Genesis Mission – Manhattan Project of the 2020s?

27.3.2026 12.12

The supercomputer LUMI, located in CSC’s data center in Kajaani, Finland, is a key part of European AI infrastructure. Photo: Mikael Kanerva, CSC

In November last year, the U.S. administration launched the Genesis Mission – a large-scale effort to develop AI infrastructure for scientific research. In an executive order, the White House ordered the Department of Energy (DOE), under an ambitious timeline, to define key scientific challenges to be solved by AI tools, and to determine resources among the National Laboratories that can be harnessed for their creation.

The main goal of the Genesis Mission is to update U.S. research infrastructures for the AI age using various federal resources and incentives for stakeholders. It aims to utilize vast datasets from the DOE’s National Labs to train scientific foundation models and create novel AI systems to test new hypotheses, design experiments, analyze results, and run autonomous research workflows.

Darío Gil, the undersecretary for science at DOE, who leads the Genesis Mission, has called it this generation’s Manhattan Project. The project aims to unite industry, academia, and the federal government’s resources, to achieve the goals of the administration.

Challenge from China

Behind the effort is the United States’ strategic competition with the People’s Republic of China (PRC), which has made massive advances in its scientific output. Other key objectives for the mission include improving national security and speeding up technological innovation – all in line with achieving the administration’s stated goals of American dominance and prosperity.

The U.S. administration is worried that U.S. scientific discovery, in contrast to the PRC, has stalled. Whatever the reason – be it the advancement of scientific discovery, all the “low hanging fruit” having been picked already, or the structuring of scientific careers directing people into constantly narrowing fields, limiting the system’s capability to solve grand challenges – the administration believes the U.S. scientific enterprise needs a major leap in productivity, which is expected to arrive in the form of AI infrastructure.

New partnerships emerge

In December, DOE announced agreements with 24 industry partners. In January, the mission got its first international partner when Japanese national research institute RIKEN singed an MoU with the Argonne National Lab, Fujitsu and NVIDIA in the fields of AI and high-performance computing aligned with Genesis Mission. In February, DOE published key scientific challenges the mission hopes to solve, and launched the Genesis Mission consortium. In March, DOE opened applications for $293 million in research funding under the Genesis Mission.

The national science and technology challenges that the Genesis Mission hopes to solve include accelerating delivery of fusion energy, securing America’s critical mineral supply, and discovering new quantum algorithms. All in all, DOE has determined over 20 national challenges spanning advanced manufacturing, biotechnology, critical materials, nuclear energy, and quantum information science.

AI workforce remains a challenge

Overall, the response to the Genesis Mission has been mostly positive. Michael Witherell, director of Lawrence Berkeley National Laboratory (LBNL) told the AAAS in November that the National Labs directors had already been discussing strengthening the Labs’ AI efforts, and a push from Washington was necessary and welcome.

Possible challenges for the mission to overcome, according to Center for Strategic and International Studies (CSIS) experts Navin Girishankar and Chris Borges, include the need for DOE to hire the necessary experts in the near term for the mission to succeed, which could prove costly, and legal and security challenges in aggregating, integrating and increasing the accessibility of federal scientific datasets.

DOE is seeking to leverage partnerships to help overcome possible challenges. In January, DOE solicited public and private sector input on strategies for meeting the technical challenges of the mission and on developing a skilled American workforce to advance AI in science and engineering. According to DOE, “a workforce of 100,000 American scientists and engineers will need to be trained over the next decade to lead the world in AI-powered science, innovation, and applications”, and the agency sought educational institutions to partner with in advancing this goal.

Anne Järvinen
anne.jarvinen(at)gov.fi

Sources: