Lifetime Citizen Portal Access — AI Briefings, Alerts & Unlimited Follows
Lab directors urge more investment in exascale computing, AI and laboratory recapitalization
Loading...
Summary
Directors told the House subcommittee that upgrades to supercomputers and research infrastructure are needed to apply AI at scale for scientific discovery and national security; witnesses highlighted recent upgrades such as Argonne's Advanced Photon Source overhaul and DOE exascale machines.
WASHINGTON — Directors of DOE national laboratories told the House Science subcommittee that modern supercomputers, AI infrastructure and upgraded experimental facilities are essential to translate basic research into technologies for energy, national security and industry.
Paul Kearns, director of Argonne National Laboratory, described recent upgrades at Argonne: "The $850,000,000 upgrade to the Advanced Photon Source replaced the facility's 30 year old electron storage ring ... increasing x‑ray brightness by up to 500 times." He also highlighted Aurora, the exascale system sited at Argonne, as part of a trio of U.S. exascale machines that include Frontier at Oak Ridge and El Capitan at Lawrence Livermore.
Kimberly Budil described El Capitan at Livermore as a leading AI‑capable machine, saying it "has 44,000 GPUs, making it an ideal machine for AI investigations." Tom Mason described partnerships that bring private AI model owners into classified test beds: Los Alamos has arranged to field raw model weights for advanced models on lab hardware in secure environments, citing work with OpenAI and NVIDIA.
Witnesses said modern infrastructure is the raw material for AI‑driven science. Mason warned that aging facilities such as the Los Alamos Neutron Science Center need recapitalization to keep the labs' experimental capabilities aligned with computing advances. Budil and Kearns said high‑performance computing coupled with AI accelerates cosmological simulations, materials discovery and medical research and that representative private‑sector collaborators provide models and hardware that labs can use under classified or restricted arrangements.
Members asked for details on how such investments translate into jobs and economic value; witnesses pointed to partnerships with industry and universities and argued that sustained multiyear funding is required to keep facilities competitive internationally. No formal funding decisions were made at the hearing; members asked for additional documentation on computing upgrades and partnerships.
