New research from CD & W highlights how artificial intelligence is reshaping health‑care IT, yet most providers still lack the infrastructure needed to support AI workloads.
AI adoption outpaces infrastructure readiness
According to the 8th Annual Nutanix Healthcare Vertical Enterprise Cloud Index, 88 percent of health‑care leaders say their current systems are not fully prepared for AI deployment on‑premises.
Scott Ragsdale, vice president of sales for health‑care and SLED at Nutanix, explained that the rapid adoption of AI is hampered by longstanding technical debt. “Healthcare has a lot of technical debt,” he said, adding that overcoming this debt is the biggest hurdle for adopting new AI workloads.
While enthusiasm for AI agents is evident, health‑care executives remain cautious. They stress the need for governance, clinical validation, security, and compliance with regulations such as HIPAA before scaling AI solutions.
Shadow AI and the push for governed environments
Ragsdale described shadow AI as a symptom rather than a root cause. Clinicians, researchers, and business users often turn to unsanctioned tools when official channels cannot meet their speed or functionality needs. “If they can’t get what they need, they’re going to go out and download something that is unsanctioned,” he said.
To curb this risk without stifling innovation, he recommends establishing governed environments where approved AI platforms can operate safely. Clear policies and standardized platforms are essential for maintaining control while still encouraging rapid problem‑solving.
Implementing such governance may involve adopting container‑based architectures. Containers, according to Ragsdale, “provide portability, scalability and consistency across the environment,” allowing AI workloads to run wherever data is generated—potentially even in patient rooms.
Data shows a growing interest in containerization, and a notable share of IT executives report that their organizations are building new applications in containers. This trend indicates a shift toward cloud‑native technologies that can support AI at the point of care.
From a practical standpoint, health‑care providers must look beyond simply adding GPUs to data centers. Modernizing the operating model, improving data governance, and creating a hybrid infrastructure that can handle AI inference wherever it is needed are critical steps.
For many clinicians, AI can reduce manual transcription and streamline documentation in electronic health records, freeing up time to see more patients.
They must act now.
What health‑care leaders can do now
Ragsdale advises organizations to start with a defined business or clinical problem, set clear success criteria, and validate outcomes through controlled pilots. Learning from peers who have advanced further in their AI journeys can also help avoid common pitfalls.
Health‑care entities should also focus on modernizing legacy systems, adopting standardized platforms, and building a strategy that supports AI inference at the point of care—whether that is in a data center or directly in a patient’s room.
In practice, this means investing in hybrid clouds, container orchestration tools, and robust data governance frameworks that can keep pace with the rapid demand for AI capabilities.
Ultimately, the report paints a picture of a sector poised to benefit from AI, yet still wrestling with the foundational work needed to make that vision a reality.