Skip to main content
v2026.11,986 entries · CC-BY 4.0

Editorial · CASRAI · AI and ML research outputs

Senators Introduce AUDIT AI Act to Require Disclosure of Off-Balance-Sheet AI Compute Leases

US Senators introduced the AUDIT AI Act requiring SEC disclosures for off-balance-sheet AI compute leases, SPV debt structures, and board risk governance.

Published 20 Nov 2026· 7 minute read
Street lamps and glass facade of the U.S. Securities and Exchange Commission headquarters building in Washington, D.C.
Photo: David (Flickr user: dbking) / Wikimedia Commons, CC BY 2.0 (cropped). Source · Licence

A bipartisan group of United States Senators introduced legislation on November 19, 2026, directing the Securities and Exchange Commission to mandate disclosure of off-balance-sheet artificial intelligence compute leases, debt arrangements, and corporate governance policies. The bill alters how academic institutions and research offices evaluate commercial cloud partnerships, compute consortium financing, and frontier model dependencies.

Bipartisan Senate Coalition Introduces the AUDIT AI Act

On November 19, 2026, U.S. Senators Mark R. Warner (Democrat of Virginia), John Kennedy (Republican of Louisiana), Angela Alsobrooks (Democrat of Maryland), and Mike Rounds (Republican of South Dakota) introduced the Advancing Uniform Disclosure of Investments and Transparency of Artificial Intelligence Act, titled the AUDIT AI Act. The legislation directs the Securities and Exchange Commission (SEC) to require public companies to reveal material financial commitments, equipment leases, and governance practices governing artificial intelligence.

Lawmakers raised concerns that technology firms are shifting billions in capital expenditures into unconsolidated partnerships and debt vehicles. According to reporting by Social News XYZ, the measure creates standardized reporting obligations for public non-financial companies across annual Form 10-K, quarterly Form 10-Q, and public offering statements.

Senator Warner, a senior member of the Senate Committee on Banking, Housing, and Urban Affairs, stressed the risk of hidden liabilities. “AI has enormous potential to transform our economy, but the trillions of dollars pouring into this technology come with real financial risks,” Warner said. “Investors and policymakers deserve to know where that money is going, how these deals are being financed, and what risks are involved. We can’t afford to have massive financial exposure building up behind the scenes, only to discover it when something goes wrong.”

Senator Kennedy warned that current reporting masks underlying exposures: “AI is the future, but investing in it shouldn’t be a guessing game. Companies are spending trillions of dollars on this technology, and investors deserve to know where that money is going and what risks they’re taking.”

Mandatory Disclosures for Infrastructure Leases and Special Purpose Vehicles

The AUDIT AI Act targets opaque accounting practices used to fund high-performance graphics processing units and data center facilities. Under Section 4 of the bill, as detailed in the official Senate Section-by-Section Summary, the SEC must issue rules within 180 days requiring covered issuers to disclose all affiliated entities, joint ventures, special purpose vehicles (SPVs), and variable interest entities (VIEs) dedicated to AI infrastructure.

The rules mandate disclosure of all associated debt instruments, capacity guarantees, servicing agreements, and compute leases. These hardware commitments correlate with operational performance benchmarks, such as a compute FLOPS estimate, allowing investors to evaluate whether computing scale matches balance-sheet capacity. In addition, Section 3 requires public companies to disclose whether they maintain a formal AI governance policy and a dedicated board committee overseeing operational risks, with committee charters filed every three years.

Statutory Section Regulatory Focus Mandated Disclosure Requirement SEC Implementation Timeline
Section 3 Corporate Governance Public availability of AI governance policies and disclosure of board oversight committees 180 days post-enactment
Section 4 Infrastructure Financing Disclosures of SPVs, VIEs, joint ventures, debt, and compute lease commitments in Forms 10-Q/10-K 180 days post-enactment
Section 5 Financial Reporting Study SEC study on AI R&D accounting, hardware useful life assumptions, and Capex disaggregation 360 days post-enactment
Section 6 Standardized Risk Factors Specific disclosures on model concentration, compute volatility, foreign models, and licensing liabilities 180 days post-enactment
Section 7 Market Structure Assessment Annual SEC reporting to Congress on automated trading algorithms, price discovery, and herding effects 120 days post-enactment (annual)

Section 6 establishes standardized risk-factor disclosures for issuers with material AI exposure. Companies must quantify vulnerabilities across five distinct categories: dependence on third-party foundation models, concentration of cloud compute providers, training and inference cost volatility, intellectual property liabilities tied to training data, and operational reliance on foreign frontier systems.

The Trillion-Dollar Buildout and Shadow Infrastructure Liabilities

The Senate proposal addresses a massive wave of capital deployment into power grids, cooling infrastructure, and semiconductor clusters. Commercial developers have increasingly relied on private credit, synthetic leases, and unconsolidated entities to expand capacity without inflating reported corporate leverage.

An empirical study published by The Brookings Institution, conducted by Columbia University economist Stijn Van Nieuwerburgh, found that cumulative U.S. investment in AI infrastructure will reach $10.3 trillion between 2025 and 2032. This represents 3.6 percent of U.S. gross domestic product annually, exceeding historic capital expansions such as the national highway system. The study warned that reliance on off-balance-sheet vehicles obscures where financial risk concentrates.

Reporting from Punjab News Express noted legislative alarm regarding circular financing structures, in which cloud providers invest equity in AI companies that concurrently commit to multi-billion-dollar compute leases from those same providers. While European regulators address operational risks through frameworks like the EU AI Act Article 10 data and data governance rules, the AUDIT AI Act enforces direct market disclosure to ensure capital markets reflect genuine counterparty risk.

Support from Former Regulators and Academic Economists

Former regulatory officials and financial governance scholars voiced strong support for the bipartisan measure, noting that current financial reporting frameworks have not kept pace with intangible tech assets.

Robert J. Jackson Jr., former SEC Commissioner and professor of law at New York University, highlighted the importance of executive accountability. “Both the enormous potential and unprecedented risks of AI are top of mind in most American boardrooms,” Jackson stated. “This bill will give investors critical insight on whether, and how, corporate leaders are overseeing the technology that will define our future. And it will shine crucial light on how enormous and opaque AI investments are actually being funded. Ordinary investors deserve to know who is behind the systemic shifts that AI will soon bring—and how they’re paying for it.”

Former SEC Chief Accountant Paul Munter noted that operational governance must accompany technological deployment: “Governance and effective control over AI development and applications are critical aspects of a company’s effective and appropriate use of AI tools. Investors need greater transparency and visibility into companies’ governance and oversight process over development and use of AI applications that are important to a company’s ongoing operations.”

Analysis published by Lokmat Times highlighted that financial stability monitors view increased transparency as vital to preventing debt shocks across private credit funds that hold data center loan obligations.

Implications for University Cloud Contracts and Research Offices

Although the legislation directly regulates public registrants, the disclosure mandate carries operational repercussions for academic medical centers, university computing clusters, and sponsored research offices. Higher education institutions increasingly rely on commercial hyperscalers for large-scale scientific computation and regional research consortiums.

University research administrators must verify the stability and compliance posture of technology vendors. Sponsored projects managing dual agency requirements, such as those evaluated in NIH vs NSF vs ERC AI in grant writing, depend on continuous cloud access. When vendors are required to report SPV structures, research offices can better assess whether critical research infrastructure is vulnerable to financial distress.

Furthermore, Section 6’s mandated risk reporting on data licensing liabilities affects institutional data sharing. University offices drafting cloud contracts must define clear data use boundaries. Administrators negotiating vendor terms should consult the distinctions outlined in Data Sharing Agreement vs Data Processing Agreement to ensure academic intellectual property and clinical datasets are protected from undisclosed commercial training integration.

Section 5 also directs the SEC to examine corporate accounting for research and development, including asset lifespans for specialized hardware. Sourced guidance on hardware depreciation could influence how university supercomputing facilities calculate equipment depreciation for federally negotiated indirect cost rates.

Legislative Trajectory and Rulemaking Milestones

The AUDIT AI Act has been referred to the Senate Committee on Banking, Housing, and Urban Affairs. With senior members Warner, Kennedy, Alsobrooks, and Rounds as sponsors, the bill enters committee with bipartisan standing.

If enacted, the statute imposes strict agency deadlines. The SEC will have 180 days to issue proposed and final regulations covering corporate AI policy disclosures, board committee charters, and off-balance-sheet SPV and compute lease reporting. The Commission will also have 360 days to complete a formal study on research and development accounting standards in consultation with the Financial Accounting Standards Board, and 120 days to deliver its initial annual report on algorithmic market structure and automated trading.

Sources

Ask CASRAI · free to try

Ask about this story

2 free questions a day, no account.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →