Million Dollar Bike Ride 2025 Pilot Grant Program

Pilot Grant Program (for various rare diseases including CADASIL, UGDH, ZC4H2, CDKL5 Deficiency Disorder, Adult Polyglucosan Body Disease, Bloom Syndrome) is sponsored by Orphan Disease Center - University of Pennsylvania. The Orphan Disease Center offers pilot grants for various rare diseases, supporting research related to specific conditions. For example, the CDKL5 Pilot Grant Program provides funding for research related to CDKL5 Deficiency Disorder, and the APBD Pilot Grant Program supports research related to Adult Polyglucosan Body Disease. A Natural History Study grant is also offered to understand disease progression over time.

GeoPEX is AFRL's standing open broad agency announcement for geospatial intelligence processing and exploitation, and it is one of the more practical routes into Air Force AI funding for teams working on imagery and sensor analytics. The announcement covers research, development, integration, test and evaluation of technologies providing geospatial intelligence in all its forms and from any source - imagery, imagery intelligence, or geospatial data and information. The AI-relevant content sits in its emphasis on full-spectrum GEOINT, AI and machine learning techniques for imagery exploitation, multi-source data integration and sensor fusion, autonomous geospatial reasoning, intelligent processing pipelines, cloud-based high-performance computing and photogrammetry. The mechanics are what distinguish it from a conventional solicitation. It is an open, two-step BAA effective until 30 September 2026: initial submissions are white papers only, and full proposals are accepted by invitation. That structure is a significant advantage for smaller teams, because the cost of entry is a white paper rather than a full proposal package, and AFRL uses the white paper stage to steer effort toward technically relevant work before anyone invests in a formal proposal. Submissions are accepted on a rolling basis rather than against a single closing date, so timing against AFRL fiscal-year funding cycles matters more than any published deadline. Individual awards normally run 250,000 to 10,000,000 dollars over up to 24 months, against approximately 99.9 million dollars of total BAA funding. The full announcement and submission instructions are posted on SAM.gov.

This is the responsible-AI and governance strand of the NIH Common Fund's PRIMED-AI program, and it is by a wide margin the most accessible of the program's five announcements - two years, up to $300,000 in direct costs per year, roughly five awards, and no requirement to build a clinical tool or assemble an industrial partnership. What it funds instead is methodological infrastructure: standardized, reusable frameworks that address the recurring problems in multimodal AI clinical decision support, drawn from a defined list of topics including data quality, data harmonization across heterogeneous sources, model transparency, patient privacy, and regulatory navigation. A specific structural requirement shapes the whole application - proposals must develop and test two or more distinct frameworks on separate topics, so a deep single-topic proposal will not satisfy the announcement. The word testing carries real weight here as well; NIH is not funding position papers or consensus documents but frameworks that are empirically exercised against realistic multimodal health data and imaging scenarios, so applicants should plan concrete validation work into a compressed two-year timeline. Because the output is meant to serve the wider PRIMED-AI consortium and eventually the field, teams that combine informatics or AI methods expertise with genuine regulatory, bioethics or data-governance depth are better positioned than pure methods groups. The award mechanism is U01 with clinical trials not allowed, and foreign entities are ineligible. For research groups interested in trustworthy and responsible health AI but lacking the scale for the program's multi-million dollar center or tool-development awards, this is the natural entry point into PRIMED-AI, and the modest budget makes it realistic for a single well-composed interdisciplinary team.