Inside AI-Native Hospitals With NVIDIA's David Niewolny

David Niewolny, Director of Business Development for Healthcare and Medical at NVIDIA, has spent much of his career working across healthcare and technology. 

Equipped with over 18 years of experience building and leading healthcare-focused divisions within AWS, Real Time Innovations, NXP and Freescale Semiconductor, he is well-versed in emerging technologies and innovations entering the healthcare sector.

Today, much of his focus at NVIDIA centres on helping health systems, medtech companies, digital health partners and researchers accelerate the development and deployment of AI and robotics across healthcare, from medical imaging and clinical documentation to surgical robotics, hospital automation and physical AI at the point of care. 

One of the technological shifts of particular interest to him, and indeed the wider NVIDIA healthcare team, is AI-native hospitals. These next-generation medical facilities are designed to integrate AI into their core operational and data architecture, rather than operating isolated AI tools or laying AI capabilities onto existing digital infrastructure.

From isolated to integrated AI

One core, defining difference separates existing digitally-integrated hospitals and emerging AI-native hospitals.

David explains that “many hospitals use AI tools, but largely as individual point solutions. One application helps with documentation. Another supports imaging. A third assists with scheduling. They're valuable, but they operate independently.”

AI-native hospitals, on the other hand, become “a connected care environment where digital agents, physical AI, robotics and simulation work together as a coordinated layer around clinicians and patients”. 

These hospitals also prioritise AI quality over quantity. “The goal isn't simply to add more AI,” David says. “It's to build an adaptive, orchestrated healthcare system that extends workforce capacity, improves hospital efficiency and helps clinicians focus on patient care. The hospital of the future will be measured less by how much AI it has, and more by how well that AI coordinates care.” 

David points to significant progress in Taiwan to advance the transition. For example, Foxconn and other leading medical centres have moved beyond pilot programmes to deploy coordinated AI agents, collaborative robots, and digital twins to support everything from cancer screening and ECG analysis to surgical workflows and hospital operations.

A changing macro landscape

David believes three core forces are converging to expedite the development of AI-native hospitals: healthcare capacity limitations, AI maturity and simulation advances. 

“Around the world, health systems are dealing with staffing shortages, aging populations and rising healthcare costs, so the pressure to do more with less is very real,” says David. 

Adding to this practical need, large language models, agentic AI, accelerated computing, simulation and physical AI have reached a stage of maturity where they're practical enough to deploy in real healthcare settings. 

David explains, “We're seeing hospitals use digital twins to safely train, test and validate new workflows before they ever reach a live clinical environment, which helps accelerate innovation while reducing risk. Digital twins let hospitals rehearse change before patients ever experience it.”

Hospitals and clinicians have also become increasingly willing to evaluate and deploy these technologies than they were previously.

Use cases

David identifies four areas facilitating healthcare’s evolution from isolated AI applications to AI-native systems, with the next chapter of healthcare AI revolving around orchestration. In other words, ensuring agents, robots, devices and workflows work together in harmony.

The four impact areas David identifies are:

1. Agentic AI – Hospitals are moving beyond standalone AI tools toward coordinated AI systems that can reason, plan and act across clinical and operational workflows. Rather than supporting a single task like documentation or scheduling, specialised AI agents will increasingly help coordinate care, streamline workflows and assist clinicians across the patient journey.

2. Physical AI – Physical AI brings intelligence out of the screen and into the clinical environment. Healthcare has enormous operational demands that don't require clinical judgment but still consume valuable staff time, and collaborative robots are beginning to support nursing teams, automate logistics and assist in surgical environments, freeing clinicians to spend more time with patients. Surgical robotics leaders are already using NVIDIA healthcare physical-AI tools to train, test and validate next-generation systems before deployment.

3. Simulation and digital twins – Before deploying AI or robotics in live clinical environments, hospitals increasingly want to train, test and validate these systems in realistic virtual environments. For example, Apian and NVIDIA are deploying photorealistic digital twins at NHS hospitals to safely develop and validate autonomous logistics robots before they ever enter patient care environments — a simulation-first approach helping to accelerate innovation while maintaining the safety standards healthcare requires. 

4. Healthcare-specific foundation models – Healthcare has unique requirements around governance, privacy and regulation, driving growing demand for sovereign AI built specifically for the sector. OneAdvanced's work with NVIDIA to develop what it describes as the UK's first private sovereign healthcare LLM, trained on NHS primary care data, shows how organisations are building AI grounded in local clinical workflows while meeting stringent governance and data residency requirements.

The common thread connecting these examples is that AI is becoming foundational infrastructure for healthcare. “Whether it's helping researchers discover new medicines faster, enabling hospitals to safely deploy robotics, or giving clinicians AI systems that work alongside them, the greatest impact will come from AI augmenting human expertise and helping healthcare systems scale to meet growing demand,” explains David.

Alleviating global health challenges

David believes that AI-native hospitals can alleviate recurring operational bottlenecks present across global health systems: too much demand, too few staff and too many hours lost to work that does not require clinical judgment. 

“AI can help by taking on many of the repetitive tasks that don't require clinical judgment – things like documentation, workflow coordination and information retrieval. That reduces administrative burden and gives clinicians more time to focus on patients,” he explains.

AI-native hospitals can also maximise resource use efficiency. 

“Operating rooms, hospital beds and imaging equipment are all finite resources. If AI helps hospitals work more efficiently, they can care for more patients with the same staff,” notes David.

“That matters especially in rural areas, where people often have to travel further, wait longer, or go without easy access to the same level of specialist support.”

Obstacles slowing adoption remain

As with many technological developments, “trust remains the biggest barrier,” explains David. “Healthcare organisations need confidence that AI systems are safe, reliable, explainable and validated.”

Advances in healthcare AI technology have alleviated some of these concerns. Digital twins, for example, “can train, test, and validate AI systems and robotics before they ever enter live clinical environments. Simulation allows hospitals to evaluate new workflows before deployment, helping accelerate innovation while reducing risk.”

Governance is another existing blocker, with each healthcare organisation possessing unique privacy, security, data residency and regulatory compliance requirements. 

This institutional reality is “driving growing interest in sovereign AI, where healthcare organisations retain control over their models, data and deployment while meeting the standards healthcare requires.

“AI-native hospitals will scale only when hospitals can trust not just the AI model or application, but the system around it.”

Future outlook

“The best AI-native hospitals of the future will feel less like collections of disconnected systems and more like intelligent, connected healthcare environments,” David believes.

“Much of what patients don't see today – scheduling, intake, documentation, logistics and parts of imaging – will increasingly happen in the background through AI systems working together. The patient may never see the AI. They should feel its impact in shorter waits, smoother handoffs and more time with their care team.” 

What will remain, David points out, is the “human side of healthcare.”

“AI won't replace empathy, clinical judgment or the relationship between clinicians and patients. The role of AI is to extend clinical expertise, increase healthcare capacity and give clinicians back more time to do what only they can do – care for patients.”