elenabs/istock via getty images
AI, interoperability to reshape health IT infrastructure in 2026
Health IT leaders share top infrastructure priorities, including cloud migrations, data governance and security.
As AI adoption accelerates and interoperability requirements expand, health IT leaders are taking a closer look at whether their infrastructure can keep pace. The challenge in 2026, they say, is less about deploying new technologies and more about sustaining them. Several leaders noted that AI demand, interoperability mandates and EHR platform capability growth are forcing infrastructure decisions sooner rather than later.
"Infrastructure is no longer just a utility. It is a strategic enabler," said Joseph Gimigliano, chief technology officer at Northwell Health.
AI and data-sharing advances hold promise for improving efficiency and care coordination. To realize those gains, leaders say they are reassessing their infrastructure -- looking beyond individual tools to cloud capacity, network performance, data governance and resilience, while evaluating whether systems can securely support demands over the next 12 to 24 months.
Cloud and scale: Preparing for unpredictable AI demand
For many organizations, AI has exposed the limits of traditional infrastructure models. Multiple leaders say flexibility, rather than ownership, is now the priority.
At Seattle Children's, that realization recently led to a shift away from running infrastructure internally.
"About 12 months ago, we concluded that as a hospital system, our job was to take care of patients, not to do IT," said Zafar Chaudry, MD, the hospital's senior vice president and chief digital officer.
Instead of building new data center capacity, Seattle Children's moved its infrastructure and networking services to external cloud partners.
Chaudry reported that the move has eliminated procurement delays, staffing constraints and supply chain uncertainty -- barriers that were incompatible with growing clinician and research demand for AI experimentation.
Children's Hospital of Philadelphia (CHOP) is taking a similar approach.
"We are beginning to execute a large migration to cloud for many of our application workloads," said Shakeeb Akhter, senior vice president and chief digital and information officer of CHOP.
Cloud, he said, is foundational not just for AI, but for long-term scalability.
For Jefferson Health, AI scale planning is less about rapid expansion and more about cohesion.
"You've got to plan your infrastructure not to address a simple point solution, but what's required across the entire enterprise," said Luis Taveras, senior vice president and chief information officer.
Taveras said Jefferson is pursuing this cohesion through enterprise data architecture and a platform-first approach, organizing data into a centralized warehouse and prioritizing partners that can deliver multiple capabilities rather than one-off tools.
"I don't want to put together point solutions and end up with a thousand different AI solutions," he said. "That's going to be chaos."
Data readiness takes center stage
As health IT infrastructure evolves, leaders say limitations in data quality and governance are often more constraining than technical capacity when scaling AI.
"We are making investments in our infrastructure to ensure our data is 'AI-ready,'" Akhter said.
At CHOP, that work has included deploying a data quality platform and completing a major migration of its data warehouse to Snowflake, a cloud-based data platform, along with a transition to Power BI, a reporting and analytics tool, to support AI-augmented analytics. The goal, Akhter said, is not just access, but trust.
Taveras echoed concerns about data quality and trust.
"What you don't want is to provide an AI model or a dashboard and have people question the validity of the data," he said. "Once you lose trust, it's very hard to get it back."
At Northwell Health, Gimigliano emphasized that data governance is a prerequisite for safe AI adoption.
"As this data begins to feed large language models, the basic hygiene still holds true -- least privilege, audit trails and understanding who has access to what," he said. In practice, that means limiting data access to only what users or systems need and keeping clear records of how data is accessed and used.
These disciplines become even more critical as AI tools gain broader access to patient data and interoperability expands.
EHR as infrastructure, not just an application
Multiple leaders reported looking to their EHR platform as both infrastructure and a primary source for new AI capabilities.
"I look at Epic as the engine that runs the hospital," Gimigliano said, describing Northwell's Epic migration.
The health system recently completed its first wave, covering roughly 35,000 staff. The next wave, planned for May, will be "the largest Epic migration ever," he said.
For Gimigliano, the value of EHR modernization lies in standardization and embedded capability. Rather than bolting AI tools onto fragmented workflows, Northwell is prioritizing AI that is integrated directly into clinical and operational processes.
"When AI is part of the workflow, and you have the human in the loop, that's when we see productivity accelerate," he said.
Akhter echoed that platform-first approach at CHOP.
"We are evaluating the new AI features Epic is releasing and prioritizing them with our clinical and operational partners," he said.
Governance, security and resilience in the AI era
As data sharing expands and AI use cases multiply, healthcare leaders are prioritizing security and governance in infrastructure decisions.
At CHOP, Akhter said the organization launched a multidisciplinary AI governance committee in 2024 to review all AI use cases.
"All AI use cases are reviewed by our AI governance committee to ensure SAFE AI -- secure, accountable, fair and explainable," he said.
Beyond formal governance, leaders described managing AI risk by limiting access to enterprise data, favoring platform-embedded capabilities over standalone tools and exercising downtime and recovery scenarios before scaling new use cases.
Taveras emphasized that growth cannot come at the expense of cybersecurity fundamentals.
"Everything that we're doing is for naught if we get hit with a major cyberattack," he said, noting that recovery planning and drills remain top priorities at Jefferson Health.
Chaudry also highlighted security as a key benefit of shifting infrastructure to external partners.
"The amount of money that cloud providers put into security, small hospitals cannot," he said.
In addition, Northwell's Gimigliano stressed the importance of testing, not just planning.
"You can read all the books you want, but unless you actually exercise, you're not going to get fit," he said.
Similarly, testing downtime procedures and infrastructure through drills and simulations helps ensure systems will perform as expected when disruptions occur. The organization conducts regular disaster recovery tests, ransomware tabletop exercises and business continuity drills involving executive and clinical leadership.
Going into 2026, the definition of innovation is shifting. Progress is less about experimentation and more about building systems that can withstand constant change. Infrastructure, once treated as a background utility, is now setting the pace for what health systems can safely deploy -- and how far they can scale what comes next.
Elizabeth Stricker, BSN, RN, comes from a nursing and healthcare leadership background, and covers health technology and leadership trends for B2B audiences.
