Category i: fabric for AI-driven science (fabaid)

The National Science Foundation awarded The Morgridge Institute for Research, Inc. a $9 million Cooperative Agreement on May 15, 2026, to operate the Fabric for AI-Driven Science (FABAID), a national-scale platform that manages the lifecycle of scientific datasets and enables AI-driven research across academic and research communities.

Under the Computer and Information Science and Engineering program (CFDA 47.070), FABAID operates the Open Science Data Fabric (OSDF) to deliver data to and from compute resources, expand dataset discovery capabilities, support active metadata extraction, and integrate archive services. The platform integrates scientific data repositories, dataset producers, curators, and processing capacity to enable researchers to manage data-centric workflows. FABAID's data-centric workload services integrate with distributed computing infrastructures to allow researchers to transform or simulate billions of datapoints and curate datasets for AI-driven workloads. The services are built on the Pelican platform and HTCondor software suite, distributed and supported under open-source licenses for use across autonomous entities. Facilitation activities provide training and collaboration with user communities to enable effective use of the national integrated data services and systems.

Performance runs through April 30, 2031, at Madison, Wisconsin. The University of Wisconsin System, through University of Wisconsin–Madison, is a sub-recipient responsible for the architecture of FABAID technologies, ensuring alignment between Pelican and HTCondor software, developing the metadata approach, and overseeing overall project coordination and personnel supervision at UW-Madison. The sub-award was dated July 28, 2026.

Abstract

The national-scale fabric for AI-driven science (fabaid) weaves services and technologies to support the life cycle of scientific datasets. Through fabaid, researchers effectively manage and leverage data-centric workflows that are a foundation of AI-driven research. By integrating scientific data repositories, dataset producers and curators, and processing capacity, the fabric presents the nation's educators and researchers with a coherent set of powerful and dependable capabilities. These capabilities are powered by the integrated NSF cyberinfrastructure ecosystem and initiatives such as the national AI research resource (nairr). Fabaid leverages decades of experience in the integration of national scale services and systems and innovates in data-related methodologies and technologies for the AI ERA. Fabaid operates the open science data fabric (osdf) to deliver data to and from compute, expand dataset discovery capabilities, support active metadata extraction, and integrate archive services. Fabaid's data-centric workload services integrate with distributed computing infrastructures to enable researchers to transform or simulate billions of datapoints, curating datasets for AI-driven workloads. Fabaid's services are built on the pelican platform and the htcondor software suite, and distributed and supported under open-source licenses, allowing them to be used across autonomous entities. The project establishes a national scale platform for AI-driven research by continuously integrating these technologies with the emerging AI methodologies, services and tools. Facilitation activities in fabaid provide training and collaboration with user communities, enabling them to effectively use the national integrated data services and systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are planned for this award.