NVIDIA CEO Jensen Huang personally attended the unveiling of the US Department of Energy's next-generation supercomputer, Doudna. This supercomputer is equipped with NVIDIA's upcoming AI chip platform, Vera Rubin, and is manufactured by Dell's AI server cluster, with plans to launch next year.

Doudna supports not only AI and simulation computing but also quantum algorithm development. Huang described its powerful computing capabilities as a "time machine for science", capable of compressing years of research into just days.

Doudna was formally unveiled at the Lawrence Berkeley National Laboratory in California, named after Nobel laureate Jennifer Doudna, a pioneer of CRISPR gene editing. The Department of Energy stated that the system is built on Dell's infrastructure and is expected to be operational by 2026, tailored to address the most urgent scientific missions of the department. Facilities under the Department of Energy across the US, such as the Fermi National Accelerator Laboratory and the Joint Genome Institute, will rely on Doudna's capabilities.

Sudip Dosanjh, director of the National Energy Research Scientific Computing Center (NERSC), indicated that Doudna was designed to accelerate various scientific workflows and will be connected to the Energy Sciences Network (ESnet) for seamless data transmission nationwide for near-real-time analysis. Nick Wright, the chief architect of Doudna, expects that it will drive advancements in fields like nuclear fusion, novel superconductors, drug discovery, and astronomy.

The Department of Energy estimates that the scientific output of Doudna will exceed that of the previous system, Perlmutter, by more than tenfold, while consuming only 2 to 3 times the energy, representing a 3 to 5 times improvement in efficiency per watt. Doudna supports workflows across traditional high-performance computing (HPC), AI, real-time streaming, and quantum computing. This includes support for scalable quantum algorithm development and the design of future integrated quantum and high-performance computing systems via NVIDIA's CUDA-Q platform.

Due to its remarkable computing power, Huang has described Doudna as a "time machine for science", suggesting that this supercomputer could reduce years of research discoveries to just days, empowering solutions to some of the world's toughest problems.