Healthcare and research institutions have never had more data, more computing power, or more specialized expertise. Yet many of their biggest challenges remain stubbornly human.
Researchers struggle to discover work happening elsewhere in their own organizations. Clinical data sits apart from research data. Industry partners and academic teams often operate at different speeds and under different constraints. Valuable knowledge exists, but it remains difficult to connect.
A recent discussion featuring leaders from Purdue University and the University of Rochester revealed a common theme. The institutions seeing the greatest value from modern infrastructure are not treating it as a collection of technologies. They are using it as a foundation for collaboration.
Collaboration Starts with a Shared Environment
Betsy Hillery, Director of Data Initiatives at The Lilly Purdue Innovation Institute, sees this challenge firsthand.
Academia has the knowledge and all the talent, but we don’t have the tooling in place,” she explained. “On the reverse side, industry has all the controls, so they’re prohibited from doing the things that will get them farther in the science.
The result is a familiar problem. Researchers, data scientists, IT teams, and industry partners often spend more time navigating organizational boundaries than advancing research itself.
One of Purdue’s initiatives focuses on accelerating pharmaceutical workflows through agentic AI. The technical challenge was significant, but Hillery described an equally important lesson learned during the project. Faced with a growing collection of tools and custom integrations, the team realized complexity was becoming an obstacle.
We started pulling way back,” she said. “What do we have in our toolbox, and what can platforms actually provide to us even though we want to engineer everything?
That shift in thinking reflects a broader change taking place across research organizations. The goal is no longer simply building more infrastructure. It is creating environments where expertise, workflows, and data can move more freely between groups.
The impact extends beyond operational efficiency. Hillery noted that when scientific workflows become visible across an organization, duplication becomes easier to identify.
If one scientist is doing something that another scientist is also doing, traditional systems isolate them,” she said. “Now we have one streamlined workflow instead of four separate ones.
The challenge, in other words, is not simply making research faster. It’s making research more collaborative.
Helping Researchers Find Each Other
At the University of Rochester, CIO Julie Myers believes many research institutions face a different version of the same problem.
“Traditional research environments are plagued by isolated compute across the institution,” she said.
Over time, departments acquire their own storage, their own computing resources, and often their own support staff. The result is a fragmented environment where researchers may be unaware of datasets, expertise, and projects that could accelerate their work.
Myers shared a story that illustrates the problem perfectly.
During a university research event, dozens of researchers presented their work to one another. During a break, the conversations shifted away from the presentations themselves.
Their main comment was, ‘I had no idea that we were doing this kind of research across the institution.
That realization had nothing to do with storage performance or GPU availability. It reflected a visibility problem.
Researchers were producing valuable work, but the institution lacked an effective way to expose that knowledge across organizational boundaries.
The implications become even more significant as AI enters the research process. Modern scientific discovery increasingly depends on combining datasets from multiple domains. Imaging data, genomics, clinical records, simulation outputs, and experimental results often need to be analyzed together.
The ability to discover and access those resources may become as important as the compute infrastructure used to process them.
As Myers put it, the opportunity is to make a large institution feel smaller by reducing the barriers that separate people from information and from one another.
Breaking Down the Wall Between Clinical and Research Data
For James Forrester, CTO of the University of Rochester Medical Center, the next collaboration challenge lies between clinical operations and research.
Historically, healthcare organizations have treated these environments separately. Clinical systems were optimized for patient care and regulatory compliance. Research systems focused on analysis and discovery. Data moved between them slowly, if at all.
That model is becoming difficult to sustain.
Forrester pointed to medical imaging as an example. Healthcare systems collect enormous volumes of diagnostic images that are critical for patient care. Traditionally, those images have been stored in archives designed primarily for retention and retrieval.
Researchers rarely had direct access to them.
All of those images, for example, have been landlocked in the past,” Forrester said. “Now they're available to our research community, and you’re talking billions and billions of images.
That shift changes what becomes possible.
Researchers studying disease progression, treatment outcomes, or predictive models can begin working with clinical datasets at unprecedented scale. At the same time, insights generated by researchers can more quickly influence clinical practice.
Forrester described this evolution as a move beyond traditional hypothesis-driven research.
“The old model is analyzing what we think we already know,” he said.
A more connected data environment creates opportunities for discovery that researchers may not have anticipated when they began their work.
The Next Measure of Infrastructure
Infrastructure conversations seemingly have always centered on performance, capacity, reliability, and cost. Those metrics remain important. Healthcare and research organizations cannot compromise on any of them.
But the discussion among Purdue and Rochester leaders pointed to an additional measure that is becoming increasingly important: how effectively does infrastructure help people work together?
The most valuable systems are no longer those that simply store information or process workloads. They are the systems that expose data to new audiences, connect researchers across disciplines, and allow institutions to operate as cohesive environments rather than collections of isolated departments.
The institutions seeing the greatest value from modern infrastructure are not treating it as a collection of independent technologies. They are treating it as a shared foundation for collaboration.
As AI continues to reshape healthcare and scientific research, that distinction may prove more important than any individual technology decision.
