Epic Systems (MyChart)

Epic Systems is a privately held healthcare software company founded in 1979 by Judith Faulner in Madison, Wisconsin. It builds the electronic health record (EHR) platform used by the majority of major U.S. hospital systems—the clinical, billing, and patient-facing software that runs modern American healthcare. Despite doing no marketing, no sales outreach, never raising venture capital, never making acquisitions, and never losing a customer, Epic has grown to roughly $5.7 billion in annual revenue with dominant market share among large and academic health systems. The episode traces how a combination of architectural discipline, regulatory timing, customer obsession, and founder continuity turned a one-woman programming project into what may be the most durable vertical software company in history.

• The founder: Judith Faulner

  • Born in 1943 in New Jersey; grew up around her father’s pharmacy and soda fountain, and her mother—who was part of a group that shared the 1985 Nobel Peace Prize for anti-nuclear advocacy.
  • A math prodigy who taught herself Fortran in a week during a summer job at the University of Rochester’s particle physics lab in the mid-1960s, becoming one of the best programmers there.
  • Enrolled in the brand-new computer science PhD program at the University of Wisconsin–Madison, where she took a pioneering “Computers in Medicine” course taught by Dr. Warner Slack.
  • While working part-time for $5–10/hour building scheduling and data applications for UW’s medical center, she had an epiphany in her living room: a single integrated database at the center of a patient’s entire care journey. She called it Chronicles, and its core architecture still underlies Epic today.
  • Started the company in 1979 (originally named Human Services Computing) with $70,000 in bank debt and $70,000 from friends and family—the only primary capital Epic has ever raised. Pre-money valuation: $70,000.
  • At 81, she still runs the company. She has structured a purpose trust so that upon her death, voting control passes to a group of family, senior Epic leaders, and customer CEOs—with ironclad provisions that the company can never go public or be acquired, and the next CEO must be a longtime employee and software developer.

• Why healthcare records were a problem—and why digitization was inevitable

  • For most of the 20th century, medical records were paper-based, fragmented, and non-standardized. Efforts to standardize date to 1912 at Massachusetts General Hospital, but adoption was voluntary and doctors resisted being told how to document.
  • The real inflection point was 1965: Congress created Medicare and Medicaid. Suddenly, hospitals and doctors had to document care in standardized, auditable ways to get paid by third-party payers (insurance companies and the government). This created an existential need for electronic medical records tied to billing.
  • In the mid-1960s, only about 10–13% of healthcare facilities had any kind of electronic system. The infrastructure (mainframes, then minicomputers) was expensive and rare outside academic medical centers.

• The technical foundation: MUMPS and the single-database architecture

  • In 1966, a team at Massachusetts General Hospital built COSTAR (Computer Stored Ambulatory Record), the first real computerized medical record system. They created their own programming language, MUMPS (Massachusetts General Hospital Utility Multi-Programming System), because existing languages couldn’t handle the high-concurrency, multi-user transaction needs of a hospital.
  • MUMPS’s key innovation: the language and the database are integrated, making it efficient for constant reads and writes across many simultaneous users—critical when multiple departments are updating the same patient record.
  • Epic still uses MUMPS (and its modern descendant, Caché, made by InterSystems). This single-database architecture—where every application (clinical, billing, scheduling, patient portal) talks to one unified Chronicles database—is Epic’s core technical differentiator.

• Epic’s early years: slow, bootstrapped, and small (1979–1992)

  • For its first decade, Epic was a tiny Madison business. By 1988, it had only 24 customers and $1.5 million in revenue.
  • The company didn’t launch its billing module, Resolute, until 1987. Before that, it only handled clinical records—not the revenue-critical function that hospitals most needed.
  • In 1983, the company was renamed Epic Systems. The name reflects Faulner’s creative, literary sensibility: a patient’s life is an epic story.
  • Epic’s early strategy: target the most complex, upmarket customers—academic medical centers and large integrated delivery networks—because they had the computing infrastructure (Unix-based minicomputers) and the most demanding needs.

• The product comes together: Epicare, MyChart, and the integrated suite (1992–2003)

  • In 1992, Epic launched Epicare, the industry’s first Windows-based graphical EMR application. Before this, EMR access was through 80-character green-screen terminals. Now doctors and nurses could use a GUI on a PC.
  • Epicare initially covered only ambulatory (outpatient) care. The inpatient version didn’t launch until 2001.
  • In 2000, Epic launched MyChart, a web-based patient portal—years ahead of its time in a HIPAA-regulated industry. Patients could view records, message doctors, and (later) self-schedule appointments. Today MyChart has 191 million active users and is a major source of patient-side virality for Epic.
  • By 2001, Epic had the full suite: Chronicles (single database), Epicare Ambulatory, Epicare Inpatient, Resolute (billing), Cadence (scheduling), and MyChart. Revenue crossed $50 million with 88 health system customers.

• The watershed: winning Kaiser Permanente (2003)

  • Kaiser Permanente is the largest integrated managed-care system in the U.S. (30 hospitals, 400+ clinics, 11,000 physicians, 8.5 million patients in 2003). It’s both the insurance plan and the hospital system—closest thing to a single-payer system in America.
  • Kaiser initially tried to build its own EMR with IBM and Accenture. That project failed. They then put out an RFP that came down to Epic and Cerner (a much larger, public, nearly $1-billion-revenue company that had been built through dozens of acquisitions).
  • Faulner made a bold, self-serving argument: don’t split the contract between two vendors (Cerner for inpatient, Epic for outpatient). Pick one integrated system. She was a “pipsqueak” compared to Cerner, but she knew Epic could do both.
  • In a pivotal technical due diligence session, Epic’s team (led by then-president Carl D’vorak) pulled an all-nighter to build a detailed Excel model proving their system could handle Kaiser’s transaction volume. Cerner couldn’t prove the same.
  • Kaiser asked both companies for equity (warrants) as part of the deal. Cerner offered 10% of the company. Faulner refused—Epic would never give up equity to a customer. Epic won anyway.
  • The deal roughly doubled or tripled Epic’s revenue overnight (from ~$50M to $162M). After a successful go-live, Epic became the new gold standard. By 2007, revenue hit $500 million.

• The HITECH Act and “Meaningful Use” (2009): government pulls forward the future

  • In the wake of the 2008 financial crisis, the Obama administration passed the American Recovery and Reinvestment Act, which included the HITECH Act (Health Information Technology for Economic and Clinical Health). It allocated ~$36 billion in incentive payments to hospitals that adopted EHRs and used them in “meaningful” ways.
  • Hospitals received $44,000–$64,000 per physician for adopting qualifying EHRs. After the stimulus period, penalties kicked in for non-compliance.
  • Impact on Epic: This was a massive tailwind. When the government makes your product effectively free and penalizes customers for not using it, the most reliable, integrated vendor wins. Epic’s single-platform architecture meant fewer implementation risks and seamless compliance with meaningful-use requirements.
  • EHR adoption went from 9% of hospitals in 2009 to 95% by 2014.
  • But the side effects were severe:
    • “Meaningful use” was defined by Congress in ways that effectively legislated how doctors do their jobs—mandating specific fields, checkboxes, and workflows. Doctors now spend roughly 2 hours on data entry for every 1 hour of patient care.
    • The regulatory burden accelerated hospital consolidation (small practices couldn’t afford compliance), which ironically benefited Epic because its strategy had always been to target the largest, most complex systems.
    • Interoperability (data sharing between different EHR systems) was a stated goal but was not incentivized with the same force as adoption. It largely didn’t happen—conveniently for Epic, which had a closed, integrated system.
    • Obama himself later said the legislation didn’t live up to expectations: digitization succeeded, but digital transformation (cost reduction, efficiency) did not.

• Epic’s culture and campus: the “software factory”

  • Faulner calls Epic a “software factory”—it takes in smart young graduates and turns them into medical software developers. The company has almost no middle managers, no budgets in the traditional sense, very light titles, and no sales or marketing department (about 8 people who only handle inbound requests).
  • Epic’s Verona, Wisconsin campus spans 1,700 acres with 89 buildings, designed with whimsical themes (Wizard of Oz, Harry Potter, Alice in Wonderland) by the firm that renovated Disneyland California Adventure. It includes an 11,400-seat underground auditorium (Deep Space), a replica of Grand Central Station, a treehouse, and a barn.
  • The campus is a recruiting tool: it makes Epic feel like a fun, college-like paradise for smart young graduates in rural Wisconsin.
  • Epic’s Ten Commandments are posted in every bathroom: (1) Do not go public, (2) Do not acquire or be acquired, (3) Software must work, (4) Reality equals expectations, (5) Keep commitments, (6) Focus on competency, (7) Have standards, (8) Have courage, (9) Teach philosophy and culture, (10) Be frugal. None mention healthcare—they’re about how to run a company.
  • Hiring is test-based: Faulner’s teenage son wrote the original programming test used for 18 years. Today, every role has a standardized test. Many hires never interview—they just score well on the test.
  • Every employee does “immersion trips”—spending time in clinical settings (operating rooms, etc.) to observe workflows firsthand.
  • Software methodology is hyper-prescriptive: developers must fix their own bugs immediately to prevent compounding errors. The goal is zero bugs because lives are on the line.
  • The culture is “up or out”—aggressive performance management with significant early-career attrition, but those who stay are deeply trusted. It resembles early Microsoft in intensity and camaraderie.

• Customer obsession and the “BFF” model

  • Every Epic customer has a dedicated technical specialist team for each product they use, plus a single BFF (Best Friend Forever) whose sole job is that customer’s success.
  • Epic sends customers annual “report cards” grading them on usage and best practices, benchmarked against peer hospital systems, sent directly to the CEO, CIO, and CFO.
  • Epic is highly opinionated about implementation: customers are strongly encouraged to use standard configurations (not custom builds) so updates and interoperability work smoothly. Discounts are tied to how standard your implementation is.
  • Epic has never raised prices on MyChart and keeps annual price increases around 2% (below inflation). They recommend third-party tools to customers as stopgaps, then build their own versions and bundle them for free—a bundling strategy reminiscent of Microsoft.
  • Despite being expensive, Epic has never lost a customer in 47 years (one customer left for 6 months and came back). There is no “getting fired for buying Epic” equivalent of the old IBM saying—it’s beyond that.

• Competitors’ struggles: Cerner, Oracle, and the DoD/VA contracts

  • Cerner (founded 1979, same year as Epic) took the opposite path: dozens of acquisitions, going public, pursuing large government contracts. It was built from 36+ merged companies.
  • Cerner won the Department of Defense ($4.3 billion, 2015) and VA ($10 billion, 2017) EHR contracts. Both went massively over budget and behind schedule. The DoD system only fully went live in 2024 (9 years). The VA system is still not fully live, with the earliest estimate of 2031 (14 years).
  • Cerner’s founder Neil Patterson died of cancer in 2017. The company cycled through leadership and was acquired by Oracle in 2021 for $28 billion. Oracle no longer breaks out Cerner’s financials; analysts consider it one of Oracle’s worst acquisitions.
  • Epic, unburdened by these government contract quagmires, kept winning every major health system RFP: Mayo Clinic, Johns Hopkins, Mass General/Harvard (Partners Healthcare), UCSF, Cedars-Sinai, Intermountain Health, CommonSpirit Health, and international expansions (UK’s NHS, starting with Guy’s and St. Thomas’ in London).

• Cosmos: the data platform and the AI future

  • Epic launched Cosmos, a massive anonymized research database aggregating data from Epic customer instances. It contains 295 million patients and 15 billion individual patient encounters.
  • Any Epic customer can opt in and query Cosmos for free. Use cases include: finding “lookalike” patients for rare disease diagnosis, clinical trial recruitment, and population health research. It’s the realization of the long-held dream of what digital records could enable.
  • Ambient AI scribes (from Microsoft/Nuance, Abridge, Suki) are emerging as a major innovation—AI that listens to doctor-patient conversations and auto-populates the EHR, reducing the documentation burden that causes physician burnout. These are Epic partners, and Epic is the “choke point” determining which software innovations reach hospitals.
  • The “mega bull case” for AI in healthcare: ambient AI could eventually make the EHR interface itself fade into the background, with AI handling data capture, billing adjudication, and prior authorizations automatically—potentially reducing the $800 billion of annual waste in the U.S. healthcare system.

• The business today

  • Revenue: $5.7 billion (2024), up 13% YoY; 16% growth from 2023 to 2024.
  • Customers: 607 health systems, 3,200 hospitals, 590,000 physicians, 495,000 staffed beds, 325 million patients worldwide (280 million in the U.S.).
  • Market share: 42% of hospitals, but the largest and most prestigious ones. 79% of Americans receive care at an Epic-connected facility. 90% of medical students train on Epic.
  • Employees: 14,000. R&D is ~35% of operating expenses (comparable to Apple, higher than Oracle). Zero spent on sales and marketing.
  • Profitability: Estimated EBITDA margins of 30–35%, implying ~$1.7 billion EBITDA.
  • Pricing: Average customer pays ~$10 million/year. This is a small fraction of the total cost of an EHR implementation (which can run hundreds of millions to billions for large systems when you include consultants, internal staff, and lost productivity during transition).
  • International: ~10–15% of business, growing (UK, Netherlands, Middle East, etc.), but most international markets have lower willingness to pay.

• Valuation and Faulner’s wealth

  • Forbes estimated Faulner’s net worth at $7.6 billion in 2024 (implying a ~$15 billion company valuation), which the hosts consider far too low.
  • Using Cerner’s acquisition multiple (30x EBITDA) as a conservative benchmark implies a ~$51 billion valuation. A 9x revenue multiple (reasonable for durable software) gives a similar number.
  • The hosts estimate that if Epic were public—which it never will be—it would likely be valued in the $100 billion range, given its durability, strategic value, and data assets. This would make Faulner’s shares worth ~$50 billion, making her the wealthiest self-made woman in history and one of the most successful entrepreneurs period.
  • For comparison: UnitedHealth Group does $400 billion in revenue and $35 billion in EBITDA—six times Epic’s entire revenue in profit dollars. Epic is a linchpin in the system without capturing a large share of the industry’s profit pool.

• Power analysis: why Epic is so defensible

  • Switching costs: This is the highest-switching-cost software in existence. Replacing Epic means replacing the “central nervous system” of a hospital. Implementations cost hundreds of millions to billions and take years. If it fails, people die.
  • Scale economies: Epic’s breadth of modules (300+ applications) is only possible because it amortizes development across 607 large health systems. No new entrant can afford to build this much software.
  • Network economies: Most major hospitals use Epic, so new customers benefit from seamless data sharing (Care Everywhere: 20 million patient records exchanged daily). Doctors are trained on Epic in medical school (90%), creating a labor-market network effect.
  • Branding: “Don’t get fired for buying Epic.” The most prestigious institutions use it, and it’s known to work reliably.
  • Process power: Epic’s proprietary development methodology, training pipeline, and use of MUMPS/Caché create a unique organizational capability that is hard to replicate.

• Playbook themes

  • R&D as the only spend that compounds: 35% of operating expenses go to R&D. Zero on sales and marketing. In software, R&D compounds; other forms of spend don’t.
  • Hill-climbing growth: Faulner describes her approach as focusing on the next hill, not the whole mountain—identical to Paul Graham’s advice in “Startups = Growth.” She discovered the same principle independently.
  • Enterprise sales on steroids: Healthcare adds HIPAA, life-or-death stakes, and regulatory compliance to the already-complex enterprise sales model. This makes bottom-up, product-led growth nearly impossible—every sale is a CIO-level decision requiring a business associate agreement.
  • Aggressive litigation: Epic aggressively defends its IP, enforces non-competes, and won a landmark Supreme Court case on forced arbitration. It is currently facing an antitrust lawsuit from Particle Health.
  • Founder continuity: 47 years of the same founder-CEO imbuing personality and values into the organization is extraordinarily rare and powerful—comparable only to Buffett at Berkshire Hathaway and Jensen Huang at NVIDIA.

• Bull case

  • Expanding from EMR into adjacent markets: payers (insurance companies), pharma, prior authorization automation, home health—leveraging their dominant position on the provider side.
  • Ambient AI and the potential for AI to absorb administrative bloat, making Epic the AI operating system for healthcare.
  • Pricing power: they likely have room to raise prices but choose not to in order to maintain customer lock-in and loyalty.
  • Regulation as a moat: new interoperability rules (TEFARA, Cures Act) are complex to implement, and Epic has the resources to comply first and best.
  • International expansion, though limited by lower willingness to pay outside the U.S.

• Bear case

  • International markets may not be very profitable due to lower healthcare spending and price sensitivity.
  • The Particle Health antitrust lawsuit (Sherman Act violation) is a real risk if it results in a material adverse ruling.
  • The Cures Act and information-blocking rules make it legal to scrape Epic data via screen scraping, Chrome extensions, and RPA. This could theoretically allow a third-party UI layer to sit on top of Epic and eventually replace its backend—though the hosts consider this near-zero probability.
  • Interoperability could shift the market toward best-of-breed point solutions rather than integrated suites—but Epic’s bundling strategy (offering point solutions for free within the enterprise agreement) counters this.
  • A paradigm shift (AI-native care delivery, value-based care) could theoretically make traditional EMRs obsolete, though it’s hard to imagine Epic not being central to whatever replaces them.

• Quintessence: why did the “build everything yourself” strategy uniquely work here?

  • This is vertical market software, not horizontal. In vertical software, you serve one category of user, and the deeper you go into their operations, the better. Listening to customers and building exactly what they want is the correct strategy—the opposite of horizontal software (where listening to customers can lead to bloated, unfocused products).
  • Healthcare is uniquely suited to the single-vendor approach because the cost of failure is catastrophic: bugs can kill people, data leaks violate HIPAA, and billing errors are federal crimes. The “sticks” (regulatory penalties, patient safety risks) are so severe that customers rationally choose the most reliable integrated system, even if it takes decades to build.
  • Epic is likely the most valuable vertical market software company in the world—a rare case where the vertical (U.S. healthcare, 18% of GDP) is large enough to support a $50–100+ billion company serving only one industry.