The U.S. National Science Foundation has launched a $100 million program to build state and regional AI infrastructure hubs across the country, with NVIDIA (NVDA ), AMD, Intel (INTC ), and Dell Technologies (DELL ) among the private-sector partners lined up to support it. The NSF State and Regional Artificial Intelligence Infrastructure Hubs program, announced August 4, 2026, will fund up to 10 regional consortia that pool compute, data, and expertise for researchers, students, and educators who currently sit outside the frontier of AI-enabled science.
The structure is a public-private cost share. State or multistate consortia, drawing contributions from universities, state and local governments, philanthropies, and industry, will build and operate the actual compute. NSF’s money goes to the connective tissue: consortium coordination, workforce development, and faculty training. One award will be made per state or region.
“Artificial intelligence is transforming how we conduct research, accelerate scientific discovery and address complex challenges across disciplines,” said Brian Stone, performing the duties of the NSF director, in the agency’s announcement. “Through this effort, we can empower researchers, students and communities to advance AI-enabled science and drive the next generation of scientific breakthroughs.”
NVIDIA said it will contribute training resources, educator enablement, applied learning content, technical guidance, and partner platforms to participating hubs. The other named supporters are AMD, Intel, Dell Technologies, Hangar, and the Secunda Innovation Fund, with NSF saying it welcomes additional organizations.
The hub model traces to a 2020 University of Florida deal
NVIDIA’s post points to a working template for what these hubs are supposed to become. In 2020 the company, its cofounder Chris Malachowsky, and the University of Florida formed a public-private partnership to turn UF into what the company calls the country’s first true AI university, extending AI compute access to all Florida public universities. Since then UF has grown to more than 300 AI-focused faculty, embedded AI across all 16 of its colleges, and received more than $511 million in AI research awards since 2017.
The new program also builds directly on the NSF-led National Artificial Intelligence Research Resource pilot, where NVIDIA has been a leading contributor. Over two years the NAIRR pilot backed more than 700 research projects, spanning protein prediction to infectious disease outbreak management, according to NVIDIA’s account of the program. NVIDIA’s contribution took a concrete form: a cloud-based resource giving research teams dedicated access to a minimum of four DGX nodes for at least a month, plus onboarding support. One University of Michigan team trained its MIST molecular foundation models on a 40-GPU DGX cluster obtained through a NAIRR allocation, supplemented by 200,000 GPU hours on the Polaris supercomputer at Argonne’s leadership computing facility.
NSF says the hubs are encouraged to integrate with NAIRR, which would let regional consortia surge beyond local capacity and share datasets, and to engage with the agency’s TechAccess: AI-Ready America initiative on workforce development. The program also supports the White House-led Genesis Mission, the national AI-for-science effort established by executive order in November 2025.
Flexible infrastructure, one award per region
On infrastructure design, NSF is leaving the architecture to the regions. Consortia can run on-premises systems, cloud computing, or a combination, designing resources around regional needs and economic priorities, according to NVIDIA’s description of the program. NSF’s release frames the federal role as catalytic: the agency funds the people and coordination layers, including AI infrastructure professionals with the technical expertise to help researchers apply the compute, while the consortia and their partners supply the hardware and operations.
The program responds to Science: A New Golden Age, a report published in July 2026 by White House Office of Science and Technology Policy Director Michael Kratsios, which called for expanding world-class R&D infrastructure. The administration’s fiscal year 2028 R&D priorities memorandum, issued alongside the report, calls for investing in AI for science as a national mission.
“Regional partners who share in the benefits of discovery will pool their resources to unlock compute at a scale that no individual stakeholder, and no federal program, could achieve alone,” Kratsios said in the NSF announcement.
What comes next for the hubs
The program’s funding page links to solicitation NSF 26-513; prospective proposers are directed to the AI Infrastructure Hubs program team for details. NSF’s stated cap is an initial cohort of up to 10 hubs with one award per state or region, so the first awards will define which regional groupings, and which compute architectures, set the pattern for any expansion.
The workforce side runs on a parallel track. NSF’s release also encourages hubs to engage with its TechAccess: AI-Ready America initiative on workforce development; the agency has not detailed that program’s deadlines in the announcement. NVIDIA says its education role in the infrastructure hubs will center on repeatable, openly available programs covering degree pathways, short-form certificates, and stackable credentials in fields including physical AI and automation, healthcare, energy, agriculture, manufacturing, quantum computing, and cybersecurity.
Compute access for academic research has been one of the persistent bottlenecks in U.S. AI policy: the universities with the researchers frequently lack the clusters, and the clusters concentrate where the capital already is. Programs like this one, NAIRR, and the private buildout happening in parallel, from reactor-powered data center bets to consolidation up the compute stack, are all attempts to close different parts of that gap. NSF’s hubs target the regional layer specifically, and the solicitation’s arrival will show how much of the $100 million lands as hardware versus coordination.