The University of Minnesota will lead three projects funded by the Department of Energy's inaugural Genesis Mission program, selected after drawing the largest response in the agency's history. This August 5 announcement places Minnesota among 278 institutions nationwide that will use Phase I grants to build and test workflows combining artificial intelligence with scientific discovery, signaling a shift toward embedding machine learning directly into experimental design.
Building AI workflows for materials and geothermal systems
The awards support distinct approaches to integrating computation with physical science. Uwe Kortshagen will direct a team developing an AI-driven engine to evaluate semiconductor manufacturing methods and identify production errors faster than traditional trial-and-error cycles. Qizhi He is building a generative model that creates real-time "digital twins" of underground geothermal reservoirs, allowing operators to interpret sensor data without waiting for slow simulation runs.
Peter Kang combines laboratory experiments with predictive algorithms to map how fluids move through fractured rock during in situ mineral recovery. Each project treats AI as a functional layer in the research loop rather than a post-hoc analysis tool. Dean Andrew Alleyne said the initiative brings together the scientific expertise, engineering capability and advanced computing needed to accelerate discovery.
Funding structure and institutional support
Phase I grants provide initial resources for researchers to prototype AI-integrated workflows. The competition drew submissions from across academia, national laboratories, industry and nonprofit sectors, ultimately distributing awards to 168 universities, 87 national laboratories, 19 companies and four nonprofits. Executive Vice President and Provost Gretchen Ritter said the university has a responsibility to shape how AI advances discovery and benefits society.
The College of Science and Engineering secured all three awards, which feed into a broader campus push through the recently launched AI Hub. That initiative coordinates curriculum updates, workforce development and cross-departmental collaboration to standardize responsible AI practices across the research enterprise.
Why this matters for science and research professionals
Researchers should monitor how Phase I prototypes evolve into deployed tools that change daily lab and field operations. The DOE requirement to test integrated workflows means future grant cycles will likely prioritize proposals that show measurable time-to-insight gains over purely computational benchmarks. Teams that build reproducible AI pipelines now will have a structural advantage when applying for larger phase II allocations or partnering with national laboratories. Keeping track of these methodology shifts helps research leaders allocate equipment, hire computational staff and adjust data management strategies before standard practices solidify.
Your membership also unlocks:
