The Number That Explains Everything
Somewhere in America right now — in a fluorescent-lit exam room in Cleveland, a neonatal ICU in Houston, an oncology clinic in Boston, a rural urgent care in Montana — a doctor is clicking through a patient's chart. The interface is familiar: a particular shade of blue, a specific arrangement of tabs, a distinctive way of rendering lab values and medication lists. The odds are overwhelming that this doctor is clicking through Epic.
Three hundred and twenty-five million patients. That's the number Epic Systems claims for its current electronic health record database — roughly the population of the entire United States, minus children too young to have seen a physician and adults who have somehow avoided the healthcare system entirely. The company holds, by its own accounting, medical records covering approximately 90% of all Americans. It commands 42% of the acute care hospital market. Among large academic medical centers — the institutions that train the next generation of physicians, that conduct the clinical trials, that treat the sickest patients — Epic's share is closer to 58%. In 2024 alone, the company added 176 facilities and nearly 30,000 hospital beds to its network, its largest annual gain on record.
And yet Epic Systems has never been public. It has never taken a dollar of venture capital. It has never made an acquisition. It has no debt. It has 14,000 employees working on a 1,670-acre campus in Verona, Wisconsin — population 16,400 — where buildings are themed after
The Wizard of Oz,
Harry Potter, and
Charlie and the Chocolate Factory, and where a yellow brick road winds through emerald-green hallways. Its 82-year-old founder and CEO,
Judy Faulkner, took the stage at the company's 2024 annual Users Group Meeting dressed as a swan, complete with a plume of feathers in her hair.
This is the company that runs American healthcare's central nervous system.
By the Numbers
The Epic Empire
$5.7BRevenue in 2024
325MPatient records worldwide
42%U.S. acute care hospital market share
3,300+Hospitals on Epic
71,000+Clinics on Epic
14,000Employees
$0Venture capital raised, ever
1979Year founded in a Wisconsin basement
A Basement on University Avenue
The origin is almost too perfect in its modesty. In 1979, Judith Faulkner — a computer programmer with a master's degree from the University of Wisconsin-Madison, daughter of a pharmacist father and a mother who directed Oregon Physicians for Social Responsibility — co-founded Human Services Computing, Inc. in a basement at 2020 University Avenue in Madison, Wisconsin. The startup capital was $70,000, raised from friends and family. The staff consisted of a handful of part-time employees. Faulkner wrote the original code herself, on a Data General Eclipse 16-bit minicomputer about the size of a refrigerator, which used magnetic tape for storage.
The idea had germinated in graduate school, where Faulkner had taken what she believes may have been the first "computers in medicine" course ever offered in the world. One of her professors, Warner
Slack — a pioneer of computerized patient interviewing — asked her to join his research team. She said yes. The intersection captivated her: healthcare, one of the most consequential domains of human activity, meeting information technology, a field that was mentally stimulating and creatively boundless. "Sometimes I think it's a river you're floating down," Faulkner later told the
Capital Times. "Life just takes you different places."
Her co-founder, Dr. John Greist, a psychiatrist, shared an early vision for applying computing to patient data. But within four years, a fundamental disagreement surfaced. Greist wanted to bring in venture capital, to build faster. Faulkner refused. The argument was not merely tactical — it was philosophical, and it would define the next four decades. Greist stepped down from the board in 1983, retaining some shares, and Faulkner was left in sole command.
The company she began building was, from the outset, an expression of a particular personality: disciplined, idiosyncratic, allergic to outside influence, obsessive about customer service, and stubbornly committed to writing every line of code in-house. It was renamed Epic Systems Corporation in 1983, the same year it introduced its first patient scheduling software. By 1992, the company launched EpicCare, the industry's first Windows-based electronic medical records system — a bet on a platform that most of healthcare IT was still ignoring.
If we take good care of our customers, they take good care of the patients, and our job is to help the customers take good care of the patients. So it's indirect.
The Ten Commandments of Verona
Walk into any bathroom or breakroom on Epic's campus and you will encounter the company's Ten Commandments. The first three tell you almost everything:
- Do not go public.
- Do not acquire or be acquired.
- Software must work.
These are not aspirational statements. They are operating constraints that have shaped every major strategic decision the company has made. The prohibition against going public is not a preference — it is scripture. The refusal to acquire is not a limitation of ambition — it is a theory of software development. The insistence that software must work is not a platitude — it is a competitive weapon in an industry where implementations routinely fail, where rival systems crash for days, where hospitals have been forced to return to paper records after botched deployments.
Faulkner has been remarkably consistent in articulating why. "Why be owned by people whose interest is primarily return of equity?" she said in a 2025 interview with CNBC. The logic cascades: if you go public, you answer to shareholders on a quarterly cadence, which distorts long-term decision-making. If you acquire, you inherit different codebases, different cultures, different technical debt — and you lose the coherence that comes from building everything yourself. If your software doesn't work, nothing else matters, because a hospital system installation is, as one observer put it, "like doing a massive house renovation while the house is still occupied."
The commandments create a self-reinforcing system. Because Epic doesn't go public, it doesn't need to optimize for quarterly revenue recognition, which means it can invest in multi-year implementation projects without Wall Street pressure. Because it doesn't acquire, it maintains a unified codebase, which means fewer integration failures and more reliable deployments. Because it builds everything in-house, it controls quality. And because the software works — reliably, on time, with fewer catastrophic failures than competitors — hospitals keep choosing Epic, generating the revenue that allows the company to remain private and keep building in-house.
This is the flywheel. It has been spinning for 46 years.
The Stimulus That Changed Everything
For the first three decades of its existence, Epic grew steadily but slowly, serving a niche market of forward-thinking hospitals willing to invest in electronic records when most of healthcare still ran on paper. By 1995, the company had 100 clients and 125 employees. By 2000, revenue was approximately $47 million with a workforce of 396.
Then, in 2003, came the deal that announced Epic's arrival as a major force: Kaiser Permanente, the largest managed care consortium in the United States, selected Epic for its 30 hospitals and more than 11,000 physicians. The contract was enormous, and its signal value was greater still — if Kaiser, with its reputation for operational rigor, chose Epic, the product had passed the most demanding test in American healthcare.
But the true inflection point came from Washington. On February 17, 2009, President Obama signed the American Recovery and Reinvestment Act, which included the HITECH Act — $36 billion in federal incentives for hospitals and physicians to adopt electronic health records. The legislation transformed EHR adoption from a gradual, optional modernization into something approaching a mandate. In 2008, only 1.5% of U.S. hospitals had a comprehensive electronic records system implemented across all major clinical units. Less than 10% had even a basic system in one unit.
The money arrived, and the gold rush began.
How federal policy transformed Epic's addressable market
2003Kaiser Permanente selects Epic for 30 hospitals and 11,000+ physicians — the signature deal of Epic's pre-HITECH era.
2008Epic revenue reaches $600M, up from $500M in 2007. Only 1.5% of U.S. hospitals have comprehensive EHR systems.
2009HITECH Act passes, providing $36B in federal incentives for EHR adoption. The entire industry accelerates.
201475.5% of U.S. hospitals have at least a basic electronic records system, up from under 10% in 2008.
2016Epic employs approximately 9,500 people and earns revenues of roughly $1.8 billion.
2024Epic revenue hits $5.7 billion. The company holds 42% of the acute care hospital market.
A prevalent narrative holds that Epic engineered the HITECH Act for its own benefit — that Faulkner, through political connections and lobbying, manipulated the regulatory framework to favor Epic's products. The reality is more nuanced and, arguably, more interesting. HITECH didn't specifically advantage Epic over competitors. The certification requirements applied equally to all vendors. What HITECH did was massively expand the addressable market in a compressed timeframe, and in a sudden gold rush, the vendor with the best reputation for reliability and the most battle-tested product at large hospitals won a disproportionate share. Epic was winning approximately 40% of new contracts at major hospitals even before HITECH passed, according to Klas Enterprises. The stimulus didn't create Epic's advantage. It amplified it.
By 2008, Epic's revenue was $600 million. By 2016, it had tripled to $1.8 billion. By 2024, it reached $5.7 billion. The curve is not exponential — it is the steady, compounding growth of a company that captures an increasing share of an expanding market, one multi-million-dollar hospital contract at a time.
The Woman in the Basement
To understand Epic, you have to understand Faulkner — which is difficult, because she doesn't especially want to be understood. She rarely gives interviews. She has never written a memoir. The company, for years, had no public relations department at all; calls to Epic were returned, when they were returned, by an outside agency. Forbes once tried to verify her age and couldn't. In 2009, Faulkner was described as "traveling and unavailable" when a reporter sought comment. The company once put up a billboard that read: "Marketing Sucks ... Epic Systems."
What is known: born Judith Greenfield in August 1943, raised in the Erlton neighborhood of Cherry Hill, New Jersey. Graduated from Moorestown Friends School in 1961. Bachelor's degree from Dickinson College. Master's in computer science from the University of Wisconsin-Madison — the degree that led to the Warner Slack connection and, ultimately, to everything. Married to Dr. Gordon Faulkner, a pediatrician. Three children, none of whom work at Epic. Liberal politics. A collector of quirky artwork for the campus. An obsessive reader of science fiction. Forbes estimates her net worth at $7.8 billion, based on an estimated 43% ownership of Epic. In 2015, she signed the Giving Pledge, committing 99% of her assets to philanthropy. She reportedly sells approximately $100 million of her shares annually back to the company, with proceeds going to the Roots & Wings Foundation, a charitable organization she established with her husband.
I've described her as a female cross between
Bill Gates and Willy Wonka.
The Willy Wonka comparison lands because of the campus, but the Gates comparison is more analytically precise. Like Gates, Faulkner is a programmer-CEO who built a category-defining software company, maintained an iron grip on product direction, refused to cede strategic control to outside investors, and proved willing to compete ruthlessly behind a veneer of mission-driven earnestness. Like Gates, she understood early that in enterprise software, the product that achieves critical mass becomes the standard — and that switching costs in her market are not merely high but effectively infinite. A hospital that has spent $20 to $30 million and two or more years implementing Epic's system, retraining thousands of clinicians, migrating millions of patient records, and rewiring every clinical workflow from pharmacy to radiology — that hospital is not switching. Period.
The succession question looms. Faulkner turned 82 in August 2025 and has said she still has more to accomplish as CEO. She has split her stock into voting shares (which cannot be sold and have gone into a trust controlled by family and employees) and nonvoting shares (which will go to the Roots & Wings Foundation). The structure ensures that Epic will not be acquired after her departure — the voting control prevents it. In August 2024, she named Sumit Rana as president, and sources have identified him as the likely successor. But Faulkner shows no sign of stepping aside. At 81, she dressed as a swan. At 82, she is still running the company she started with $70,000 and a refrigerator-sized computer.
The Cathedral and the Bazaar
The most consequential strategic choice Epic ever made — more consequential than staying private, more consequential than refusing acquisitions — is the decision to build everything itself. Every module. Every feature. Every integration. One codebase. One architecture. One development team in Verona, Wisconsin.
In software, this is the cathedral model — as opposed to the bazaar of open-source, modular, API-driven platforms where third-party developers build on top of a core system. Most of the technology industry has converged on the bazaar. Apple has its App Store. Salesforce has its AppExchange. Even Oracle, Epic's largest competitor in EHR, allows (and encourages) third-party development. Epic took the opposite path.
The reasons are partly historical. In the company's early years, Faulkner and her team had several formative experiences with partners and competitors who, in their view, stole or misappropriated Epic's intellectual property. These experiences — which included aggressive patent disputes and partnerships gone wrong — created a deep institutional suspicion of external developers. Epic became fiercely protective of its code, its documentation, and its interfaces. No direct sandbox access for developers. Proprietary API documentation. Until recently, an almost complete absence of direct partnerships.
The benefits of this approach are real. A unified codebase means that every module — scheduling, billing, clinical documentation, pharmacy, radiology, lab, patient portal — talks to every other module natively, without the fragile middleware and integration layers that plague multi-vendor hospital systems. When a physician in an Epic hospital orders a medication, the pharmacy module sees it immediately. When a lab result comes back, it flows directly into the patient's chart. When a billing code is generated, it maps to the clinical documentation without manual reconciliation. This seamless integration is Epic's core product advantage, and it exists precisely because Epic builds everything itself.
The costs are also real. Epic's user interface, designed in-house, has been criticized by clinicians for decades as clunky, click-heavy, and contributing to burnout. The company's resistance to third-party innovation means that startups with better ideas for specific clinical workflows — ambient listening for clinical documentation, specialized AI tools for radiology, more intuitive patient-facing interfaces — face enormous barriers to integration. And Epic's control over its platform gives it extraordinary power over the healthcare technology ecosystem: it can effectively decide which innovations reach clinicians and which don't.
This is the paradox at the center of Epic's strategy. The same architectural choice that makes the software reliable also makes it rigid. The same control that ensures quality also stifles innovation. The same walls that protect intellectual property also create a walled garden that some hospitals, startups, and policy-makers increasingly view as a monopoly.
The Oracle at Redwood Shores
For most of its history, Epic's primary competitor was Cerner Corporation, founded in 1979 — the same year as Epic — in Kansas City, Missouri. Where Epic stayed private, Cerner went public in 1986. Where Epic refused acquisitions, Cerner grew through them. Where Epic built everything in Verona, Cerner assembled capabilities through purchases. The rivalry was fierce, personal, and defining for both companies.
Then, in 2022, Oracle acquired Cerner for approximately $28 billion — the largest acquisition in Oracle's history — and the competitive landscape transformed.
On paper, Oracle Health (as Cerner was rebranded) should be a formidable challenger. It contributed $5.9 billion in revenue to Oracle's fiscal 2023, putting it in the same revenue neighborhood as Epic. Oracle's founder,
Larry Ellison, has made healthcare a personal priority, spending extensive time at the company's annual Health Summit talking up AI-powered EHR capabilities. "Our user interface is so different than Epic's," Ellison told investors on a quarterly earnings call in September 2024.
The reality on the ground has been less impressive. In 2024, Oracle Health lost 74 hospital sites and 17,232 beds — while Epic gained 176 facilities and 29,399. For the first time, Oracle declined to share its list of new contracts with Klas Research, the industry's standard-bearer for vendor analysis, forcing Klas to estimate Oracle's market share. Customer loyalty ratings dropped more than 10 points after the Oracle acquisition, with hospital systems citing "poor partnership and a lack of follow-through on promises" as primary concerns.
Oracle's EHR software has been plagued by outages and rocky deployments. In one episode, Oracle engineers mistakenly caused a five-day software outage at several Community Health Systems hospitals, forcing facilities to activate downtime procedures and temporarily return to paper-based patient records. "Over the last decade, Epic has been the only vendor chosen by large health systems making go-forward EHR decisions," the Klas report concluded.
Beyond strictly technological considerations, Epic's reputation for customer partnership has brought them to the forefront of most EHR considerations.
The irony is thick. Oracle spent $28 billion to acquire Epic's chief rival, and the acquisition appears to have accelerated Cerner's decline. The integration challenges, the culture clash between Silicon Valley database engineers and Kansas City healthcare specialists, the distraction of merging systems — all of it created openings that Epic exploited. Ellison's ambition in healthcare is genuine, and Oracle's cloud infrastructure and AI capabilities are formidable. But enterprise healthcare software is not a database problem. It is a relationship problem, a workflow problem, a trust problem — and in those dimensions, Epic's 46-year head start and obsessive customer focus have proven nearly impossible to replicate.
The Cosmos Gambit
If Epic's first strategic era was building the best EHR, its second may be building the best data asset in healthcare.
In 2019, Epic launched Cosmos — a de-identified clinical data research platform that aggregates data from across Epic's customer base. The numbers are staggering: because Epic's software touches 325 million patient records across thousands of hospitals and clinics, Cosmos offers researchers, public health officials, and life sciences companies access to one of the largest longitudinal clinical datasets ever assembled. During the COVID-19 pandemic, Cosmos became a critical tool for tracking disease spread, treatment efficacy, and vaccination outcomes in near-real time.
The strategic implications go beyond research. As AI and machine learning reshape healthcare — in clinical decision support, drug discovery, population health management, and operational efficiency — the value of training data increases exponentially. Epic sits on a data asset that is, by at least one measure, unmatched: the most comprehensive, longitudinal, multi-specialty clinical dataset in the United States, possibly the world.
This is where Epic's flywheel takes on a new dimension. Every hospital that joins Epic adds its patient data to the network. Every patient record that flows through Epic's system enriches the Cosmos dataset. Every researcher or pharmaceutical company that uses Cosmos validates the platform's value, which makes Epic more attractive to the next hospital considering an EHR purchase. The data becomes a moat within a moat — a competitive advantage that compounds with every new customer, and one that no competitor can replicate without first achieving comparable market share.
The question is whether Epic will be an open steward of this data or a gatekeeper. The company's history suggests the latter. In 2024, Particle Health, a health data startup, filed an antitrust lawsuit against Epic, alleging that the company had blocked access to patient data through proprietary interfaces and restrictive practices. The lawsuit touched a nerve: the tension between Epic's role as the custodian of the nation's health data and its instincts as a private company protecting its competitive position is perhaps the defining tension of its next decade.
The Campus as Culture
Epic's 1,670-acre headquarters in Verona is not merely a workplace. It is a physical manifestation of Faulkner's worldview — a controlled environment designed to attract, retain, and acculturate a very specific type of employee.
The themed buildings — an Oz-inspired complex with a yellow brick road, a chocolate factory with giant chocolate chips at the entry, a Wizards Academy, a structure guarded by life-sized playing cards — are the most visible expression of Epic's culture, but they are not the most important. More telling are the practices: the mandatory monthly all-hands meetings in an underground auditorium called Deep Space (which some employees jokingly call "work church"), where executives discuss company news, objectives, and — remarkably — conduct grammar lessons on prepositions and pronoun usage. The promote-from-within policy, which means nearly every senior leader at Epic started as a recent college graduate in Verona. The expectation that employees live near campus. The rejection of remote work.
Faulkner noted during her 2024 keynote that the campus buildings and their upkeep account for 8% of the company's total expenses — a number she presented not as extravagance but as efficiency. It's cheaper to build in Verona than in San Francisco, Seattle, or New York. And in this small midwestern town, far from the churn and poaching of tech hubs, employees are less likely to leave.
The culture has its critics. Former employees describe an environment that can feel insular, even cultish — a sealed ecosystem where young recruits are immersed in Epic's way of doing things before they've had the opportunity to develop alternative frameworks. The promote-from-within policy, while creating deep institutional knowledge, also means that fresh perspectives from outside the Epic bubble are rare. And the centralization in Verona, while reducing costs and improving coordination, limits the diversity of the talent pool.
But the results are hard to argue with. Epic's employee retention, while not publicly disclosed, is widely regarded as exceptional for the tech industry. The company's ability to execute complex, multi-year hospital implementations — which require deep domain expertise, institutional continuity, and enormous coordination — is directly linked to a workforce that has spent years, sometimes decades, learning the Epic system from the inside. The campus is a factory for producing the specific type of human capital that Epic's business model requires.
Interoperability and Its Discontents
The single most politically charged word in healthcare technology is interoperability — the ability of different software systems to exchange patient data seamlessly. The ideal is that a patient's records should follow them wherever they go: from their primary care physician (on Epic) to an emergency room (on Oracle Health) to a specialist (on Athenahealth) to a rehab facility (on MEDITECH). The reality has been, for most of healthcare's digital era, a mess.
Epic's critics argue that the company has been the primary obstacle to interoperability. Ken Glueck, an executive vice president at Oracle, went so far as to call Faulkner "the single biggest obstacle to EHR interoperability" in a May 2024 blog post. The argument: Epic benefits from data lock-in. If patient data flows freely between systems, the switching costs that keep hospitals on Epic would diminish. Therefore, Epic has an incentive to make data exchange difficult — and, the critics allege, it has acted on that incentive through proprietary interfaces, restrictive data-sharing agreements, and foot-dragging on open standards.
Epic's defenders — and the company itself — counter that Epic has been an early adopter of interoperability standards, including FHIR (Fast Healthcare Interoperability Resources) and, more recently, TEFCA (Trusted Exchange Framework and Common Agreement), the federal government's framework for nationwide health data exchange. Epic's MyChart patient portal allows patients to access their records across institutions, and the company's Care Everywhere network facilitates data exchange between Epic sites and, increasingly, non-Epic sites.
The truth, as is often the case with Epic, sits in a productive ambiguity. The company has genuine incentives both to enable interoperability (customers demand it, regulators require it, and the Cosmos data platform benefits from broader data flows) and to constrain it (data lock-in is a real competitive advantage, and every third-party integration creates a potential attack surface). Faulkner has navigated this tension with characteristic pragmatism — moving far enough on interoperability to satisfy regulators and customers, while maintaining enough control over data flows to preserve Epic's strategic position.
[Judy Faulkner is] the single biggest obstacle to EHR interoperability.
The AI Frontier
Every major technology platform is, in 2025, making its AI bet. Epic is no exception — but its bet looks different from most.
The company has invested in AI across multiple dimensions: ambient clinical documentation (using AI to listen to doctor-patient conversations and automatically generate clinical notes), clinical decision support (surfacing evidence-based recommendations in real time), predictive analytics (identifying patients at risk of deterioration or readmission), and operational tools (optimizing scheduling, staffing, and revenue cycle management). At its 2024 Users Group Meeting, AI was a central theme, with Epic showcasing integrations with large language models and announcing partnerships with companies like Abridge, a startup specializing in AI-powered clinical documentation.
But the relationship between Epic and the AI startup ecosystem is, as Business Insider characterized it, "the messiest relationship in healthcare AI." Epic's platform control means it can bless or block AI startups that want to integrate with the EHR. Companies like Abridge have built businesses on Epic integration, only to find that Epic can develop competing functionality and route it through its own interface. The dynamic is familiar — it's the same platform leverage that Microsoft exercised in the 1990s, that Apple exercises through the App Store, that Amazon exercises in marketplace commerce. But in healthcare, the stakes are literally life and death, and the regulatory environment adds layers of complexity that don't exist in consumer technology.
Epic's data advantage through Cosmos positions it as potentially the most important AI company in healthcare — not because it builds the best models (it doesn't, and likely won't), but because it controls the data on which those models must be trained and the workflow into which they must be embedded. In AI, distribution and data are more valuable than model architecture. Epic has both.
The Weight of the System
The paradox of Epic's dominance is that the company that was supposed to modernize healthcare may now be the thing that healthcare needs to modernize around. Clinicians complain about click fatigue — the sheer volume of clicks required to document a patient encounter in Epic, a problem so widely documented that it has been linked to physician burnout. Startups complain about platform control — the difficulty of integrating novel tools into a system that was designed to be self-contained. Policy-makers worry about concentration — a single private company holding the health records of 90% of Americans, with no public shareholders, no SEC filings, and limited transparency into its operations.
Hospital consolidation has amplified Epic's position. When health systems merge — as they have at an accelerating rate over the past two decades — the combined entity typically standardizes on a single EHR platform. Because Epic is already the dominant player among large academic medical centers, it is almost always the platform that survives the merger. Each consolidation event further concentrates the market, creating a dynamic where Epic's share increases not just through new sales but through the structural reorganization of the American healthcare system.
And yet. Hospitals that use Epic consistently rate it higher than alternatives. The software, for all its flaws, works — in a domain where "working" means that medication orders don't get lost, that lab results reach the right physician, that billing codes match clinical documentation, that a patient's allergy information is visible to every caregiver they encounter. In a field littered with catastrophic implementations, where competing systems have caused medication errors, missed diagnoses, and — in at least one controversial study — higher death rates, Epic's reliability is not a marketing claim. It is a competitive moat.
The company doesn't follow a preordained budget. It has no board of directors in any conventional sense. Faulkner makes the major decisions. The Ten Commandments hang in the breakrooms. And the flywheel keeps spinning: better software attracts more hospitals, which generates more revenue, which funds more development, which produces better software, which attracts more hospitals.
A Swan in Verona
In August 2024, thousands of healthcare executives gathered on Epic's campus for the annual Users Group Meeting. Judy Faulkner, 81 years old, walked onto the stage of an arena-sized underground auditorium dressed in a swan costume. She told the assembled hospital CEOs, CIOs, and CMOs that the campus buildings account for 8% of Epic's total expenses. She reminded them that it's cheaper to build in Verona than in any tech hub in America. She talked about AI and interoperability and the future of clinical documentation. And then she went back to running the company she started in a basement 45 years earlier.
Somewhere in the audience was Sumit Rana, the newly named president, the man most likely to eventually succeed her. Somewhere in Oracle's headquarters in Austin, Texas, Larry Ellison was preparing his next pitch about the future of healthcare AI. Somewhere in a congressional office, a staffer was drafting language about health data interoperability. And somewhere in a hospital in Houston, or Cleveland, or Boston, a doctor was clicking through a patient's chart on a screen that was unmistakably, recognizably Epic.
Forty-two percent of the nation's hospital beds. Three hundred and twenty-five million patient records. $5.7 billion in revenue. Zero acquisitions. Zero investors. One founder, one campus, one codebase, one set of commandments printed in every bathroom.
The Data General Eclipse minicomputer is long gone. The basement on University Avenue is someone else's problem now. But the code Judy Faulkner started writing on that refrigerator-sized machine — rewritten, extended, rebuilt a thousand times over, running on modern infrastructure across thousands of hospitals — is, in a meaningful sense, still running. It holds the medical history of nearly every American who has ever seen a doctor. And it belongs, still, to one woman in a small town in Wisconsin who once dressed as a swan.
Epic Systems offers an operating manual for a type of company that barely exists in modern technology — the sovereign private enterprise, built without outside capital, unconstrained by public markets, and led by a single founder for nearly half a century. Its principles are not universally replicable. They are, in many cases, the precise opposite of what venture capital and Silicon Valley orthodoxy would recommend. That is what makes them interesting.
Table of Contents
- 1.Stay private to stay patient.
- 2.Build everything yourself.
- 3.Win the customer, not the press release.
- 4.Let the government create your market.
- 5.Make switching costs infinite.
- 6.Centralize the factory.
- 7.Turn your product into a data asset.
- 8.Write commandments, not strategy decks.
- 9.Let the competitor acquire its way into weakness.
- 10.Control the platform, control the ecosystem.
Principle 1
Stay private to stay patient.
Epic's decision to remain private is not merely a capital structure preference — it is the foundational architectural decision from which every other strategic choice follows. Without quarterly earnings pressure, Epic can invest in multi-year hospital implementations that would look terrible on a public P&L. Without institutional shareholders, Faulkner can make decisions that optimize for customer satisfaction and product quality over revenue growth. Without public market scrutiny, Epic can keep its financials, strategy, and competitive moves opaque to rivals.
The proof is in the contrast. Cerner went public in 1986, and by the early 2020s was grappling with activist investors, short-term performance pressure, and a CEO succession crisis that ultimately led to its sale to Oracle for $28 billion. The acquisition, which was supposed to strengthen Oracle's healthcare position, instead created integration chaos, customer defections, and a 10-point drop in loyalty scores. Cerner's public-market path led, with a certain grim inevitability, to its absorption by a larger entity whose interests were not aligned with Cerner's customers.
Epic, meanwhile, used the same four decades to compound revenue from $47 million to $5.7 billion, building everything organically, never diluting Faulkner's control, and never giving a competitor insight into its strategic direction through SEC filings.
Benefit: Freedom to optimize for 10-year outcomes over 10-quarter outcomes. In an industry where implementations take years and customer relationships span decades, this temporal freedom is an enormous structural advantage.
Tradeoff: No access to public equity markets for capital, limited liquidity for employees, and a dependency on internal cash generation that could become constraining if the company ever needed to invest beyond its means. The absence of external governance also means there is no check on founder decision-making if that decision-making deteriorates.
Tactic for operators: If your business has long implementation cycles, high switching costs, and customers who care more about reliability than novelty, seriously consider staying private longer than conventional wisdom suggests. The optionality of public markets is real, but so is the distortion.
Principle 2
Build everything yourself.
Epic's commitment to a unified, in-house codebase is the most operationally expensive and strategically valuable decision in the company's history. Every module — clinical documentation, pharmacy, billing, scheduling, radiology, lab, patient portal, population health — is built by Epic engineers in Verona, on a single architecture, with native integration between every component.
In 2008, a $20 to $30 million hospital contract might involve dozens of modules, all of which needed to interoperate seamlessly. At competitors using multi-vendor or acquisition-assembled systems, this integration was a perpetual source of failure. IBM made money simply advising hospitals on how to combine disparate software systems. Microsoft sold middleware to extract data from multiple platforms. Epic's pitch was: you don't need any of that. It all works together, because we built it all together.
This approach is predicated on a bet about complexity. Healthcare workflows are so deeply interconnected — a medication order triggers pharmacy checks, which triggers insurance verification, which triggers billing codes, which maps back to clinical documentation — that the integration layer is not a nice-to-have. It is the product. By building every piece, Epic doesn't just control quality. It is the integration layer. And in a domain where integration failures can literally kill patients, that control is worth the enormous R&D cost.
How Epic's approach contrasts with the industry
| Dimension | Epic | Typical Competitor |
|---|
| Codebase | Single, unified | Multi-vendor, acquired |
| Acquisitions | Zero, ever | Frequent, strategic |
| Integration | Native, built-in | Middleware-dependent |
| Third-party development | Highly restricted | Encouraged via APIs |
| Implementation reliability | Industry-leading | Variable |
Benefit: Product coherence that competitors cannot replicate through acquisition. The unified codebase is Epic's deepest moat — deeper than market share, deeper than data, deeper than brand.
Tradeoff: Enormous R&D cost, slower feature development in any given area than a focused point-solution competitor, and a dependency on internal talent that creates scaling constraints. The build-everything approach also means Epic's user interface reflects the sensibilities of Verona engineers rather than the broader design talent available in the market.
Tactic for operators: When your product's value proposition is integration across complex workflows, own the integration. Don't outsource your most important product dimension. The cost is high, but the compounding value of a coherent architecture is immense.
Principle 3
Win the customer, not the press release.
Epic once put up a billboard that said "Marketing Sucks ... Epic Systems." It does no advertising. It issues no press releases. It has roughly 180 to 3,300 customers (depending on whether you count individual hospitals or health systems), and it gets business almost entirely by word of mouth.
This anti-marketing stance is not affectation — it is strategy. Epic's customers are hospital CIOs, CMOs, and CEOs making $20 to $100 million purchasing decisions with multi-year implementation timelines. These are not impulse buyers influenced by brand advertising. They are sophisticated operators who call their peers at other hospitals and ask: "Does the software work? Did they deliver on time? How's the support?" Epic wins when those conversations go well. It loses when they don't. No amount of advertising substitutes.
They deliver what they promise.
The customer focus extends to Epic's implementation methodology, which is widely regarded as the most hands-on in the industry. Epic doesn't just sell software — it embeds teams of implementation specialists at client sites, provides extensive training, and maintains ongoing support relationships that can last for decades. The high prices that some customers cite are, in this context, not a bug but a feature — they fund the service model that ensures successful deployment.
Benefit: In enterprise markets with long sales cycles and high stakes, customer reference is the ultimate marketing channel. Epic has turned every successful implementation into a sales tool, creating a self-reinforcing loop of reputation and revenue.
Tradeoff: No marketing means no narrative control. When critics characterize Epic as a monopoly, a data gatekeeper, or an obstacle to innovation, the company has no established public communications apparatus to respond. The "Marketing Sucks" ethos that served Epic well when it was a scrappy challenger becomes a liability when it is the dominant incumbent under scrutiny.
Tactic for operators: Invest in customer success before you invest in marketing. In B2B enterprise sales, one delighted customer who answers a peer's phone call is worth ten whitepapers. Build the reference base first. The brand will follow.
Principle 4
Let the government create your market.
Epic did not lobby for HITECH. But it was ready when HITECH arrived. The company had spent 30 years building and refining electronic health records software when the federal government decided to spend $36 billion incentivizing hospitals to adopt that exact product category. Being prepared when policy opens a market is not the same as manipulating policy — but it is equally valuable.
The lesson is broader than healthcare. Across industries, government action regularly creates, expands, or restructures markets: clean energy subsidies, defense procurement, financial regulation, infrastructure spending. The companies that benefit most are not the ones that lobby hardest, but the ones that have been building the best product in anticipation of the shift.
Epic's pre-HITECH market share at major hospitals — approximately 40% of new contracts, according to Klas — meant that when the floodgates opened, Epic captured a disproportionate share of the surge. By 2014, 75.5% of U.S. hospitals had adopted electronic records, and Epic was the leading vendor by a widening margin.
Benefit: Government-driven market expansion provides a demand shock that rewards incumbents with the best products and implementation track records. There is no faster or cheaper way to grow a market than having a government pay your customers to buy your product category.
Tradeoff: Dependence on government policy is a double-edged sword. Policy giveth and policy taketh away. The same regulatory apparatus that funded EHR adoption now imposes interoperability requirements, data-sharing mandates, and certification standards that constrain Epic's strategic flexibility.
Tactic for operators: Study pending legislation and regulatory shifts in your industry. The companies that capture policy-driven demand are those that have spent years building the right product before the policy arrives. You cannot build readiness retrospectively.
Principle 5
Make switching costs infinite.
A typical Epic implementation at a large hospital system costs $20 to $100 million, takes one to three years, and requires retraining thousands of physicians, nurses, pharmacists, and administrative staff. The system becomes embedded in every clinical workflow, every billing process, every compliance function. Patient data — years or decades of medical history — lives in Epic's database. Interfaces to labs, pharmacies, imaging centers, and insurance companies are all configured to Epic's specifications.
This is not merely a high switching cost. It is, for practical purposes, an infinite switching cost. No hospital that has successfully implemented Epic will switch to a competitor unless that competitor offers an almost inconceivably superior product — and even then, the implementation risk and operational disruption might not be worth it.
Epic has amplified this lock-in through design choices that deepen entrenchment over time. The Cosmos data platform ties each hospital's data into a broader network that becomes more valuable with scale. The MyChart patient portal creates a consumer-facing brand association that patients expect at every provider. Clinical decision support tools trained on aggregate Epic data improve with each additional hospital in the network.
Benefit: Near-permanent customer retention. Epic's churn rate is essentially zero among large health systems. Lifetime customer value is measured in decades and hundreds of millions of dollars.
Tradeoff: The same lock-in that protects Epic's revenue base also generates resentment, regulatory scrutiny, and the antitrust allegations that are beginning to surface. Infinite switching costs are, from the customer's perspective, infinite captivity. The political risk of being perceived as a monopoly is real and growing.
Tactic for operators: Design your product to become more valuable to each customer over time, through data accumulation, workflow integration, and network effects. Switching costs should be a consequence of genuine value creation, not artificial barriers. The distinction is critical — one creates loyalty, the other creates lawsuits.
Principle 6
Centralize the factory.
Epic employs 14,000 people. Virtually all of them work on a single 1,670-acre campus in Verona, Wisconsin. In an era of remote work, distributed teams, and global talent acquisition, this centralization is deliberately anachronistic — and, Epic would argue, deliberately effective.
The benefits are structural. Co-location enables the kind of rapid, informal communication that distributed teams struggle to replicate. A software engineer working on the pharmacy module can walk across campus to discuss a data model with a colleague working on clinical documentation. An implementation specialist returning from a client site can brief the product team in person the same day. The grammar lessons at monthly all-hands meetings — absurd to outsiders — serve as a cultural bonding mechanism that would be impossible over Zoom.
And Verona itself is a competitive advantage. Real estate is cheap: Faulkner noted that the campus costs 8% of total expenses, a fraction of what a comparable footprint would cost in a tech hub. The talent pool is narrow but deep — the University of Wisconsin-Madison, a major research university, is 10 miles away, and Epic has become the region's dominant technology employer. The remoteness reduces poaching. The lack of big-city distractions reduces turnover.
Benefit: A unified, culturally cohesive workforce with deep institutional knowledge, lower real estate costs, reduced attrition, and the coordination advantages of co-location.
Tradeoff: A limited talent pool. Difficulty attracting senior hires from outside the Epic ecosystem. A homogeneity of perspective that can become an echo chamber. And a dependency on a single geographic location that creates concentration risk — though the probability of a natural disaster destroying Verona and the nation's healthcare IT infrastructure simultaneously is, admittedly, low.
Tactic for operators: If your product requires deep cross-functional coordination and long employee tenure, consider the advantages of centralization. Remote work optimizes for individual flexibility. Co-location optimizes for institutional capability. The right choice depends on which dimension matters more for your business.
Principle 7
Turn your product into a data asset.
Epic's Cosmos platform represents a strategic evolution from software vendor to data platform. By aggregating de-identified clinical data from 325 million patient records across thousands of institutions, Epic has built a data asset that grows more valuable with every new customer and every new patient encounter.
The implications for AI are profound. In healthcare, model performance is constrained by training data — the volume, quality, and diversity of clinical examples available to train algorithms. Epic's Cosmos dataset is arguably the largest, most diverse clinical dataset in the world. Any AI company that wants to build healthcare-specific models will need data of this quality and scale. Epic can either build those models itself, license the data to partners, or — most likely — use its data advantage to ensure that AI in healthcare flows through Epic's platform.
Benefit: A compounding, defensible asset that increases in value with every new customer. Data network effects create a secondary moat around Epic's primary software moat.
Tradeoff: Data stewardship carries immense responsibility. Patient data privacy is legally protected and politically sensitive. Any breach, misuse, or perceived overreach could trigger regulatory backlash, customer defections, and reputational damage that would take years to repair. The Particle Health antitrust lawsuit signals that the tension between Epic's data control and industry expectations of openness is already reaching the courts.
Tactic for operators: If your product generates data as a byproduct of its core function, invest early in aggregating, structuring, and deriving value from that data. The marginal cost of data collection is zero once the product is in use. The value compounds. But treat data stewardship as seriously as you treat product quality — because a data scandal can destroy a company faster than a product failure.
Principle 8
Write commandments, not strategy decks.
Most companies have mission statements. Epic has commandments. The distinction is not semantic. A mission statement is aspirational — it describes what a company hopes to be. A commandment is binding — it describes what a company will not do, regardless of circumstances. "Do not go public" is not a strategy. It is a constraint. And constraints, paradoxically, generate strategic clarity.
Faulkner's Ten Commandments function as a decision-making framework that operates independently of any individual strategic plan. When a private equity firm approaches about an investment, the answer is predetermined. When an acquisition target presents itself, the answer is predetermined. When a customer asks for a feature that would compromise software reliability, the answer is predetermined. The commandments eliminate entire categories of decision-making overhead, freeing management to focus on execution.
Benefit: Cultural coherence and strategic consistency over decades. In a company where the founder is 82 years old and succession is approaching, the commandments provide continuity that individual leadership cannot.
Tradeoff: Rigidity. A commandment that was correct in 1979 may not be correct in 2035. "Do not go public" made sense when Faulkner had 43% ownership and the company was cash-flow positive. If healthcare AI requires capital expenditure that exceeds internal generation, the commandment could become a constraint on competitiveness rather than a source of strength.
Tactic for operators: Identify your three to five non-negotiable constraints and codify them. Not as mission statements, but as things you will not do. The strategic value of knowing what you won't do is at least as great as knowing what you will.
Principle 9
Let the competitor acquire its way into weakness.
Oracle's $28 billion acquisition of Cerner in 2022 was supposed to create a formidable challenger to Epic. Instead, it accelerated Cerner's decline. Customer loyalty ratings dropped. Hospital sites defected. Outages and rocky deployments plagued the combined entity. Oracle engineers caused a five-day outage at multiple hospitals. For the first time, Oracle declined to share its new contract list with Klas Research.
Epic, which has never made a single acquisition, watched this unfold with something approaching satisfaction. The lesson is not that acquisitions always fail — many succeed. The lesson is that in complex enterprise software, where the product is deeply embedded in customer workflows and the competitive advantage is integration quality, acquisitions introduce precisely the kind of technical debt, cultural fragmentation, and execution distraction that erode the advantages the acquirer hoped to gain.
Benefit: Epic's zero-acquisition track record means zero integration risk, zero cultural conflict, and a unified codebase that has been refined over 46 years. While competitors assemble capabilities through M&A, Epic compounds capabilities through organic development.
Tradeoff: Organic-only growth is slow. It took Epic 25 years to reach $47 million in revenue. A well-executed acquisition strategy could have accelerated growth and expanded capabilities in areas where Epic is weak — analytics, consumer-facing design, international markets. The commandment against acquisition is only wise if Epic can build everything it needs faster than the market evolves.
Tactic for operators: Before acquiring, ask: will the integration cost exceed the capability gain? In industries where product coherence is the competitive advantage, the answer is often yes. Watch your competitor's acquisitions — the integration period is often your best window to take share.
Principle 10
Control the platform, control the ecosystem.
Epic is, in the language of platform economics, a gatekeeper. Its EHR is the operating system of the American hospital, and every clinical application — whether built by Epic or by a third party — must either integrate with Epic or be irrelevant. This gives Epic extraordinary power over the healthcare technology ecosystem: it can determine which innovations reach clinicians, on what terms, and at what cost.
The company has used this power cautiously but decisively. Its restrictive approach to third-party development — no direct sandbox access, proprietary API documentation, limited partnership programs — has drawn criticism from startups, interoperability advocates, and regulators. But it has also preserved the product coherence that is Epic's core advantage. Every integration point is a potential failure point. Every third-party application that touches the EHR introduces risk. By controlling the platform, Epic controls the risk.
The AI era will test this strategy. The pace of innovation in healthcare AI is faster than Epic can match internally. If Epic restricts AI integration too aggressively, it will frustrate customers who want access to the best tools. If it opens the platform too broadly, it risks the integration failures that define its competitors. The balance is delicate, and the stakes — for Epic, for hospitals, and for patients — are immense.
Benefit: Platform control enables quality control, strategic positioning, and long-term competitive advantage. In healthcare, where system failures can harm patients, the gatekeeper function has genuine clinical justification.
Tradeoff: Platform control generates resentment, antitrust risk, and the perception of monopolistic behavior. The Particle Health lawsuit, the Oracle blog posts, the congressional interest in health data interoperability — all of these are consequences of Epic's platform power. The company's next decade will be defined by how it balances control with openness.
Tactic for operators: If you are building a platform, understand that gatekeeper power is both an asset and a liability. Earn the right to control by ensuring that control serves your customers' interests, not just your own. The moment customers feel captive rather than served, the regulatory clock starts ticking.
Conclusion
The Architect's Dilemma
Epic's playbook is a masterclass in building durable competitive advantage through deliberate constraint. By refusing capital, acquisitions, and publicity, Faulkner created a company that is optimized for the specific demands of enterprise healthcare software: long implementation cycles, mission-critical reliability, deep workflow integration, and multi-decade customer relationships.
The playbook's greatest strength is also its greatest vulnerability. The same constraints that enabled 46 years of compounding advantage may become the constraints that prevent adaptation. Healthcare is entering an era of AI-driven transformation, consumer-centric care delivery, and regulatory pressure for data openness. Each of these trends challenges a different pillar of Epic's strategy — the closed platform, the build-everything approach, the Verona-centralized workforce.
Faulkner built an institution, not just a company. The question that haunts every institution is whether the commandments that created it will be the commandments that sustain it — or the ones that calcify it. The answer will be written not by Faulkner, who cannot lead forever, but by the successors who must decide which commandments are eternal and which were right for their time.
Part IIIBusiness Breakdown
The Business at a Glance
Current Vital Signs
Epic Systems, 2024
$5.7BAnnual revenue (2024)
~$4.9BAnnual revenue (2023)
~16%YoY revenue growth
14,000Employees
42%U.S. acute care hospital market share
3,300Hospitals on Epic
71,000Clinics on Epic
325MPatient records worldwide
Epic Systems is, by revenue, among the five largest private technology software and services companies in the United States. Its $5.7 billion in 2024 revenue represents a growth trajectory from $3.8 billion as recently as 2022 — a roughly 50% increase in two years, driven by continued hospital wins, expansion within existing customers, and the rollout of new products and services. The company is profitable, though it does not disclose margins. It has no debt, no outside investors, and generates sufficient cash flow to fund all operations, campus construction, and product development internally.
The company's strategic position is unprecedented in enterprise software: dominant market share in a mission-critical category with near-infinite switching costs, compounding data network effects, and a regulatory environment that effectively mandates the product category. The closest analogy might be Bloomberg in financial data — another private, founder-led company that built a terminal-centric platform with enormous switching costs and became the indispensable infrastructure of its industry.
How Epic Makes Money
Epic's revenue model combines perpetual software licensing, implementation services, ongoing support and maintenance, and increasingly, subscription-based and cloud-hosted offerings. The company does not disclose a detailed revenue breakdown, but the major streams can be identified from industry analysis and customer disclosures.
Epic's major revenue streams
| Revenue Stream | Description | Characteristics |
|---|
| Software Licensing | Perpetual licenses for EHR modules (clinical, billing, pharmacy, scheduling, etc.) | High-margin, upfront |
| Implementation Services | Multi-year deployment, configuration, training, and go-live support | Labor-intensive, high-value |
| Support & Maintenance | Ongoing technical support, software updates, and system optimization | Recurring, predictable |
| Hosting / Cloud Services |
The economics of an Epic deal are revealing. A contract to install electronic records at a 400-bed hospital might be worth $20 to $30 million over several years — and that figure was from 2009. Large health system deployments involving thousands of beds, dozens of facilities, and comprehensive module coverage can run into the hundreds of millions. The University of Michigan, Stanford, and other major academic centers have disclosed Epic implementation budgets well in excess of $100 million.
The critical insight is the revenue lifecycle: a large initial licensing and implementation fee (Year 1–3), followed by decades of recurring maintenance, support, and expansion revenue. Epic's existing customer base of 3,300 hospitals and 71,000 clinics represents an enormous annuity stream — even if Epic never won another new customer, the support and maintenance revenue from its installed base would sustain the company for years.
Unit economics are exceptionally favorable for Epic once a customer is live. The marginal cost of supporting an additional user on an existing Epic installation is low relative to the per-user license and maintenance fee. The marginal cost of adding a new module to an existing Epic deployment is low relative to the licensing revenue. And the marginal cost of including a customer's data in Cosmos is essentially zero, while the marginal value of that data to researchers and AI developers increases with every additional record.
Competitive Position and Moat
Epic's competitive moat is multi-layered and, in its current form, among the widest in enterprise software.
Epic vs. major EHR competitors
| Vendor | Market Share (Acute Care) | 2024 Net Bed Change | Ownership |
|---|
| Epic Systems | ~42% | +29,399 | Private, founder-controlled |
| Oracle Health (Cerner) | ~23% | -17,232 | Oracle subsidiary |
| MEDITECH | ~16% | Flat to modest gain | Private |
| Athenahealth | Small/ambulatory focus | N/A | Private (Bain/Hellman & Friedman) |
Moat Source 1: Switching Costs. As discussed, implementations cost tens to hundreds of millions of dollars, take years, and require total organizational transformation. No large health system has switched away from Epic once live.
Moat Source 2: Integration Depth. Epic's unified codebase means that every clinical, administrative, and financial workflow is natively connected. This integration quality — the reason "software must work" is commandment number three — is the product advantage that competitors assembled through acquisition cannot replicate.
Moat Source 3: Data Network Effects. Cosmos, MyChart, and Care Everywhere all become more valuable as more hospitals join the Epic network. A researcher using Cosmos benefits from every new hospital's data. A patient using MyChart benefits from seeing their records across every Epic provider. These network effects compound with scale.
Moat Source 4: Reputation and Reference Network. In a market where the sales cycle runs 12 to 24 months and purchasing decisions involve board-level risk assessment, Epic's 46-year track record of successful implementations is an asset no amount of marketing spend can replicate. Hospital CIOs call their peers. The peers say it works.
Moat Source 5: Hospital Consolidation. The structural reorganization of the U.S. healthcare system — through mergers, acquisitions, and affiliations — overwhelmingly benefits Epic. When two hospital systems merge, they standardize on one EHR. Because Epic dominates among the largest, most prestigious health systems, it is almost always the surviving platform.
Where the moat is vulnerable: Epic's dominance among large health systems does not extend to smaller community hospitals, rural facilities, and ambulatory care settings, where MEDITECH, Athenahealth, and others retain meaningful share. The interoperability push, if successful, could erode data lock-in over time. And if a truly disruptive AI-native clinical platform emerges — one that offers a fundamentally better physician experience — even infinite switching costs might eventually yield to clinical demand.
The Flywheel
Epic's competitive advantage compounds through a multi-stage reinforcing cycle:
How each element feeds the next
Step 1Epic builds reliable, integrated software in-house, maintaining a unified codebase with native cross-module integration.
Step 2Successful implementations at major hospitals generate peer-to-peer referrals and word-of-mouth reputation.
Step 3New hospital wins expand the installed base, generating licensing, implementation, and recurring maintenance revenue.
Step 4Revenue funds continued R&D investment without external capital, maintaining product quality and expanding the module portfolio.
Step 5Each new hospital adds patient data to the Cosmos network and expands MyChart's consumer reach, strengthening data network effects.
Step 6Hospital consolidation absorbs smaller systems into Epic-using networks, expanding market share without direct sales effort.
The flywheel's most powerful property is that several of its links operate independently of Epic's direct effort. Hospital consolidation (Step 6) is driven by macroeconomic forces in healthcare — declining reimbursement, regulatory burden, labor shortages — that have nothing to do with Epic's sales strategy. Data network effects (Step 5) compound automatically with every patient encounter. Peer referrals (Step 2) happen in conversations Epic doesn't control and can't observe. The flywheel spins partially under its own momentum.
Growth Drivers and Strategic Outlook
1. Continued Hospital Market Share Gains. Epic added 176 facilities and 29,399 beds in 2024 — its largest annual gain ever. The pipeline of hospitals switching from Oracle Health, MEDITECH, and legacy systems remains robust. Oracle's integration difficulties create a multi-year opportunity for Epic to capture defecting Cerner customers.
2. International Expansion. Epic's patient record count of 325 million includes growing international deployments. The company has won contracts in the UK (NHS trusts), the Netherlands, Denmark, Australia, and other countries. The international EHR market is earlier in its digitization cycle than the U.S., representing a large untapped addressable market.
3. AI and Clinical Intelligence. Epic's Cosmos dataset and EHR platform position give it a unique advantage in healthcare AI. The company is investing in ambient documentation, predictive analytics, clinical decision support, and operational AI. AI features that improve physician efficiency and reduce documentation burden could accelerate adoption and deepen engagement among existing customers.
4. Consumer Health / MyChart Expansion. MyChart is already one of the most widely used healthcare consumer applications in the United States. Expanding MyChart's capabilities — telehealth, scheduling, prescription management, insurance integration, wellness tracking — could create a consumer-facing business that complements the enterprise platform.
5. Data Monetization via Cosmos. The Cosmos platform is in its early stages as a revenue stream. As pharmaceutical companies, payers, public health agencies, and AI developers increasingly need clinical data, the value of Epic's aggregated dataset will grow. The potential TAM for healthcare data analytics is estimated in the tens of billions, and Epic controls the most comprehensive dataset.
Key Risks and Debates
1. Antitrust and Regulatory Scrutiny. The Particle Health lawsuit, filed in 2024, alleges that Epic engages in anticompetitive practices by blocking access to patient data. Oracle's public attacks on Epic's interoperability practices have attracted congressional attention. If regulators determine that Epic's platform control constitutes monopolistic behavior, mandatory data-sharing requirements or structural remedies could erode the company's competitive position. The risk is heightened by Epic's 42% acute care market share and 90% patient coverage, numbers that cross thresholds that historically attract antitrust attention.
2. Founder Succession. Judy Faulkner is 82. She has been CEO for 46 years. While Sumit Rana has been named president and is the likely successor, no leadership transition of this magnitude — from a founding CEO who is the company's culture, decision-making framework, and strategic vision — proceeds without risk. Faulkner's commandments provide structural continuity, but the commandments themselves may need reinterpretation in the post-Faulkner era.
3. Physician Burnout and UX Backlash. The relationship between EHR documentation burden and physician burnout is well-documented. Epic's click-heavy interface has been specifically cited in burnout studies. If a competitor — or a startup building on AI-native clinical documentation — delivers a dramatically better physician experience, the clinical community's frustration with Epic could create an opening. AI-powered ambient documentation may address this, but Epic is not the only company building it.
4. Oracle's Brand-New EHR. In October 2024, Oracle announced a brand-new, ground-up EHR built on modern cloud-native architecture with integrated AI. If Oracle executes — a significant if, given its track record — this represents the first potentially credible architectural challenge to Epic in a generation. Larry Ellison's personal commitment to healthcare, Oracle's cloud infrastructure, and the clean-sheet design could, over a 5–10 year horizon, create a product that is genuinely competitive. The risk is not immediate but it is real.
5. Data Privacy and Security. Epic holds the medical records of 325 million people. A data breach at Epic's scale would be the largest healthcare data incident in history. While Epic's track record on security is strong, the increasing sophistication of cyberattacks and the growing value of healthcare data make this a persistent, existential risk. The company's centralized architecture — a strength in many dimensions — means that a breach of the core system could expose the entire patient database simultaneously.
Why Epic Matters
Epic Systems is the most consequential software company that most Americans have never heard of, even though most Americans are in its database. It is proof that you can build a dominant technology enterprise without venture capital, without acquisitions, without going public, without moving to Silicon Valley, and without doing any marketing — if you are building in a domain where reliability matters more than novelty, where switching costs are structural rather than psychological, and where the customer's alternative to your product is not a competitor's product but a return to paper.
For operators, the lessons are specific and transferable. Build your moat through integration, not feature count. Invest in customer success before marketing. Codify your constraints, not just your aspirations. Let your competitors' acquisitions create your opportunities. And understand that in enterprise software, the most powerful sales channel is a phone call between two CIOs — one who has your product, and one who is considering it.
For investors, Epic represents the archetype of the business you cannot buy — and the competitive dynamics you must understand to evaluate the companies you can. Oracle's $28 billion bet on Cerner is, in significant part, a bet against Epic's flywheel. Every hospital system that has chosen Epic over Oracle in the past decade is a data point on whether that bet will pay off. The evidence so far suggests it will not.
The central tension of Epic's next chapter is whether the commandments that built the cathedral can sustain it. Healthcare is entering an era where the pace of innovation — AI, interoperability, consumer health, genomics — may outrun the capacity of any single company to build everything in-house. If Epic opens its platform, it risks the coherence that defines it. If it keeps the gates locked, it risks being routed by a more open ecosystem. The answer will determine whether Epic remains the operating system of American healthcare or becomes the mainframe that the next generation of builders works around.
In Verona, the yellow brick road still winds through emerald-green hallways. The grammar lessons continue in Deep Space. And somewhere in a server farm that holds the medical histories of 325 million Americans, the code Judy Faulkner started writing in 1979 is still running.
