The U.S. National Science Foundation (NSF) announced a $100 million investment, made jointly with Capital One and Intel, to fund five National Artificial Intelligence Research Institutes and a community hub.
The goal of the pact is to enable the institutes to pursue fundamental AI research, aiming to translate discoveries into practical applications. The program also supports national AI workforce development through outreach to high schools, universities and industry, according to the announcement.
AI-MI at Cornell
The AI‑Materials Institute (AI‑MI), led by Cornell University, will use AI to accelerate discovery of materials for energy, sustainability and quantum technologies. The institute plans a cloud‑based portal called the AI Materials Science Ecosystem that integrates large‑language models with experimental data, simulations, images and scientific literature. Building on the popular arXiv research paper repository that Cornell hosts, AIMS-EC will facilitate work on several fronts, including discovering two-dimensional moiré structures with properties suitable for robust quantum bits, designing new superconductors and developing molecules for removal of microplastics from the environment. The institute will also train students at all levels through partnerships with schools, universities and industry. NSF’s news release says that AI‑MI will “create the AI Materials Science Ecosystem, a cloud‑based portal that integrates large language models with experimental data, simulations, images and scientific literature.”
The Institute for Foundations of Machine Learning (IFML)
The Institute for Foundations of Machine Learning (IFML), led by University of Texas at Austin, builds on work begun in 2020 to develop new foundational tools for generative AI, including diffusion models that power generative‑AI tools like Stable Diffusion 3 and Flux. In its next phase the institute will expand generative‑AI research into domains such as protein engineering and clinical imaging and develop methods for handling noisy data and improving model reliability, particularly for health applications. The institute comprises researchers from a string of other research centers. The NSF notes that IFML’s work on diffusion models underpins widely used generative models and that the new award will develop tools to “expand generative AI to new domains … including protein engineering and clinical imaging” and “develop new methods to handle noisy data and improve model reliability”. IFML members were instrumental in developing coursework for a new Master of Science in Artificial Intelligence (MSAI) degree program at UT Austin, addressing the demand for a highly skilled AI workforce.
The Institute for Student AI‑Teaming (iSAT)
The Institute for Student AI‑Teaming (iSAT), led by University of Colorado Boulder, develops AI partners that help student groups learn together by facilitating discussion, exploration and reasoning. More than 6,000 middle‑school students and educators have used iSAT’s tools. The center’s researchers have developed two AI “partners,” CoBi (Community Builder) and the Jigsaw Interactive Agent (JIA), which they’ve tested in real-world classrooms. The next phase will develop a semester‑long curriculum and expand AI literacy. NSF’s summary describes iSAT’s AI partners that facilitate group learning and notes that over 6,000 middle‑school students and educators have participated. It adds that the institute will “co‑develop a semester‑long curriculum” to build AI literacy. The institute brings together researchers from nine universities spanning 15 research areas, working with school district partners.
The Molecule Maker Lab Institute (MMLI)
The Molecule Maker Lab Institute (MMLI), led by University of Illinois Urbana‑Champaign, uses AI and machine learning to speed up discovery and creation of molecules for medicine, materials and clean energy. In its first five years, the institute has resulted in 166 journal and conference papers, 11 patent disclosures. That includes six that have been licensed and two start-up companies. Accomplishments include the creation of AlphaSynthesis, an AI-enabled platform that helps researchers plan and execute chemical synthesis. In its next phase the institute will develop advanced AI tools, including new language models and intelligent agents, that can reason, predict and help design useful molecules such as drugs, catalysts and new materials. The institute brings together a team of chemists, engineers, and AI experts from the University of Illinois Urbana-Champaign, Pennsylvania State University and the Rochester Institute of Technology.
The AI Institutes Virtual Organization (AIVO)
The AI Institutes Virtual Organization (AIVO), led by University of California Davis, serves as a national hub that connects federally funded AI institutes, government stakeholders and the public. Building on a pilot launched in 2022, AIVO coordinates events, networking tools and collaboration support to create a cohesive innovation ecosystem and serves as a nexus for the 29 AI Institutes, organizing annual summits for AI institutes’ leadership. AIVO will amplify the work of the institutes and promote public engagement. NSF states that AIVO will expand on a 2022 pilot to connect AI institutes and government stakeholders, foster communication through events and networking tools, help form public‑private partnerships and raise awareness of how AI can address real‑world challenges.
The AI Research Institute on Interaction for AI Assistants (ARIA)
The AI Research Institute on Interaction for AI Assistants (ARIA), led by Brown University, focuses on accelerating the development of next‑generation AI assistants that are safer, more effective and adaptable to individual user needs. The institute’s work will focus on the potential for use in mental and behavioral health, where trust and safety are of the utmost importance, combining research on human and machine cognition to create AI systems that can interpret a person’s unique behavioral needs and provide helpful feedback in real time. The NSF announcement summarizes ARIA’s mission as developing “next‑generation AI assistants that are safer, more effective, and better able to adapt to individual user needs”.
Context within U.S. policy
The investment aligns with the White House AI Action Plan and Executive Order 14277 (see America’s AI Action Plan for more) on advancing AI education. According to the NSF, the AI institutes are intended to “translate cutting‑edge research into scalable, practical solutions that improve lives” and to build a national infrastructure for AI education and workforce development; the program will train researchers, empower educators and reach into communities.
Internationally, Europe, Israel, the UAE and other countries are launching national AI institutes, supercomputers and safety programs, signaling a global race to develop AI capabilities and governance frameworks. At the same time, the U.S. Department of Defense’s contracts with Microsoft, OpenAI, Anthropic, Google and xAI illustrate how commercial AI leaders are becoming essential defense suppliers, challenging incumbents such as Palantir.
