DOE selects first 278 Genesis Mission projects for AI-enabled research cycles

DOE selects 278 Genesis Mission projects for AI-enabled research

On July 22, the U.S. Department of Energy selected 278 Genesis Mission projects for award negotiations. The projects involve 342 institutions, including national laboratories, universities, companies, and nonprofit organizations.

The Genesis Mission provides research teams with shared resources for AI-enabled work. DOE reported that national laboratories lead 87 of the selected projects, universities lead 168, companies lead 19, and nonprofit organizations lead four. Selection begins the award negotiation process. Once agreements are reached, the teams will gain access to the program's resources.

DOE evaluated applications in two stages. It first assessed whether each proposal addressed one of the agency's 26 scientific priorities, then considered how AI could accelerate the proposed work. The competition received more than 5 000 applications, according to Nextgov.

The teams will receive tools for working with AI agents, models, software from industry partners, and computing capacity at national laboratories. This infrastructure is intended to shorten the scientific research cycle by helping teams formulate an approach, test it, and refine it based on the results.

Princeton's projects provide specific examples. One group will train a neural network to reproduce measurements in plasma physics that are too expensive or difficult to obtain through laboratory experiments. Another will build a model of mobile DNA transfer among microbes to predict which bacteria will acquire a genetic fragment. The portfolio also includes nuclear energy, critical mineral extraction, chip design, and nuclear fusion.

In March, the Genesis Mission connected 17 national laboratories, supercomputers, and scientific instruments, while Argonne demonstrated an AI research collaborator, autonomous laboratories, and robotic assistants within this system. DOE is now allocating this infrastructure across 278 research tasks.