The MATRIX AI Consortium at the University of Texas at San Antonio (UTSA) recently announced it has received a $4 million grant from the National Science Foundation (NSF) to support “The Neuromorphic Commons (THOR)” project.
This project is a collaborative initiative across multiple universities, providing researchers with access to large-scale, diverse neuromorphic computing hardware systems.
The goal of the THOR project is to promote interdisciplinary research into the neural foundations of biological intelligence, including areas such as perception, decision-making, and continuous learning in the real world.
UTSA believes the THOR project will revolutionize algorithm design, hardware and software co-design, and neuromorphic applications—similar to the way high-performance computing systems transformed engineering research. The system will be available for research in areas like artificial intelligence, machine learning, physics, life sciences, and computational neuroscience.
Dhireesha Kudithipudi, the principal investigator for THOR and director of the MATRIX AI Consortium, explained that the initiative will help create a national hub for large-scale neuromorphic platforms with support from industry partners.
“This field is at a critical juncture, and broadening access to researchers is essential. This initiative represents a community-driven approach, creating a framework that is shaped by and for the community,” she noted.
The THOR team also plans to develop and share training materials on neuromorphic learning algorithms and systems through open platforms, ensuring accessibility to a wider audience.
Neuromorphic computing mimics the structure and function of the human brain, particularly the neocortex, which is responsible for functions like spatial reasoning, sensory perception, and language. These systems use advanced spiking neural networks of chips with artificial neurons and synapses to process data and solve problems. The networks simulate the way biological neurons transmit information through discrete spikes, enabling the system to handle temporal data patterns.
Despite recent focus on quantum computing and the rise of GPU-driven hardware, neuromorphic computing remains promising for energy-efficient and low-latency computing, which could boost technologies like AI. Large tech companies are also embracing this field. For instance, IBM launched its NorthPole neuromorphic chip in 2023, claiming it is 25 times more energy-efficient than current chip technologies. In addition, Intel unveiled its Hala Point system earlier this year, powered by 1,152 Loihi 2 neuromorphic processors.
NSF Program Director Andrey Kanaev emphasized the importance of the grant in advancing the NSF’s mission to promote innovation and increase access to research resources. “By providing bio-inspired computing resources to a broader group of researchers in computer science, neuroscience, and computational physics, this project will help democratize access to cutting-edge tools and support breakthroughs in energy-efficient, resilient AI,” he said.
The core team behind this interdisciplinary effort includes Dhireesha Kudithipudi (Principal Investigator, UTSA), Catherine Schuman (Co-Principal Investigator, University of Tennessee Knoxville), Gert Cauwenberghs (Co-Principal Investigator, University of California San Diego), and Vijay Janapa Reddi (Senior Personnel, Harvard University).
