LegisTrack
Back to all bills
HR 5332119th CongressIn Committee

Liquid Cooling for AI Act of 2025

Introduced: Sep 11, 2025
Sponsor: Rep. Obernolte, Jay [R-CA-23] (R-California)
Technology & Innovation
Standard Summary
Comprehensive overview in 1-2 paragraphs

Liquid Cooling for AI Act of 2025 would require the Comptroller General (GAO) to conduct a comprehensive technology assessment of liquid-cooling systems for AI compute clusters and high-performance computing facilities, and to develop government-wide best-practice guidance for federal agencies. The bill emphasizes the growing energy footprint of data centers and the potential for liquid cooling—such as direct-to-chip cooling and immersion cooling—to improve thermal performance, enable higher compute density, and reduce cooling loads. It also calls for collaboration with industry and research stakeholders, and designates a liquid cooling advisory organization to aid the review. The Department of Energy (DOE) would later review the GAO findings and provide Congress with recommendations on maintaining U.S. leadership in AI and pursuing heat-reuse opportunities. In short, if enacted, the bill would steer federal agencies toward standardized, evidence-based liquid cooling practices for AI and HPC, assess economic and energy implications, explore heat reuse, and strengthen oversight and coordination across government, industry, and national laboratories. It does not itself fund projects but creates a structured assessment, guidance, and coordination framework to inform future federal decisions.

Key Points

  • 1GAO technology assessment on liquid cooling for AI/HPC: Study research and development needs, costs/benefits, energy impacts, compute capacity gains, performance, resiliency, and cybersecurity effects; compare direct-to-chip cooling with immersion cooling; analyze coolant options and lifecycle considerations.
  • 2Federal guidance and standards: Develop government-wide best-practice guidance, consider federal use and ongoing research, and promote security, reliability, resilience, and education of operational best practices in liquid cooling.
  • 3Stakeholder engagement and advisory organization: Require input from government, private sector, academia, and national laboratories; designate a liquid cooling advisory organization (jointly by the Secretary of Energy and GAO) to coordinate and promote standards and adoption.
  • 4Scope of analysis: Include heat-reuse opportunities, failure scenarios (pump failures, leaks) in co-location environments, and reference architectures for rack/row/room cooling by density and approach; assess safety, interoperability, and lifecycle costs.
  • 5DOE follow-up and congressional reporting: Within 90 days, GAO reports findings to the Secretary of Energy and committees; within 180 days, DOE evaluates the GAO report and provides a congressional assessment with recommendations on research priorities and maintaining global AI leadership through liquid cooling and heat reuse.

Impact Areas

Primary group/area affected: Federal data centers and AI/HPC compute infrastructure across federal agencies; energy and technology policy oversight; DOE and National Laboratories involved in testing and guidance development.Secondary group/area affected: Private sector stakeholders (hardware manufacturers, data center operators, AI platform developers), academia, and public research institutions, which would be consulted and potentially align with new standards and best practices.Additional impacts: Potential influence on national energy efficiency goals, cybersecurity and resilience practices for critical computing facilities, and climate-related considerations through heat-reuse opportunities and more efficient cooling. May shape future federal procurement and deployment decisions for AI infrastructure.
Generated by gpt-5-nano on Oct 8, 2025