DOE Awards $68 Million for AI Foundation Models in Science

The Department of Energy has selected 43 awards totaling $68 million to build artificial intelligence foundation models for scientific research — a bet that purpose-built AI can accelerate everything from drug discovery to climate modeling faster than general-purpose commercial models.

The awards, announced under the "Advancements in Artificial Intelligence for Science" funding opportunity, span 11 multi-institution collaborations running up to three years. Initial funding draws on $20 million in Fiscal Year 2024 dollars, with the remainder contingent on congressional appropriations.

What the Money Buys

The funded projects target three capability gaps that commercial AI providers haven't addressed. First, scientific foundation models trained on domain-specific data rather than internet-scale text. Second, privacy-preserving approaches that let institutions collaborate on sensitive datasets without exposing raw data. Third, energy-efficient AI algorithms designed for DOE's supercomputing infrastructure rather than cloud data centers.

"Progress in AI is inspiring us to imagine faster and more-efficient ways to do science," said Ceren Susut, DOE Associate Director of Science for Advanced Scientific Computing Research.

Who Should Pay Attention

Researchers at national laboratories and universities with existing DOE relationships are the primary beneficiaries, but the multi-institution structure means subcontracting opportunities exist for smaller teams with specialized AI expertise. The program sits under DOE's Office of Science, which has historically favored proposals demonstrating clear pathways from AI capability to measurable scientific outcomes.

DOE cautioned that selection does not guarantee funding — awards are finalized during negotiation. Researchers interested in future rounds should track Office of Science funding opportunities and explore grant discovery tools on the Granted blog for updates on federal AI funding.