The Los Alamos National Laboratory, established in 1943 to coordinate the United States’ effort to build the first atomic bomb, is now focusing its historical scientific expertise on artificial intelligence (AI) for both scientific and national security purposes. The Lab has created a National Security AI Office led by senior director Jason Pruet, who is tasked with advancing AI as a cornerstone of modern scientific discovery and national strength. In a statement published in the Lab’s science and technology magazine, 1663, Pruet emphasizes that the Lab’s investment in AI has reached an unprecedented level this year, with funding surpassing all previous years. This marks a significant shift in the Lab’s strategy, moving from its traditional focus on nuclear and energy research to a renewed emphasis on AI as a transformative force in science and security.
Pruet argues that AI is now becoming a foundational element of scientific progress, citing the increasing capabilities of large AI models in areas as diverse as hypothesis generation, complex simulations, and biological design. He notes that the cost of running these models at the forefront of AI development is prohibitively high, with only major national research facilities capable of providing the necessary computational power. For example, the Lab’s Venado supercomputer, equipped with 250,000 GPUs, represents a significant resource that is being shared with academic institutions such as the University of California and the University of Michigan. This initiative reflects a broader effort to democratize access to high-performance computing resources, allowing universities to engage in large-scale AI research that was previously beyond their reach.
Pruet highlights the importance of fostering collaboration between the government, academia, and industry to ensure that the United States maintains a competitive edge in AI innovation. In a recent interview, he stated that the scale of AI research is now so vast that no single university can afford to develop the necessary infrastructure on its own. Therefore, the Lab is working closely with multiple institutions to provide shared computational resources, enabling researchers to explore cutting-edge applications in fields such as mathematics, physics, and bioengineering. Pruet also underscores the growing global competition in AI, noting that the U.S. must invest heavily to ensure that its scientific and technological leadership remains intact. The Lab’s AI initiatives are part of a broader strategy to secure national security and technological dominance through strategic investment in AI research and development.