DOE Rolls Out Phase II Genesis Mission Projects, New Educational Opportunities

The Department of Energy today unveiled a dozen Phase II awards for its Genesis Mission initiative to accelerate scientific discovery and engineering with AI. It also announced its four-year doctoral pathway as part of the Graduate Fellowship Pilot, and awarded $96 million to 67 early career scientists.

It’s been nearly a year since the DOE launched the Genesis Mission, its ambitious plan to imbue AI technologies and techniques directly into the heart of American science and engineering. The DOE originally launched with 26 science and technology challenges in areas like accelerating fusion energy research, better material discovery, and drug development, but it later expanded that to 33 challenges.

In July 2026, the DOE unveiled awards for nearly 300 Phase I projects at the Genesis Mission Summit, giving the government the greenlight to fund smaller projects with up to $750,000 over nine months to determine the feasibility of ideas. Around the same time, it also announced two Phase II projects, which are more involved and can range up to $15 million over a three-year period.

DOE Under Secretary for Science Darío Gil at Genesis Mission Summit in Washington, D.C. July 2026

The 12 new Phase II projects demonstrate the big ambitions of the DOE with its AI efforts. The projects target a range of topics, from fusion energy and decoding RNA to developing semiconductors that work in space and building quantum magnets.

The projects will be led by a range of organizations, including DOE National Labs like Argonne, Oak Ride, and Lawrence Berkeley and universities like the University of California San Diego, MIT, and University of Illinois Urbana-Champaign. There is even one private company, Commonwealth Fusion Systems, an MIT spinout.

“We’ve had an embarrassment of riches in good proposals and ideas,” Darío Gil, the DOE’s Under Secretary for Science, said in a press conference. “It has been very demanding for the staff and leadership, because there was a great deal to review, but it is a very impressive portfolio that will now be executed.”

Many Phase I projects are already underway, with the labs, schools, and companies behind them setting up and executing experiments to see if their idea pans out. Award recipients get access to the various computing, software, and data resources that are provided by the Genesis Mission, such as the American Science and Security Platform (ASSP), the American Science Cloud (AmSC), and Transformational AI Models Consortium (ModCon).

“This is about building a portfolio,” Andrew Schwartz, the deputy associate director for basic energy sciences in the Office of Science, said in the press conference. “It’s about starting with the science and then integrating AI as a powerful tool into the scientific workflow. That’s what we asked for in both the Phase I and Phase II proposals, and it’s what we looked for in selecting them: the strength of the science, its importance, and how AI would be a truly transformative addition.”

The DOE checks in with Phase I project leaders around the six-month mark to see how the project is going, and to decide whether to continue it, end it, or transition it into a Phase II project. Phase II projects, meanwhile, typically have a yearly check-in cadence, Schwartz said.

“We designed the program to allow flexibility, because anyone following AI knows you have to keep up with it day by day or week by week,” Gil said. “As things change, and they will, we’ll be able to adapt. We’re very confident we’ve selected a portfolio of some of the most extraordinary people doing truly meaningful, impactful work. But we also expect those teams and the program managers to change course when change is needed, and that’s a good thing.”

Training the next generation of scientists is also a DOE priority, and the federal agency made two announcements on that topic today.

For starters, the DOE announced the Genesis Mission Graduate Fellowship, a new academic program that it will embark upon with university partners to build accelerated, four-year doctoral pathways for graduate students with dual-competency training in both AI and science and engineering.

The Graduate Fellowship program will start as a pilot in 2027, with $20 million in funding for about 100 fellows in American universities. The goal is to establish a model with the universities, national labs, and industry partners that will eventually scale to train 100,000 domestic students within the United States, said Harriet Kung, deputy director for science programs in the DOE’s Office of Science.

“The general thesis is that we need to shorten the training path, because it is a very leaky path,” Kung said. “We lose a lot of exceptional talent, especially domestic talent, through the lengthy PhD track. The goal is to accelerate the PhD from the usual six to seven years to four years.”

The program will revolve around dual competencies. Each fellow will focus on AI and their domain disciplines, Kung said. They will have immersive research experience, including an internship at one of the national labs as well as an internship in industry.

“These are fully paid, with the usual stipends and tuition covered as part of the fellowship,” she said. “This will be a highly prestigious, premier fellowship program for the Genesis Mission.”

A big goal of the Graduate Fellowship program is to train and retain American science and engineering talent, Gil said.

“Nine out of 10 undergraduates who earn science and engineering degrees are domestic students,” he said. “In graduate school, on average, it’s one in two, and in some fields it can be as low as one in three. So we go from nine in 10 to, at best, half, and sometimes just a third.”

The DOE wants to shrink PhD programs from six or seven years to four as part of its Genesis Mission Graduate Fellowship (Pete Spiro/Shutterstock)

By shortening the length of the PhD program from six or seven years to four, providing a diversity of experiences, and focusing on AI in addition to core scientific areas, the DOE is confident it can attract a solid contingent of American talent to this new graduate program.

The DOE also announced that it has allocated $96 million in funding for 67 scientists as part of its Early Career Research Program (ECRP), a DOE Office of Science program that began in 2010 to support exceptional early-career scientists and stimulate research in critical disciplines.

“The program is highly selective,” Kung said. “The typical success rate is 6% to 7%, which likely makes it one of the most prestigious and competitive awards in the nation. To date, we have funded about 1,100 awardees, and we’re adding 67 with this year’s cohort.”

The 67 2026 ECRP recipients are affiliated with 39 universities, 11 DOE national laboratories and SC user facilities in 26 states. The ECRP awards are worth $875,000 over five years at universities, while awards to scientists at a DOE laboratory or Office of Science user facility are approximately $2,750,000 over five years, according to the DOE.

Kung said she is especially pleased with the diversity of topics these 67 awards cover. From applied math to developing surrogate and inference models, or from quantum algorithms for fault-tolerant quantum computers to 3-D tokamak prediction, this year’s class has a very deep bench, she said.

“To us, these are not only current leaders but future leaders of the Office of Science portfolio,” Kung said. “We are training them to be at the forefront of their fields and to contribute to the Office of Science and DOE missions. We’re very excited about this progress.”