HR-2385-119
Ordered to be Reported in the Nature of a Substitute (Amended) by the Yeas and Nays: 29 - 0.
Sponsored by Jay Obernolte (R-CA)
What it does
This bill would formally establish the National Artificial Intelligence Research Resource (NAIRR), a federally managed infrastructure that would provide researchers, students, educators, and qualifying small businesses with shared access to computing power, datasets, AI testing tools, and educational resources. The National Science Foundation would oversee a Program Management Office to run day-to-day operations, and a competitively selected nongovernmental Operating Entity would handle actual operations. The bill would also create an interagency Steering Subcommittee, set user eligibility rules, establish privacy and security requirements, and authorize the NAIRR to accept private-sector donations.
Who benefits
Academic researchers and university students who currently lack access to large-scale computing resources. Small businesses that have received federal funding and need AI infrastructure. Nonprofit research institutions. Federal agencies seeking AI research capacity. Historically underrepresented groups in STEM who face barriers to expensive private computing resources. Researchers focused on AI safety, ethics, and trustworthiness, who would receive priority access. Renewable or startup AI companies that could benefit from a more diverse research ecosystem. The broader public, to the extent that wider AI research participation produces more broadly beneficial AI systems.
Who is hurt
Large technology companies that currently hold a near-monopoly on cutting-edge AI computing resources may face increased competition from a publicly subsidized research ecosystem. Private cloud computing providers could see reduced demand if researchers shift to NAIRR resources, though they may also participate as resource providers. Taxpayers would bear costs of establishing and operating the infrastructure. Foreign-affiliated researchers — particularly those connected to countries on the national security exclusion list — would be explicitly barred from access. Researchers at for-profit companies not qualifying as small businesses would be ineligible.
Supporters argue
Supporters argue that access to the computing power and data needed for cutting-edge AI research is currently concentrated among a handful of large technology companies, effectively locking out universities, nonprofits, and small businesses. They contend that a shared national resource — modeled on existing federal scientific infrastructure like national laboratories — would broaden the AI research base, improve diversity in who develops AI systems, and strengthen U.S. competitiveness against state-backed AI programs in countries like China. The bill's unanimous 29-0 committee vote suggests broad bipartisan agreement that the current resource gap is a genuine structural problem.
Opponents argue
Opponents argue that creating a new federally managed AI infrastructure layer risks duplicating existing private-sector and academic computing resources, potentially crowding out private investment rather than complementing it. They contend that the bill's funding mechanism — relying heavily on private-sector donations and unspecified appropriations — leaves the NAIRR's long-term financial sustainability uncertain, and that a resource-constrained or poorly governed NAIRR could become a bureaucratic bottleneck rather than an innovation accelerator. Critics may also raise concerns that exempting Advisory Committees from the Federal Advisory Committee Act (FACA) reduces public transparency and accountability over key governance decisions.
Constitutional context
Congress has broad authority to fund and organize scientific research infrastructure under the Commerce Clause (Art. I, §8, cl. 3) and the Intellectual Property Clause (Art. I, §8, cl. 8), which together support federal promotion of science and technology. No active constitutional tension is implicated: the bill creates a voluntary-access research resource, does not compel private-sector participation, and does not involve government surveillance or content regulation that would trigger Fourth or First Amendment scrutiny under Riley v. California (2014) or Moody v. NetChoice (2024).
Checks and balances
The Executive Branch — primarily NSF and OSTP — gains administrative authority to establish and oversee the NAIRR, subject to congressional appropriations control, annual public reporting requirements, independent performance assessments, and competitive procurement rules for the Operating Entity.
Historical precedent
The National Science Foundation's existing supercomputing centers and the Department of Energy's national laboratory network provide analogous models of federally funded shared scientific infrastructure made available to academic and nonprofit researchers.