and local governments, who would receive data to inform land use and utility planning decisions. Competing businesses that already hold LEED Gold or Platinum certifications and would not lose the tax benefit. Renewable energy companies and green building certification bodies (U.S. Green Building Council). Environmental and public health advocates seeking transparency on industrial resource use. Electric utilities and water authorities, who gain legal protection from liability for disclosures made under the bill.
Who is hurt
Operators of large AI data centers that do not hold LEED Gold or Platinum certification, who would lose access to bonus depreciation and face higher near-term tax burdens. Technology companies — including major cloud providers — that have invested heavily in rapid data center expansion and rely on bonus depreciation to reduce capital costs. Data center operators in states that have not elected to participate, who would report directly to federal agencies with less local flexibility. Smaller data center operators with fewer compliance resources who may face disproportionate administrative costs. Landlords and real estate investors in data center properties subject to new leased-property depreciation rules.
Supporters argue
Supporters argue that AI data centers are among the fastest-growing consumers of electricity and water in the United States — with some projections estimating the sector could account for 8% of U.S. electricity demand by 2030 — and that taxpayers should not subsidize this growth through bonus depreciation without any environmental accountability. They contend that the LEED certification carve-out creates a direct financial incentive for operators to adopt more efficient designs, and that the disclosure requirements fill a genuine information gap: local governments and utilities currently have no standardized way to anticipate the infrastructure demands of incoming data centers, making grid and water planning difficult.
Opponents argue
Opponents argue that removing bonus depreciation from AI data centers singles out a specific industry for disfavored tax treatment, potentially slowing domestic AI infrastructure buildout at a time when the U.S. competes with China and other nations for technological leadership. They contend that the LEED certification standard was not designed for industrial-scale computing facilities and may be poorly suited as a tax eligibility threshold, and that the disclosure regime — with penalties up to $100,000 per day — imposes significant compliance costs and litigation risk on an industry that already faces complex state and federal regulatory requirements, without a demonstrated record that such disclosures produce measurable environmental improvements.