Aug. 8, 2025 — The University of Pennsylvania is expanding its artificial intelligence research capacity with the help of a new off-campus supercomputer named “Betty.”
Developed by NVIDIA and hosted in a data center in Collegeville, roughly 30 miles from Penn’s campus, Betty combines high-performance GPUs and CPUs to run AI workloads capable of interpreting massive datasets and refining results over time. The system adapts to researchers’ needs, delivering the computational muscle required for advanced projects.
Kenneth Chaney, Associate Director of AI and Technology at the Penn Advanced Research Computing Center (PARCC), called Betty the result of a cross-campus effort involving the School of Engineering and Applied Science, the Perelman School of Medicine, and the School of Arts and Sciences.
“The needs for modern AI research have grown to a scale where it is no longer feasible for any one school to maintain such resources,” Chaney explained.
Representatives from multiple schools, together with the Office of the Vice Provost for Research, built this shared platform to enable new kinds of research that demand Betty’s computing power.
Marylyn Ritchie, Vice Dean of Artificial Intelligence and Computing at the Perelman School of Medicine, sees Betty as both a practical tool and a catalyst for interdisciplinary collaboration.
“In the future, we will be leveraging AI in all areas of research,” Ritchie said. “While GPU computing may be limited to certain fields today, I expect we’ll see it adopted across most disciplines.”
A Shift to Shared High-Performance Computing #
Betty marks a move away from the traditional model where individual labs purchase and maintain their own servers — a setup Ritchie described as expensive and inefficient.
“Most labs do not use their computers 24/7, so they sit idle much of the time,” she noted.
“With a computing center like Betty, the campus community can share resources for optimal usage. This enables computation at a scale that’s hard to achieve in a single lab.”
Why Off-Campus? #
Locating Betty in Collegeville was a deliberate choice. The system requires specialized cooling and energy infrastructure, including 1 megawatt of power, which Chaney said would be impossible to scale in urban Philadelphia. Despite being off-site, Betty — and all future PARCC systems — will be fully accessible to researchers via Penn’s network.
“If you can access the network at Penn, you can access Betty,” Ritchie confirmed.
Honoring a Computing Pioneer #
Betty is named after Frances “Betty” Holberton, one of the six original programmers of the Electronic Numerical Integrator and Computer (ENIAC), developed at Penn in the 1940s. The name celebrates Penn’s historical role in computing innovation.
Early Use and Future Outlook #
Currently in a pilot phase, Betty supports 10 research labs and 47 active researchers. Early projects include:
- Project Eureka — led by Computer and Information Science Assistant Professor Dinesh Jayaraman.
- Large models for biomolecular methods — developed by Bioengineering and CIS Assistant Professor Pranam Chatterjee.
- NSF AIRFoundry — led by CIS Department Chair and Adani President’s Distinguished Professor Zack Ives.
Chaney believes this is just the beginning:
“Betty’s rollout marks the start of a very bright future for supercomputing at Penn, opening the door to new resources and collaborations made possible by this system.”