Powering the Next American Century: US Energy Secretary Chris Wright and NVIDIA's Ian Buck on the Genesis Mission
Key point
The US Department of Energy and NVIDIA emphasized the power needed for AI scaling and their collaboration on the Genesis Mission.
Details
Chris Wright, the US Secretary of Energy, and NVIDIA Vice President Ian Buck emphasized at the SCSP AI+ Expo that AI leadership is built on top of energy leadership. Wright said that the more energy is abundant and cheap, the greater the opportunity for society.
The centerpiece was the US Department of Energy's (DOE) Genesis Mission. DOE contributes scientists from 17 national laboratories, national-scale challenges, and data, while NVIDIA contributes not just chips but algorithms, methodologies, and 20 years of collaboration experience with the national laboratories. Buck said NVIDIA is 100% committed to Genesis, and that expectations are high across both the laboratories and industry. Wright explained that in the five-layer structure of AI that Jensen Huang described, his department is responsible for the bottom layer: energy.
- Two AI supercomputers are being built at Argonne National Laboratory.
- Equinox is currently under construction and will be composed of 10,000 NVIDIA Grace Blackwell GPUs, using the same GPUs and software currently used for AI training.
- Solstice will use 100,000 GPUs and NVIDIA Vera Rubin, which Buck described as being on the scale of 5,000 exaflops—5 times larger than the combined total of all TOP500 supercomputers.
Buck also introduced an open-source model trained on 1.5 million physics papers and further fine-tuned on 100,000 fusion papers. DOE researchers can query it to accelerate fusion research. Buck also said that in moving from Hopper to Blackwell, performance improved 30x, and performance per watt improved 25x.
Wright pointed out that over the past 20 years, the US has greatly increased oil and natural gas production but has barely increased electricity generation. As a result, his department is moving to bring back natural gas, nuclear, and coal all together, and said that 3 SMR units will reach criticality before July 4th this year.
He also noted that AI can shrink tasks like grid interconnection review, which used to take years, down to weeks or even hours. Wright said he aims to deliver concrete results within 12 months in fusion, materials, and grid interconnection, and argued that expanding data centers and power generation facilities can actually lower electricity rates and make the power grid stronger.
Finally, he summarized AI not as a technology that replaces humans, but as a tool that amplifies human capability and passion.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.