Data Center Tax Accountability and Disclosure Act of 2026
- Sponsor
- Sen. Warner, Mark R. [D-VA]
- Committees
- Finance Committee (primary)
- Last action
- Jul 21, 2026
Bottom line
The bill aims to reduce tax subsidies for certain AI data centers and increase transparency regarding the environmental footprint of large data centers by mandating comprehensive reporting on their energy and water consumption.
What it actually does
This bill amends the Internal Revenue Code to exempt AI data centers from bonus depreciation, thereby reducing a tax incentive for their development. It also mandates that operators of large data centers (25 megawatts or more) provide detailed annual disclosures to federal agencies or electing states regarding their electricity and water usage, efficiency metrics, greenhouse gas emissions, and backup power. The bill includes provisions for federal data center disclosures and establishes civil penalties for non-compliance.
Proponents argue
Supporters argue the bill addresses the significant environmental footprint of rapidly expanding AI data centers by removing a tax subsidy that encourages their growth, particularly for less efficient facilities. They contend that mandatory disclosures will provide crucial data for policymakers and affected communities to understand and mitigate the strain on local resources and infrastructure, promoting greater accountability and sustainability in the data center industry.
Opponents contend
Opponents might argue that exempting AI data centers from bonus depreciation could stifle innovation and investment in a critical emerging technology sector, potentially putting the U.S. at a competitive disadvantage. They may also contend that the disclosure requirements impose an undue regulatory burden on businesses, increasing operational costs that could be passed on to consumers or hinder economic development.
The bill is moderately complex due to tax code amendments and detailed reporting requirements, but its length makes it manageable for a thorough review within a reasonable timeframe.
Section 2(a) of the bill, amending Section 168(k) of the Internal Revenue Code of 1986.
Exclusion of AI Data Centers from Bonus Depreciation
This provision modifies federal tax law to prevent 'AI data centers' from claiming bonus depreciation for property placed in service after the bill's enactment. Bonus depreciation allows businesses to immediately deduct a large percentage of the cost of eligible property, accelerating tax savings. The bill defines an AI data center as a structure primarily for IT equipment, with at least 20% used for developing or operating artificial intelligence.
Supporters argue
Supporters argue this change removes a federal subsidy that may inadvertently encourage the rapid, unchecked expansion of AI data centers, which are known for high energy and water consumption. They contend it levels the playing field and encourages more responsible investment.
Critics contend
Opponents might argue that removing bonus depreciation could deter investment in AI infrastructure, slowing technological advancement and economic growth in a critical sector. They may also claim it unfairly targets a specific industry.
Tradeoffs
The provision balances the desire to promote AI development through tax incentives against concerns about the environmental impact and resource strain of AI data centers.
Section 2(a)(2)(B), adding new paragraph (11)(B) to Section 168(k) of the Internal Revenue Code.
LEED Platinum/Gold Certification Exemption from Bonus Depreciation Exclusion
This provision creates an exception to the bonus depreciation exclusion for AI data centers that achieve LEED Green Building Rating System certification at the Platinum or Gold level. It also allows the Secretary of Energy and the EPA Administrator to establish or identify equivalent or superior alternate standards for this exclusion.
Section 2(b), adding new subsection (o) to Section 168 of the Internal Revenue Code.
Depreciation Rules for Leased AI Data Centers
This section specifies that for AI data centers subject to a lease, depreciation deductions will be determined according to existing Treasury regulations (26 CFR 1.167(a)-11(e)(3)(iii)) as of the date of enactment. This regulation generally deals with the depreciation of leased property, often impacting who claims the deduction (lessor or lessee) and how it's calculated.
Section 3(b) and 3(c).
Mandatory Electricity and Water Usage Disclosures for Covered Data Centers
This provision mandates that operators of 'covered data centers' (those with a power demand of 25 megawatts or more) submit initial disclosures before commencing operations and annual reports thereafter. These reports must detail monthly water and electricity usage, sources, efficiency metrics (Water Usage Effectiveness, Power Usage Effectiveness), greenhouse gas emissions, backup power details, and property setback requirements.
Section 3(e).
Director of OMB to Establish Federal Data Center Disclosure Standards
This section directs the Director of the Office of Management and Budget (OMB) to establish standards for compiling and submitting information on dedicated federal and dual-use data centers. These standards must promote accountability and transparency while protecting national security, including exemptions for sensitive information. Dual-use data centers will also be subject to disclosure requirements under subsections (b) and (c) as prescribed by OMB.
Section 3(f).
Invalidating NDAs that Prevent Required Disclosures
This section explicitly voids any nondisclosure or confidentiality agreements that would prevent, limit, delay, or penalize the disclosure of information required by this Act to the Secretary of Energy, EPA Administrator, or an electing State. It also preempts state contract law that would uphold such restrictive agreements and protects utilities and state agencies from liability for compliance.
Section 3(g).
Civil Penalties for Non-Compliance with Disclosure Requirements
This section establishes civil penalties for data center operators who fail to submit required disclosures or reports. Negligent violations can incur penalties of up to $50,000 per day, while knowing violations or submitting false/misleading information can result in penalties up to $100,000 per day. Electing States are also authorized to impose fines and conduct enforcement.
Defines an 'AI data center' as a structure dedicated to IT equipment, including 'not less than 1 graphics processing unit,' and at least 20% of which is used for developing or operating artificial intelligence.
Section 2(a)(2), adding new paragraph (11)(A) to Section 168(k) of the Internal Revenue Code.
Why it matters:Standard legislative practice for defining terms, but the specific technical threshold could be easily missed.
Case for: Supporters would argue that including the GPU requirement ensures that the tax exemption targets facilities genuinely involved in AI processing, which are typically the most energy-intensive.
Case against: Critics might argue that a minimum of '1 graphics processing unit' is an extremely low bar, potentially capturing many general-purpose data centers that only minimally engage in AI, thus broadening the scope of the tax exemption removal beyond truly dedicated AI facilities.
Estimated impact: This definition determines which data centers lose bonus depreciation. A low GPU threshold could mean a wider range of data centers are affected than might be immediately apparent.
Defines an 'electing State' as one that chooses to receive disclosures and reports directly from data center operators and then submits an annual compilation of this information to the Secretary of Energy and EPA Administrator.
Section 3(a)(11) and Section 3(d)(1).
Why it matters:Standard legislative practice for defining roles and responsibilities in a federal-state partnership.
Case for: Supporters would argue this allows states to take a more active role in monitoring local data center impacts, potentially leading to more tailored local policies, while still ensuring federal oversight through the compilation requirement.
Case against: Critics might argue it creates an uneven regulatory landscape, where data centers in electing states face different reporting pathways, potentially leading to inconsistencies or additional administrative burdens for multi-state operators.
Estimated impact: This mechanism decentralizes the initial data collection, potentially empowering states but also adding a layer of complexity to federal data aggregation.
Explicitly states that 'To the extent that any provision of State contract law would render enforceable a provision of a nondisclosure agreement, confidentiality agreement, or similar contract or agreement prohibited or rendered unenforceable under this subsection, such provision of State law is preempted by this Act.'
Section 3(f)(4).
Why it matters:Often, preemption clauses are placed within the specific sections they relate to, rather than in a general preemption title.
Case for: Supporters would argue this is essential to ensure the federal disclosure requirements cannot be undermined by state-level contract enforcement, guaranteeing the transparency goals of the bill.
Case against: Critics might argue this is an overreach of federal power into areas traditionally governed by state law, potentially setting a precedent for federal interference in private contracts beyond the scope of this bill.
Estimated impact: Ensures the enforceability of the bill's anti-NDA provisions, potentially overriding state legal frameworks.