Nvidia and Oracle are teaming up with the US Department of Energy to build a series of AI supercomputers, including the department’s largest to date. The partnership aims to modernize US high-performance computing and AI research, with a focus on commercial applications and national defense. Construction begins immediately, with the first supercomputer set for delivery in 2026. Nvidia will assemble the systems domestically, leveraging US-based silicon design and manufacturing.
Nvidia and Oracle will collaborate with the US Department of Energy (DOE) to build a series of AI supercomputers, including what will become the department’s largest system to date. Nvidia CEO Jensen Huang announced the plan at the company’s first GTC AI conference in Washington, DC, emphasizing that construction will begin immediately and that initial computing capacity will begin supporting DOE workloads within a week.
The effort involves seven new systems built in cooperation with Argonne National Laboratory. Nvidia and Oracle will co-develop the infrastructure, focusing on performance optimization, data throughput, and AI-driven simulation capabilities. DOE Secretary Chris Wright stated that most of the computational resources will support commercial applications that strengthen US industry, while a portion will advance national defense and scientific research. Wright initiated contact with industry partners to accelerate the department’s computing capacity and reinforce its research capabilities.
The first supercomputer is scheduled for delivery in 2026, with the largest following later in the program. The new systems will expand DOE’s portfolio of AI infrastructure, which currently includes Argonne’s Polaris and Aurora projects. Wright said the initiative aims to modernize US high-performance computing (HPC) and AI research while aligning with the administration’s broader industrial strategy.
Nvidia’s involvement deepens its long-term relationship with the federal government. The company’s AI hardware and software already support major DOE facilities, and the Washington-based GTC event underscores its growing role in government technology programs. Huang explained that Nvidia intends to assemble its new systems within the United States, stating that every stage—from silicon design to software integration—will occur domestically.
He attributed the company’s expansion to policies that encourage energy production and industrial investment. During a post-event briefing, Huang noted that the availability of reliable energy directly affects AI development. He expressed appreciation for government support of energy-intensive technologies, describing them as vital to maintaining US competitiveness in AI and advanced manufacturing.
The DOE’s next-generation AI supercomputers will draw substantial power and water resources. Such systems operate under heavy thermal loads and require efficient cooling to sustain continuous performance. Environmental concerns over the energy and water use of large-scale data centers continue to generate debate among regulators and energy providers. Wright acknowledged that the department must manage power distribution carefully to avoid stressing regional grids.
Nvidia’s decision to host a second GTC in Washington reflects a strategic shift toward closer collaboration with policymakers and federal research agencies. Traditionally held in San Jose, California, the GTC conference showcases advances in GPU computing and AI systems. Holding the event in the capital highlights the company’s interest in aligning with government initiatives related to national computing infrastructure and industrial policy.
Huang’s remarks connected Nvidia’s business outlook with evolving trade and energy conditions. He plans to meet with US officials in South Korea as part of ongoing diplomatic discussions involving semiconductor supply chains and US–China trade policy. Such interactions affect Nvidia’s long-term access to fabrication partners and materials critical to GPU production.
Huang reiterated his intention to strengthen domestic production capacity. Nvidia plans to assemble AI servers, supercomputer nodes, and networking equipment in US facilities, reducing dependence on overseas manufacturing. He emphasized that the company’s current roadmap includes US-based silicon design, packaging, and system integration for future GPU and AI server architectures.
Energy policy remains a central factor in AI infrastructure expansion. AI workloads consume vast amounts of electricity, with each new supercomputer adding megawatts of demand to regional grids. The DOE plans to pair its new AI supercomputers with optimized cooling and renewable offsets where possible, though the program still relies heavily on conventional energy sources.
Wright explained that expanding compute capacity aligns with DOE’s dual mandate—supporting national energy security and fostering industrial innovation. He said the collaboration among Nvidia, Oracle, and Argonne National Laboratory allows DOE to combine commercial expertise with federal oversight in the development of large-scale AI systems. The partnership aims to balance commercial competitiveness with the department’s public research responsibilities.
Nvidia’s continued engagement with DOE and other federal partners demonstrates how AI infrastructure now intersects with national economic and strategic interests. The planned systems will provide computational resources not only for government research but also for private-sector projects that depend on large-scale model training and simulation.
What do we think?
The partnership among Nvidia, Oracle, and the Department of Energy presents opportunities for advancing AI research and commercial applications, while also posing environmental and geopolitical threats. On the one hand, the collaboration can drive innovation, boost US competitiveness, and create new economic opportunities. On the other hand, the massive energy consumption and water usage required by these supercomputers raise concerns about environmental sustainability and strain on regional grids. Furthermore, the reliance on conventional energy sources and potential dependence on international supply chains may undermine the initiative’s long-term viability and national security implications.
LIKE IT? THINK YOUR FRIENDS AND ASSOCIATES MIGHT? PLEASE SEND IT TO THEM WITH OUR BEST WISHES.