# Meta Pledges $2 Million for AI Research at Five U.S. Universities

By John Kim

Canonical URL: https://www.tokenpost.com/news/business/43083
Published: 2026-10-29T00:01:35.000Z
Updated: 2026-10-29T00:01:35.000Z
Section: Business

> Projects span medical research and fundamental science, with universities retaining control of patient and clinical data.

Meta unveiled a $2 million research program Wednesday, Oct. 28, to fund AI projects at five U.S. universities, pairing grants with model access and collaboration with Meta researchers. The projects cover medicine and fundamental science. Grant amounts and model assignments by school have not been disclosed.

The five grantees are California Institute of Technology, The Ohio State University College of Medicine, Louisiana State University, Indiana University and UTHealth Houston. Support varies by institution and project.

Meta’s first-party API will give five Caltech faculty members access to frontier Muse models, with project grants lasting up to three years. Projects include testing whether AI systems can generate machine-checkable mathematical proofs. A project will compare AI systems against MathDB Frontier, a benchmark based on more than 86,900 open problems at the research level.

“Frontier AI may be the next one,” said Fred Farina, Caltech’s chief innovation and corporate partnerships officer, referring to AI as a potential scientific instrument.

At Ohio State, researchers are studying whether routine abdominal CT scans, longitudinal medical records and ECG data, when available, can help models identify future risk of heart failure with preserved ejection fraction. Other funded work includes chest X-ray reporting and support for Alzheimer’s diagnosis.

Louisiana State University is developing a camera-based system to monitor infants with tracheostomies for tube displacement and signs of distress. Indiana University is studying conversational AI for people with dementia.

Potential research tools include Muse Spark Open, the Muse Wonder and Muse Spark model series, and Muse Glimmer. The list also includes computer vision models SAM, V-JEPA and DINO, and the multilingual model SeamlessM4T. Muse Glimmer is a 30-billion-parameter open-weight model.

For health care projects, universities will run models in their own environments and retain control of patient and clinical data. Meta will not receive, process or otherwise have access to that information.

UTHealth Houston’s work involving 65 patient cases checked preliminary workflows; it did not evaluate clinical outcomes. A proof of concept has been published for Louisiana State University’s infant-monitoring project.