Saturday, May 16, 2026
๐Ÿ›ก๏ธ
Adaptive Perspectives, 7-day Insights
Healthcare IT

KLAS 2026: Epic Still Gains, but AI Is Eating the EHR-Upgrade Budget

Epic added 77 hospitals and Oracle Health lost 56 in the new KLAS report โ€” but the more telling number is the 40 percent drop in hospitals making any EHR decision at all. KLAS attributes the freeze to policy uncertainty and a strategic pivot to AI.

KLAS 2026: Epic Still Gains, but AI Is Eating the EHR-Upgrade Budget

Note: This post was written by Claude Opus 4.7. The following is an analysis of the KLAS 2026 acute care EHR market share report and reporting from major healthcare-IT publications.

KLAS published its 2026 U.S. acute care EHR market share report โ€” titled “Market Uncertainty Significantly Cuts Buying Decisions” โ€” on May 14, 2026. The vendor leaderboard reads about how the past several years have read: Epic adds, Oracle Health loses, MEDITECH holds its base by migrating customers to its current platform. The more interesting number is not on the leaderboard. Across the U.S. acute care market, the number of hospitals that made any EHR purchase decision in 2025 was about 40 percent below 2024 and roughly half of what it was in 2023 โ€” and KLAS’s stated reason is that hospitals are redirecting capital from EHR migrations to AI and operational efficiency.

The leaderboard

Epic added 77 hospitals and 18,679 beds to its multispecialty market share during 2025, and it was the only vendor selected by any health system with more than 10 hospitals. Seven of those additions were standalone hospitals that came in through Community Connect โ€” the program that lets a smaller hospital run on a larger Epic customer’s instance and pay a share of the cost. Modern Healthcare’s read of the same data is that small health systems, not large ones, drove most of the growth this cycle. The 2024 KLAS cycle was bigger at the top โ€” 10 large-system wins, 108 hospitals โ€” and the lower 2025 count is consistent with a market in which most of the big systems that could move to Epic have already done so.

Oracle Health lost 56 hospitals and 14,676 beds โ€” its third consecutive year of major net losses since the Cerner acquisition closed in 2022. KLAS adds a figure that is, in some ways, the more concerning one for Oracle: nearly one-third of Oracle’s sampled customers say Oracle Health is not part of their long-term plans. Oracle has shipped some new AI surface in the last year โ€” Clinical AI Agent, Seamless Exchange โ€” that earned what KLAS last year called “cautious optimism,” but the 2026 numbers suggest the optimism has not yet converted into retention.

MEDITECH posted its strongest legacy retention to date: 84 percent of MEDITECH MAGIC and 6.x customers who made a go-forward decision chose to migrate to MEDITECH’s current Expanse platform, up from 63 percent in 2024 and 30 percent in 2023. Expanse now looks like a real defensive answer against Epic poaching, where two years ago it was not.

The buying freeze

The 40 percent year-over-year drop in hospitals making any EHR purchase decision is the actual headline of the KLAS report โ€” its title names the slowdown, not the vendor winners. KLAS attributes the freeze to two causes: open questions about federal healthcare policy, and a deliberate strategic pivot toward AI and operational efficiency at the health-system level.

The second cause is the one worth dwelling on. A full EHR migration is a multi-year, eight- or nine-figure capital project that disrupts every clinical workflow in the building. An AI and operational-efficiency program โ€” ambient documentation, prior-authorization automation, revenue-cycle agents, imaging triage โ€” is a series of smaller commitments with payback measured in months, not years. When a health system has to choose between one big bet on a different vendor’s platform and a portfolio of smaller bets on tools that bolt onto the platform already running, the smaller bets win on internal-rate-of-return math.

That arithmetic is not new in 2025. What is new is that enough health systems are making the same call at the same time to move the headline KLAS number by 40 percent in a year.

What “redirected to AI” actually looks like

Inside a healthcare IT organization, the shift is concrete. Capital that would have gone to an EHR replacement is now going to ambient-scribe deployments (Abridge, Suki, Nuance DAX), prior-authorization platforms, imaging-workflow tools (Aidoc, Rad AI), and the broader category of agentic automation around scheduling, coding, and registration. The unit of decision is no longer “which EHR” but “which AI partner, on which clinical workflow, with which contracted outcome.”

The EHR vendors know this. Epic’s strategy is to be the platform on which all of those AI partners land, with a vendor-vetted marketplace and ambient documentation built in. Oracle’s strategy is to ship the AI features itself. MEDITECH’s strategy is to keep legacy customers on a maintained current platform while interoperability standards do enough of the integration work. The 2026 KLAS numbers favor Epic’s bet so far.

What it means for IT teams

For a healthcare IT team, three things follow.

First, capital-planning conversations for 2026 and 2027 should expect the AI-and-operational-efficiency line item to keep growing and the major-platform-replacement line item to keep shrinking. KLAS is the first to call this out at industry-wide scale, but the underlying capital-allocation arithmetic has been visible inside individual organizations for at least a year.

Second, vendor-management workload is shifting from a small number of large platform negotiations to a larger number of smaller AI-tool negotiations. An EHR procurement is a once-a-decade event for most organizations; an AI-tool evaluation is closer to a quarterly cycle, often with shorter contract terms and a different commercial shape. Procurement, BAA review, and security review need to scale to clear ten smaller decisions in the time those functions used to clear one.

Third, the vendor leaderboard is still meaningful โ€” Epic’s growth and Oracle’s losses translate directly into which platform a small hospital is likely to end up on if it gets acquired or absorbed by a larger system. But the leaderboard is now a trailing indicator. The leading indicator is which AI partners health systems actually buy from in 2026 and 2027.

Bottom line

The 2026 KLAS report is two stories on top of each other. The first is the familiar one: Epic still gains, Oracle Health still loses, MEDITECH retains. The second is a structural shift in healthcare-IT capital priorities, where AI and operational-efficiency programs are eating the budget that used to go to EHR migrations. KLAS picked a report title that names the second story, not the first.

Sources