Building the AI-Infused University Starts with Smarter Compute Investments
Orchestrating people, process, and compute for an AI-infused future at UGA.
I remember the fall of 2011 clearly. Alan Dorsey was interviewing to become Dean of Franklin College, the University of Georgia’s largest and most academically diverse college. After the usual pleasantries, Alan shifted the conversation toward something deeper: research infrastructure. A physicist by training, he understood the strategic value of high-performance computing and wanted to know how a partnership with my team could advance the college’s research agenda. He got the job and spent the next decade leading the college through substantial research growth and transformation.
When his tenure as Dean ended, Alan joined my organization as Associate CIO for Research, where he led our HPC operations with the same clarity and conviction he brought to Franklin. Now, two years later, he’s returning to the faculty as a Simons Foundation grantee. In today’s Dispatch, I want to reflect on what we learned under Alan’s leadership and where we’re headed next as we build on that foundation.
Why it matters
For a research-intensive university, the promise of becoming truly AI-infused is not an abstract aspiration. It is a practical challenge, and it rests on the strength of the computing infrastructure that underpins the work of our faculty and students. AI models, whether used in genomics, drug discovery, or astronomy, are computationally hungry. They demand GPUs optimized for deep learning, high-throughput storage, robust networking, and the human expertise to make it all work reliably.
That means continual investment, not just in hardware, but in facilities, power, and the staff who can run, maintain, and optimize it. Without that backbone, AI strategies stay stuck on PowerPoint slides instead of producing research breakthroughs.
What’s new
At the University of Georgia, we are strengthening this backbone. I am pleased to share that Dr. Jesse Kissinger will serve as our new Associate Chief Information Officer for Research, supervising operations for our high-performance computing center. Dr. Kissinger is a leading researcher in bioinformatics and has been a persistent, clear-eyed advocate for building research capacity across the university.
Her appointment matters because she brings the faculty perspective directly into my leadership team, which directs all technology operations at UGA. Too often, technology decisions are made without the lived experience of those doing the work. By bridging that gap with our faculty, we ensure our infrastructure plans and other operations are aligned with the real needs of our academic community.
The bigger picture
In the past, universities often equated “research computing” with the central HPC cluster: a powerful shared environment where the largest, most complex jobs could run. That central resource remains critical. But in an AI-driven era, it is no longer the only resource that matters. At UGA, we are pursuing a three-tiered approach:
Desktop powerhouses. Many breakthroughs are happening outside the cluster. In physics, infectious diseases, and other STEM disciplines, researchers are pushing boundaries with advanced desktop workstations. Today, an 80-core GPU-enabled Mac Studio with 512 GB of RAM can be purchased for around $10,000. For many research groups, that’s enough to process massive datasets, run sophisticated simulations, or train targeted AI models, all without queuing for cluster time. This keeps the central HPC available for the heaviest computational loads.
Cloud elasticity. While on-premises hardware remains more cost-effective for continuous workloads, cloud resources from Microsoft, Amazon, and Oracle can provide overflow capacity. They are especially valuable for short-term spikes in demand, specialized computational needs, or collaborations with external partners who already operate in the cloud. Cloud elasticity ensures researchers don’t have to scale back a project simply because the internal queue is full.
Central HPC. Our Georgia Advanced Computing Resource Center (GACRC) remains the backbone of our research computing portfolio. It is the home for the largest, most resource-intensive projects: climate simulations, exoplanet identification, genomic sequence alignments, and large-scale AI model training. And it benefits from the efficiency of shared investment across disciplines.
Why this matters for an AI-Infused University
An AI-infused university is one where computational resources are not a bottleneck. It is an environment where a faculty member with a promising idea can get the computing power they need quickly, in the format that makes sense for their work, without navigating a maze of delays or compromises.
That requires not only the right mix of hardware and capacity, but also thoughtful leadership to guide those investments. Dr. Kissinger will oversee operations at the GACRC and work with deans and faculty to make strategic use of all three tiers. The goal is to match the right workload with the right resource and to make sure no research effort is slowed by a lack of access to compute, storage, or expertise.
The final word
AI’s demands on research infrastructure will only grow. Models will get larger, datasets more complex, and collaborations more computationally intensive. By building a balanced portfolio of computing resources and by embedding faculty leadership into the governance of those resources, UGA is taking the steps necessary to remain competitive and innovative in this next chapter of research.
We are fortunate to have Dr. Kissinger in this critical role, and I look forward to working with her as we continue to shape UGA into a truly AI-infused university.
Great points. I am seeing that even smaller institutions need access to HPC to fully participate in AI instruction and research. Given the expense of HPC infrastructure and related support costs, HPC as a shared service might make sense for many universities.