๐๐ข๐ญ๐ญ ๐ฎ๐ฆ ๐๐ด๐ฉ๐ฎ๐ข๐ฆ๐ญ. ๐๐ฐ๐ฎ๐ฆ ๐บ๐ฆ๐ข๐ณ๐ด ๐ข๐จ๐ฐ - ๐ฏ๐ฆ๐ท๐ฆ๐ณ ๐ฎ๐ช๐ฏ๐ฅ ๐ฉ๐ฐ๐ธ ๐ญ๐ฐ๐ฏ๐จ ๐ฑ๐ณ๐ฆ๐ค๐ช๐ด๐ฆ๐ญ๐บ - ๐ฉ๐ข๐ท๐ช๐ฏ๐จ ๐ญ๐ช๐ต๐ต๐ญ๐ฆ ๐ฐ๐ณ ๐ฏ๐ฐ ๐ฎ๐ฐ๐ฏ๐ฆ๐บ ๐ช๐ฏ ๐ฎ๐บ ๐ฑ๐ถ๐ณ๐ด๐ฆ, ๐ข๐ฏ๐ฅ ๐ฏ๐ฐ๐ต๐ฉ๐ช๐ฏ๐จ ๐ฑ๐ข๐ณ๐ต๐ช๐ค๐ถ๐ญ๐ข๐ณ ๐ต๐ฐ ๐ช๐ฏ๐ต๐ฆ๐ณ๐ฆ๐ด๐ต ๐ฎ๐ฆ ๐ข๐ด๐ช๐ฅ๐ฆ ๐ง๐ณ๐ฐ๐ฎ ๐ข๐ฏ๐ต๐ช๐ต๐ณ๐ถ๐ด๐ต, ๐ ๐ต๐ฉ๐ฐ๐ถ๐จ๐ฉ๐ต ๐ ๐ธ๐ฐ๐ถ๐ญ๐ฅ ๐ด๐ข๐ช๐ญ ๐ข๐ฃ๐ฐ๐ถ๐ต ๐ข ๐ญ๐ช๐ต๐ต๐ญ๐ฆ ๐ข๐ฏ๐ฅ ๐ด๐ฆ๐ฆ ๐ต๐ฉ๐ฆ ๐ฐ๐ถ๐ต๐ฑ๐ข๐ต๐ช๐ฆ๐ฏ๐ต ๐ฑ๐ณ๐ฐ๐ท๐ช๐ฅ๐ฆ๐ณ ๐ฑ๐ข๐ณ๐ต ๐ฐ๐ง ๐ต๐ฉ๐ฆ ๐ธ๐ฐ๐ณ๐ญ๐ฅ.
Epic owns the enterprise. Its acute-care share climbed to 43.7% in 2025, up from 42.3% a year prior, and it keeps winning the few large decisions left. Their architectural approach of a single integrated monolith underpins their dominance in the enterprise: every department writing to one shared database, each new user compounding the value for everyone already on it.
But they have not won all markets. From There Will Be Bundling, we saw how system of record markets (and the EHR market) are typically fractal: concentrated at the top around two or three integrated systems of record, then fragmenting into hundreds of niche players as you move downmarket and workflow needs get more specific.
The long tail of solo practices or 5-provider groups has long been their white whale for a few reasons:
Epic's long, heavy implementations donโt match small practice staffing or expectations.
The EHRโs premium cost structure often prices these practices out entirely.
Most importantly, the main differentiator (a monolithic mega-database and associated intra-organizational network effects) that sell the enterprise need many departments and roles to matter. A solo practice has neither!
That segment belongs to athenahealth, eClinicalWorks, PracticeFusion, and a field of niche speciality specific SaaS. And so we see them chase it like health techโs Captain Ahab:
Community Connect (2007) lets a small org ride a host system's instance. Itโs been incredibly successful in aligning multiparty incentives: the host pulls affiliated practices onto its instance, deepening referral capture and earning a hosting stipend; the small practice gets Epic without the full price tag or implementation lift, plus clean data exchange with the partners it already leans on; and Epic books share in a segment it can't profitably serve itself. But thereโs a ceiling in that practice autonomy is sacrificed. Truly independent groups wanting no master have nowhere to go.
Epic Hosting (2015) runs the same full Epic, with the data center handed back to Epic. You still buy, license, and build your own instance, while Epic just operates the infrastructure in its own Verona and Rochester facilities. It removes the servers and the ops staff, not the cost or the implementation. Fittingly, it's a re-entry: Epic started in 1979 as a time-sharing firm, so vendor-operated computing is where the company began. The result: 162 health-system tenants on an
epichosted.comendpoints (30% of all branded Epic endpoints), with Johns Hopkins, Memorial Sloan Kettering, CommonSpirit, and Advocate Aurora on the list. It helps a customer big enough to own a full Epic, but doesnโt change the equation for a practice that can't.Sonnet (2018) was the slimmed-down build for small hospitals and physician groups. The logic was that Epic's implementation pain came from sheer surface area, so stripping modules down to a core feature set would shorten timelines and lower the price enough to compete with athenahealth and eClinicalWorks on their own turf. It never caught on. The orgs that wanted Epic wanted the whole thing, and the orgs that wanted lean and cheap had better-fit options already. After years of muted uptake, Sonnet was effectively shelved.
Which brings us to Garden Plot (2022), essentially Community Connect with Epic itself as the host. As a shared, cloud-hosted Epic tenant with third parties like Availity and Surescripts pre-wired, the idea has been that a small practice can land on Epic without forfeiting its neutrality. In a multi-hospital market, riding one systemโs Community Connect instance reads as picking a side and often serves as a simple precursor to acquisition.
From our exploration of the Garden Plot SMS page last summer, there were around 24 total organizations listed. Since then, that number has hardly budged, increasing to 28. Moreover, the last time the page was updated was nearly 9 months ago (on September 9, 2025). However, it could be that the SMS page is a weak proxy for the total count. If it only captures practices that opted into Epicโs stock texting program, it could undercount the real number.
