Single Source for Establishing Pilot/Opportunity program for AI Models to Accelerate Diabetes Research (U24- Clinical Trials not allowed)

The National Institutes of Health (NIH), through the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), has announced a forecasted cooperative agreement opportunity titled “Single Source for Establishing Pilot/Opportunity program for AI Models to Accelerate Diabetes Research (U24 - Clinical Trials not allowed).” This initiative seeks to address one of the most pressing challenges in diabetes research: the high level of disease heterogeneity. Diabetes affects millions of people in the United States and globally, and its chronic, long-term nature requires innovative solutions to improve prevention, diagnosis, treatment, and prognosis. NIDDK emphasizes the urgent need to integrate advanced data science methods into the diabetes research ecosystem to accelerate progress in the field. The background for this funding opportunity stems from the significant accumulation of complex data and prior scientific knowledge regarding diabetes. Despite this, there are major hurdles in fully harnessing these resources due to limited data science expertise among diabetes researchers and the lack of diabetes-specific artificial intelligence tools and models. To overcome these barriers, NIDDK proposes to establish a pilot funding program that draws on recent advances in artificial intelligence and data science. The goal is to create AI foundation models specifically for diabetes research, validate these models through key research questions, and disseminate them broadly to the research community for application and further development. The scope of this program includes four major objectives. First, it will support the development of foundational AI models that can handle diabetes-specific data. Second, it will focus on validating these models against critical questions related to the heterogeneity of diabetes. Third, it will ensure that models are widely disseminated and that the research community is engaged in their further development and validation. Finally, the program will foster the creation of use cases that demonstrate how these models can accelerate the pace of research. Expected outcomes include not only the availability of new AI tools for use by diabetes researchers but also the integration of AI experts into the diabetes research workforce. This is a single-source competition opportunity, designated for the University of California, San Diego, which will administer the Pilot/Opportunity program in collaboration with dkNET (the NIDDK Information Network). Eligibility is therefore restricted, and other institutions or applicants are not eligible to apply. Applications will undergo peer review, and only meritorious proposals will be considered for funding, consistent with NIH’s standard processes. Because this is a forecasted opportunity, no specific funding ceiling, floor, or total program funding has yet been published. The estimated timeline for this opportunity begins with a posting date of April 1, 2026. Applications are expected to be due by July 5, 2026, with awards anticipated on January 1, 2027. The project start date is set for February 1, 2027, and the program is forecasted for fiscal year 2027. No pre-application deadlines, such as letters of intent, have been indicated in the forecast. Since this is a forecast, applicants should monitor NIH updates and Grants.gov for any changes in dates or requirements as the official Notice of Funding Opportunity is finalized. The program contact for this opportunity is Xujing Wang, Ph.D., who can be reached at 301-451-2862 or via email at xujing.wang@nih.gov. Interested parties should note that all applications must be submitted through Grants.gov, following NIH’s standard application processes, and should comply with all NIH submission requirements when the opportunity is officially posted. Because this is a forecasted opportunity, applicants should prepare for additional details to be released as the official notice becomes available.