ARPA-H announces awards to advance AI for rare disease diagnosis, discovery, and treatment 

Funded teams will build transformative solutions that shorten time to diagnosis, uncover new insights into rare disease biology, and accelerate therapeutic development

The Advanced Research Projects Agency for Health (ARPA-H), an agency within the U.S. Department of Health and Human Services, today announced the teams receiving contract awards through the Rare Disease AI/ML for Precision Integrated Diagnostics (RAPID) program. This landmark initiative will invest up to $98.5 million over 4.5 years to advance AI technologies and build foundational data resources to enable earlier diagnosis, deepen understanding of disease progression, and accelerate rare disease drug development — from target discovery through clinical trial design and endpoint selection. 

“For millions of patients and families, living with a rare disease still means years without an accurate diagnosis and too few paths to effective treatment,” said RAPID Program Manager Scott Gorman. “RAPID aims to change that trajectory by building the capabilities needed to improve diagnosis and support treatment development across thousands of rare and ultra-rare conditions. This bold endeavor reflects ARPA-H’s mission to take on challenges that demand ambitious technology development, bringing together capabilities across sectors that no single organization can develop alone.” 

More than 350 million people worldwide live with one of over 10,000 rare diseases — many of which cause severe symptoms that greatly interfere with quality of life. The diagnostic odyssey faced by patients with a rare disease lasts six years on average but can extend for decades, often leading to inappropriate care, irreversible disease progression, and rising medical costs. Delayed diagnoses also slow treatment development by making it harder to recruit enough patients for clinical trials. With only about 5% of rare diseases having an approved therapy, the field urgently needs better ways to identify and characterize patient populations, understand disease progression and variability, and build the evidence needed to design more effective clinical trials and accelerate the development of new treatments. 

Current technologies address only part of the challenge. Research funded by the National Institutes of Health has demonstrated that timely access to genomic sequencing can lead to improved diagnostic rates in many rare diseases. However, many genetic test results remain inconclusive without additional clinical context, while other rare diseases lack a known genetic cause, leaving patients without a definitive diagnosis. AI-powered diagnostic tools could help identify rare diseases earlier and broaden access to diagnostic expertise, but progress is limited by the shortage of robust, representative datasets needed to develop and validate these tools. 

“Rare diseases represent one of the greatest unmet needs in medicine, and one of the most important frontiers for AI,” said Gorman. “By building AI-ready datasets and infrastructure, along with technologies designed to learn from sparse and complex rare disease data, RAPID aims to improve outcomes for patients while also advancing precision medicine more broadly.” 

RAPID will advance the full AI stack for rare disease. The program will unify and standardize dispersed data and develop scalable approaches to privacy-preserving data sharing, consent management, and interoperability — removing barriers that have long prevented rare disease data and technologies from being used across institutions. At the center of this effort will be the largest AI-ready, real-world data resource of its kind for rare diseases, linking longitudinal clinical records with genomic, patient-reported, and other emerging sources of health data. Performer teams will use this resource to develop AI-driven tools that will drastically reduce time to diagnosis, accelerate clinical trials through precise cohort identification, and lower healthcare costs through earlier intervention for rare diseases. 

Leading patient advocacy groups, including Global Genes and the National Organization for Rare Disorders (NORD), will support RAPID across key areas, including close collaboration with the program’s technical performers to drive patient involvement and engagement in the solutions developed to ensure they reflect the needs of the rare disease community. 

Additionally, RAPID will launch Rare Challenges, an open innovation platform supported by the program teams and NASA’s Center of Excellence for Collaborative Innovation (CoECI) that will use rigorous competitions to benchmark and accelerate emerging AI approaches across diagnosis, mechanistic understanding, and other rare disease priorities. The platform will enable researchers and organizations beyond the core performer teams to leverage RAPID’s secure data infrastructure and benchmarks to develop new solutions and contribute to the program mission. 

RAPID performer teams are led by: 

• University of North Carolina: The team will build the largest real-world rare disease dataset by integrating clinical, genomic, and patient-reported data across thousands of rare diseases. The resulting resource will provide a privacy-preserving foundation for research and innovation across diagnosis, disease understanding, clinical trial design, and therapeutic development. The team will be supported by a broad group of partner organizations spanning academic and research institutions, health data and technology companies, and rare disease patient advocacy and research organizations. 
• Sage Bionetworks: The team will build a secure and interoperable platform that harmonizes multimodal rare disease data and enables the development, evaluation, and benchmarking of new AI approaches for rare disease diagnosis and discovery. The platform will provide transparent performance assessment, common evaluation standards, and shared infrastructure to accelerate innovation across the rare disease ecosystem. 
• FDNA: The team will develop tools to collect longitudinal, multimodal health data, including photos, videos, voice and patient reported concerns directly from patients at national scale. The team will use these data to develop and deploy AI tools in real-world clinical and direct-to-patient settings, enabling earlier identification of rare and ultra-rare diseases with complex or variable presentations that may be missed by single-modality approaches. 
• Probably Genetic: The team will develop advanced patient-centered tools to collect multimodal data — including photos, videos, wearable data, functional assessments, clinical records, and genomic data — and generate high-fidelity synthetic datasets to support rare disease research. The team will use these resources to develop patient-facing AI tools that enable earlier rare disease identification, uncover new diagnostic indicators and connect patients with confirmatory testing, care navigation, and clinical trial opportunities. 

Additional program support will be provided by the Lawrence Berkeley National Laboratory, the National Institutes of Health’s All of Us Center for Linkage and Acquisition of Data, and the Monarch Initiative. Several leading companies in AI, data, and health care—OpenAI, Anthropic, Amazon Web Services, and Google—have also pledged in-kind support for RAPID, including computing credits for LLMs, subject-matter expertise, and engineering resources.