The 2026 HIMSS Global Health Conference and Exhibition drew more than 24,000 healthcare and technology leaders to Las Vegas from March 9 through 12, delivering four days of programming that reflected a healthcare technology industry at an inflection point. Artificial intelligence, cybersecurity, interoperability, and workforce sustainability each commanded major conference real estate, but across all of them, the underlying theme was consistent: the time for proof has arrived, and the industry’s tolerance for pilot programs and vendor promises has run out.
HotSpot Take
HIMSS26 drew 24,000 health IT leaders to Las Vegas with one overriding message: the era of AI experimentation in healthcare is over. The receipts have arrived, the governance conversation has begun, and the conference’s keynotes — from Tesla’s playbook to a patient’s 38 broken bones — framed both the urgency and the humanity of what’s at stake.
Opening Keynotes Set the Tone: Scale, Principles, and What Matters
HIMSS26 welcomed more than 24,000 healthcare and technology leaders to Las Vegas from March 9–12, 2026, for four days of programming centered on AI deployment, cybersecurity resilience, and digital health transformation. Photo courtesy of HIMSS26.
The conference opened Tuesday with a two-part keynote that framed HIMSS26’s intellectual through-line. Jon McNeill, former president of Tesla and former COO of Lyft, now CEO of DVx Ventures, drew parallels between the methodology that scaled Tesla from $2 billion to $20 billion in revenue and the organizational discipline healthcare needs to do the same with digital transformation.
McNeill’s central argument was not about technology; it was about process. His “Algorithm” framework, shared in detail in his newly published book of the same name, begins with questioning every requirement, eliminating friction before introducing automation, and setting metrics that are specific and aggressive: not “improve patient satisfaction” but “cut diagnosis cycle time in half.” As Becker’s Hospital Review captured from the floor, both McNeill and his fellow opening keynote speaker, Dr. John Halamka of Mayo Clinic Platform, “barely mentioned AI in their talks,” a pointed observation at a conference otherwise saturated with the topic.
“You won’t see the term AI in my slides. I did that purposefully, because having done this for 40 years, what you know is the technologies will change, but the engineering principles will stay the same.” — Dr. John D. Halamka, President, Mayo Clinic Platform
Dr. Halamka’s address was deliberately framed around enduring engineering principles rather than specific technologies. “You won’t see the term AI in my slides,” he told the audience. “I did that purposefully, because having done this for 40 years, what you know is the technologies will change, but the engineering principles will stay the same.” His message: build for interoperability, modularity, and data liquidity first, and the AI that runs on top will take care of itself.
Halamka’s presentation drew on Mayo Clinic Platform’s work building a dataset of more than 10 million deidentified patient records spanning structured and unstructured data, genomes, wearable data, and imaging going back decades. That foundation, he argued, is what makes AI at Mayo clinically trustworthy rather than clinically risky. He raised a question that reverberated across conference sessions throughout the week: given how rapidly AI-augmented clinical capabilities are advancing, “this may soon become our standard of care,” and with that, the ethical calculus around not deploying proven AI tools may need to be reconsidered. HealthTech HotSpot covered the opening keynotes in detail at our HIMSS26 keynote recap.
Apple, CMS, and the Consumer Health Moment
Wednesday’s keynote brought Dr. Sumbul Ahmad Desai, Vice President of Health and Fitness at Apple, to the main stage for a fireside conversation on how consumer technology is reshaping clinical care and preventive health. Dr. Desai pointed to Apple’s partnerships with Emory Healthcare, Sharp HealthCare, Stanford Medicine, and Harvard as examples of how wearable data is being incorporated into clinical discovery and used to develop features with regulatory approval. Her presence on the HIMSS26 main stage signaled something broader: the boundary between consumer health technology and clinical health IT is dissolving, and health system leaders are increasingly having to account for the data and expectations patients bring in from their devices.
“Why couldn’t we start to introduce agentic AI for every beneficiary of Medicare? Telecom companies are doing this now. Banks are doing it.” — Dr. Mehmet Oz, Administrator, Centers for Medicare & Medicaid Services
Thursday morning’s keynote brought the federal government’s perspective directly to the floor. CMS Administrator Dr. Mehmet Oz, joined by senior CMS officials Amy Gleason and Kimberly Brandt, delivered a keynote framed around a “revolutionary vision for American healthcare transformation.” The most headline-generating moment came from Dr. Oz’s call for agentic AI at the Medicare scale: “Why couldn’t we, by the end of this year, start to introduce agentic AI for every beneficiary of Medicare?” he asked. “Telecom companies are doing this now. Banks are doing it. If you can buy a mortgage with agentic AI giving you advice, you should be able to use that same technology to help you pick which Medicare Advantage plan to use or which doctor to go to.”
The statement landed in a room that was receptive but not uncritical. A KFF survey conducted in the fall found that only 31 percent of Medicare beneficiaries ages 65 and older trust AI “a great deal” or “a fair amount” to access their medical records and provide advice, a gap between the federal ambition and the patient readiness that the industry will need to bridge. Dr. Oz’s keynote also addressed the CMS “Kill the Clipboard” initiative, pushing for modern identity tools that give Medicare beneficiaries secure control of their health data across care settings, a theme that resonated directly with the Samsung/b.well demonstration elsewhere on the conference floor.
The Exhibition Floor: Execution Replaces Experimentation
Across 900 exhibitors and the dedicated Artificial Intelligence Pavilion, the dominant HIMSS26 exhibition floor message echoed the keynotes: vendors who arrived with outcomes data from named health systems commanded attention; those who arrived with feature demonstrations and roadmaps found a more skeptical audience.
HealthTech HotSpot’s HIMSS26 coverage documented this shift across four category roundups. In the EHR segment, Epic’s Agent Factory, Oracle Health’s ED clinical AI agent, MEDITECH’s native ambient tools, and athenahealth’s interoperability push collectively demonstrated that EHR platforms are being repositioned as autonomous operating environments rather than documentation repositories. In the agentic AI category, Google Cloud’s enterprise partnerships, Waystar’s $15 billion in prevented claim denials, and the emergence of AI governance as a standalone product category reflected a market where autonomy and accountability are being developed in parallel. The virtual care roundup showed Stryker’s SmartHospital Platform and a wave of rural-focused AI models attacking access gaps from every angle. And the medical imaging roundup confirmed that cloud-native infrastructure and clinical AI analytics have moved from pilot programs to enterprise deployment at network scale.
A recurring observation from health system leaders on the floor captured the procurement reality: “We have over 20 AI tools, and I don’t know what half of them do,” one CIO told Digital Health News. The market is shifting from feature shopping toward platform consolidation and long-term vendor partnerships, a dynamic that favors established players with deep EHR integration over standalone point solutions.
Cybersecurity: The Uninvited Conference Story
HIMSS26’s cybersecurity conversation received an unwanted real-world illustration when Stryker, which had just launched its SmartHospital Platform at the conference, was simultaneously hit by a destructive cyberattack on March 11. A pro-Iran hacking group called Handala claimed responsibility, describing the attack as retaliation for U.S.-Israeli military strikes in Iran. The incident disrupted Stryker’s global Microsoft environment, affecting order processing, manufacturing, and shipping. CISA launched an investigation. Stryker stated that its medical products remained safe for patient use and that the incident was contained to its internal IT environment.
The Stryker attack arrived at a conference where cybersecurity had already been identified as a strategic imperative rather than an IT concern. A Cohesity cybersecurity report released at the conference found that 94 percent of healthcare organizations had paid a ransom in a cyberattack, despite 49 percent of healthcare leaders expressing confidence in their security posture. The gap between stated confidence and actual vulnerability was a recurring theme in cybersecurity sessions throughout the week, and the Stryker incident made it viscerally concrete.
Interoperability: From Compliance to Clinical Utility
The interoperability conversation at HIMSS26 was shaped by a meaningful shift in framing. ASTP/ONC offered policy updates at the conference and reaffirmed enforcement priorities around information blocking, a reminder that the regulatory baseline for data sharing continues to rise. But the more interesting conversation was about what interoperability enables rather than what it requires.
athenahealth’s launch of athenaConnect and its patient-facing MCP server, the Samsung/b.well “Kill the Clipboard” demonstration, and Dr. Oz’s data liquidity agenda all pushed toward the same destination: a healthcare system where patient data follows the patient rather than staying siloed in whatever institution last touched it. Mayo Clinic’s Dr. Halamka framed the challenge directly from the floor: “This happens quite often, you still have to have someone call over and say, ‘Can you fax me records?'” The gap between the interoperability promise and the operational reality remains real, but at HIMSS26 it was framed more as an engineering problem with known solutions than as an intractable systemic barrier.
Clinicians Take the Wheel on AI Evaluation
One of the quieter but more substantive threads running through HIMSS26 was the push to give clinicians a more formal role in evaluating the AI tools being deployed around them. In a session on Wednesday, Nabile Safdar, chief AI officer at Emory Healthcare, and Bernardo Bizzo, senior director of artificial intelligence at Mass General Brigham, presented the case for pooling health system data to build standardized benchmarks that any clinical team can use to evaluate AI performance in their environment.
The problem they described is structural. “The biggest challenge we have is that we lack benchmarks to assess how these tools are performing and how well they can help us,” Bizzo said. Their answer is the Healthcare AI Challenge, a multi-health-system collaborative that equips clinicians and digital health leaders with hands-on evaluation frameworks, looking beyond accuracy metrics to factors such as clinical utility and workflow fit.
The session arrived alongside newly published evidence of what happens when AI tools do clear clinical scrutiny. A study published in JAMA Network Open, based on pilot data from both Emory Healthcare and Mass General Brigham, found that ambient documentation technology was associated with a 30.7% absolute increase in documentation-related well-being at Emory Healthcare at 60 days, and a 21.2% absolute reduction in burnout prevalence at Mass General Brigham at 84 days. The study involved 1,430 clinicians across both systems. Researchers believe it is the largest study of its kind to specifically examine ambient AI’s effects on burnout.
That evidence base matters beyond the ambient AI market. It demonstrates that rigorous, clinician-centered evaluation produces data capable of moving the industry, and that health systems willing to invest in measurement infrastructure are building a competitive advantage in both vendor selection and workforce retention.
The governance side of that equation was addressed directly by Jane Moran, chief information and digital officer at Mass General Brigham, in a Monday keynote at the HIMSS26 AI in Healthcare Forum. Moran described MGB’s path from ad hoc AI adoption toward a formal governance structure, including an AI executive committee and an operating committee focused on safety and ethics. Her framing of resource discipline was direct: “We can’t let 1,000 flowers grow. It’s too expensive to do that. So we’ve reduced the number of programs that we’re working on, and we’ll do this handful of things, and once those are successful, then we’ll go on to the next.”
Health Equity: AI and the Language Access Gap
HIMSS26 also made space for a patient-centered conversation that often gets overshadowed by enterprise technology announcements: the impact of language barriers on care quality. A Wednesday session titled “Parlez-Vous Patient? How to Provide Accurate Patient Discharge Notes” brought together Tripp Partain, senior director of cloud engineering at Oracle, and Dr. Albert Villarin, vice president and chief medical officer at Nuvance-Northwell Health, to discuss how AI-driven translation tools are being used to deliver clearer discharge instructions for patients who speak languages other than English.
Language barriers remain a persistent challenge in healthcare, affecting approximately 26 million individuals in the United States. The clinical stakes are specific: research has shown that non-English-speaking patients experience higher readmission rates for conditions such as heart failure and acute myocardial infarction, in part because post-discharge instructions are not fully understood. “If we can’t communicate with patients, we’re in the wrong business,” Villarin said at the session.
The solution Partain and Villarin described is not generic translation but clinical-grade language AI: systems trained on medication terminology, medical codes, and HL7 and FHIR data standards so that dosage instructions and clinical documentation are rendered correctly in the patient’s preferred language. Nuvance-Northwell’s longer-term vision, as Villarin described it, is to embed language preference directly into registration workflows, so that discharge instructions and follow-up communications are automatically generated in the right language from the moment a patient enters the system.
The Human Dimension: Jeremy Renner’s Patient Perspective
The conference’s closing keynote on Thursday delivered what may have been its most memorable moment — not from a technology executive or policy leader, but from Jeremy Renner, the Oscar-nominated actor who survived a catastrophic accident on New Year’s Day 2023 that left him with 38 broken bones after a 14,000-pound snowcat ran over him.
“I wish there was some sort of way that data got passed around, especially in emergency situations. From dispatch to first responders.” — Jeremy Renner, actor and patient advocate
Speaking in conversation with ABC News Live anchor Linsey Davis, Renner described navigating the healthcare system as a patient: the emergency response, the ICU transfers, and the communication gaps between care teams as he moved from one facility to another. “I wish there was some sort of way that data got passed around, especially in emergency situations,” he said. “From dispatch to first responders. I started working with this company called RapidSOS that is just that — uses AI to collect the data that’s necessary and passes it to the people who need it and need it the most.”
Renner’s presence and message served as a grounding counterpoint to four days of product announcements and vendor metrics. “I hold all the people who saved my life with every breath and every step I take,” he said. The room of 24,000 health IT professionals, building the infrastructure he was describing, was left with a clear reminder of who all of it is ultimately for.
Change Management as the Deciding Variable
Technology procurement was not the dominant challenge described by health system leaders on the floor. Change management was. A session at HIMSS26 captured a perspective that recurred in multiple conversations throughout the week: the organizations seeing the most durable AI results are those that redesigned workflows before deploying tools, not after.
Donna Fortson, senior vice president and chief revenue officer at WellSpan Health, described inheriting a call center operation generating consistent negative feedback from both clinicians and patients, frustrated with wait times. Rather than overlaying AI onto the existing structure, WellSpan rebuilt the call center’s primary care workflows before deploying an agentic AI solution, and the results held. Dr. Amish Desai, vice president and chief medical officer for population health at Northwestern Medicine, described a similar sequencing discipline: patient engagement campaigns that previously required weeks to complete were finishing in days after Northwestern restructured the underlying process and then introduced agentic AI to execute it.
The framing that surfaced repeatedly at the conference was a deliberate inversion of how most organizations approach automation: identify the broken workflow first, simplify it before introducing technology, and automate last. Jon McNeill had made this argument from the main keynote stage on Tuesday morning. Health system operators were making it again from session panels, two to three days later. When the same principle arrives from both the opening keynote and the clinical operations floor, it is no longer a consultant’s talking point; it is the conference’s consensus.
What HIMSS26 Leaves Behind
HIMSS26 will be remembered as the conference where healthcare AI stopped asking for permission. The shift from “what AI can do” to “what AI is trusted to do” (as Google Cloud’s Aashima Gupta framed it during the conference) captures the maturation moment precisely. Outcomes data replaced roadmaps. Governance emerged as a product category. Federal policy pushed agentic AI from a vendor conversation to a patient-access conversation.
Underlying all of it was a workforce story. Clinicians are burning out at documented rates, and the data from Emory and Mass General Brigham suggest that the right AI tools, evaluated rigorously and deployed with workflow discipline, can meaningfully reverse that trend. The language access session reminded attendees that equitable deployment matters: AI that reaches only English-speaking patients in well-resourced systems is not yet a healthcare transformation.
The threads that defined HIMSS26 (AI at production scale, interoperability as infrastructure, cybersecurity as a patient safety issue, and change management as the deciding variable) are not separate conversations. They are the same conversation about whether the healthcare system can be trusted to operate the powerful tools now available to it. The HIMSS27 community will gather in Chicago in April 2027 to take stock of how far that trust has been earned.
Photo credit: HIMSS26
— This original article was created with AI support.
