Meta announced $2 million in support for research at five U.S. universities on October 28, 2026. The support is not a stated equal cash award to each school: Meta says the mix varies by institution and project, and may include access to its AI models and technology, collaboration with Meta researchers, and funding for selected projects. Meta’s announcement and research post name California Institute of Technology (Caltech), The Ohio State University College of Medicine, Louisiana State University (LSU), Indiana University, and UTHealth Houston.
What the support covers
Meta describes a mix of tools, collaboration, and project funding rather than a single cash grant for each university. The work spans mathematics, physical sciences, medicine, and clinical research. Meta lists tools including Muse models, Segment Anything Model (SAM), DINO, V-JEPA, and SeamlessM4T; individual projects use different models and technologies.
That distinction matters when reading the $2 million headline: access to technology and time with researchers are also part of the described support, while funding is intended for selected projects. Meta does not present the total as five identical university awards.
Five universities, different research questions
At Caltech, projects explore whether AI can help with research-level mathematics, physics, chemistry, biology, and computer science. One effort builds on MathDB, a resource that Meta says contains more than 86,900 research-level open problems. Researchers plan to assess tasks such as formalizing conjectures and producing machine-checkable proofs. These are research questions, not reported solutions to those problems.
The Ohio State University College of Medicine is investigating several clinical applications. One project explores future risk prediction for heart failure with preserved ejection fraction (HFpEF) using abdominal CT scans, medical records, and ECG data when available. Other work is developing drafts of chest X-ray reports and studying how radiologists respond to correct and incorrect AI drafts in a simulated setting. A clinician-guided system is also being developed to analyze evidence related to dementia diagnosis.
LSU’s projects include Eyes-On, which explores monitoring infants with tracheostomies for tube displacement and signs of distress. Meta says a proof of concept has been published. LSU also has a working prototype for an AI coach designed to support procedural training in virtual reality; user studies are planned. Indiana University’s research includes NeoSmartML+, which aims to provide multilingual maternal and newborn-care guidance and assess infant positioning on low-cost Android devices, including for offline use. Other Indiana projects study how children learn and how conversational systems might adapt to communication patterns among older adults with dementia.
At UTHealth Houston, projects include benchmarking AI models in local infrastructure, prototyping hands-free point-of-care workflows, and developing an agent for glioblastoma-related clinical work. Meta describes early testing across 65 patient cases as initial workflow validation, not a measure of clinical outcomes. Across the five universities, Meta’s October 28 descriptions focus on research aims and early project stages; they report no initiative-wide clinical outcomes.
Meta’s stated approach to patient data
For healthcare projects, Meta says the universities will run its models in their own environments and retain control of patient and clinical data. Meta says that data will not be shared with, processed by, or otherwise made available to the company. This is Meta’s description of the arrangement.
