Epic Systems (MyChart)

2025-04-21 - source - Read full transcript

Ben Gilbert (host)David Rosenthal (host)

Key insights

Epic's entire competitive advantage traces back to a single architectural decision from the 1970s: one core database (Chronicles) that every application, clinical and billing, reads and writes to.

Judy Faulkner designed Chronicles as a patient-record database before EpicCare or Resolute existed. Because every module pulls from the same database rather than being stitched together through acquisitions (as with Cerner's 30+ mergers), Epic can promise implementations that go live on time and on budget, and clinical data never gets lost in translation to billing, which the hosts argue is the single biggest reason hospitals pick Epic over more feature-rich competitors.

vertical-software-moats

Epic won the 2003 Kaiser Permanente RFP over the much larger, publicly-traded Cerner by refusing to split the contract and refusing to give Kaiser equity, betting correctly that its integrated architecture would win a head-to-head technical bake-off.

Cerner offered Kaiser 10% of the company for the deal; Judy Faulkner declined on principle, arguing that giving one customer equity would force the same offer to every large customer. Epic's team modeled Kaiser's actual transaction load in a spreadsheet overnight before the final presentation while Cerner presented only theoretically, and Epic won a deal that roughly doubled its revenue overnight and elevated it to industry gold-standard status.

vertical-software-moats

The 2009 HITECH Act's $36 billion in EHR adoption incentives functioned as a government-funded moat-widening event for Epic rather than a market-opening one, because the payments rewarded reliability over innovation and pulled forward RFP cycles that might otherwise have given future competitors a chance to bid.

Hospitals received $44,000-$64,000 per physician for adopting and 'meaningfully using' an EHR, making price largely irrelevant and reliability the deciding factor. Market EHR penetration jumped from 9% of hospitals in 2009 to 95% by 2014. Because so many multi-decade vendor decisions got compressed into one narrow window, a theoretically superior EHR that might have emerged later in the 2010s never got a comparable greenfield bidding opportunity.

regulatory-tailwinds

Epic captures a surprisingly small slice of the value it creates: about $10 million a year per customer on average, versus hundreds of employees and far larger consultant, implementation, and downtime costs that hospitals pay to run the system.

The hosts compare this to Jeff Bezos's early AWS strategy of wanting it to be 'irrational and irresponsible' not to use the platform: Epic has historically kept price increases to roughly 2% a year (below inflation) and has never changed MyChart's price, deliberately leaving money on the table to remove any incentive for a customer to ever rebid.

healthcare-industry-economics

Epic's 47-year run of never losing a single one of its 600+ customers (aside from one six-month exception) rests on switching costs the hosts argue exceed any other software category, because a failed EHR migration can directly cause patient deaths.

Unlike an ERP or CRM swap, a botched EHR cutover breaks the operational nervous system of a hospital and can produce medication or care errors with life-or-death stakes. This makes hospitals structurally risk-averse in ways that reinforce incumbency far more strongly than typical enterprise software switching costs.

vertical-software-moats

The disastrous Department of Defense and VA EHR contracts, won by Cerner as a subcontractor, show why Epic's customers are relieved Epic lost that bid: the DoD system took nine years to fully deploy and the VA system, bid out in 2017, has an official target of 'as early as 2031.'

The hosts attribute the failure partly to perverse contracting incentives ('you make more money on failed government contracts than successful ones') and partly to compounding bureaucratic layers in defense-prime subcontracting. Cerner's founder Neal Patterson was also diagnosed with cancer and died in 2017 mid-project, and Cerner cycled through multiple CEOs afterward while Epic kept winning major health systems uncontested.

regulatory-tailwinds

Vertical market software inverts the classic horizontal-software wisdom that listening too closely to customers produces bad products, because in a regulated, high-stakes vertical, going as deep as possible into one customer type's exact workflow is the correct strategy.

David frames this as Epic's central strategic insight relative to companies like Microsoft or Salesforce: horizontal platforms must resist over-fitting to any one customer segment to stay broadly usable, while Epic's growth came from being maximally responsive to hospital CIOs, CIOs, and CFOs specifically, since HIPAA and patient-safety requirements make a single, deeply-integrated, standards-compliant system the only workable playbook.

vertical-software-moats

Judy Faulkner built Epic with almost no outside capital ($70,000 of equity plus a $70,000 bank loan as the only financing in the company's history), which the hosts argue was only possible because she personally was a cornered resource as a generational programmer, echoing Bill Gates and Microsoft's own no-primary-capital-raise history.

A mentoring visit to Meditech founder Neil Pappalardo in the 1970s taught Faulkner to run Epic like a software developer would: hardened internal processes, college-only hiring, and an explicit aversion to bringing in traditional 'business guys,' since the era's typical venture playbook was to replace technical founders with professional management once they raised capital.

founder-driven-culture

Epic's control structure, built around Judy Faulkner's Ten Commandments (posted in every bathroom) and a post-death 'purpose trust' requiring the next CEO to be a longtime employee and software developer, is explicitly designed to make the company permanently unsellable and unlistable.

The trust will be jointly administered by her family, five longtime senior Epic managers, and three customer CEOs after her death, with an ironclad rule that the company can never be sold or taken public. The hosts estimate this makes Judy Faulkner likely the wealthiest self-made female entrepreneur in history if the company were ever valued like a comparable public software company (their estimate: roughly $100 billion), a valuation that will never be tested because no liquidity event can occur.

founder-driven-culture

Ambient AI scribe products (from Microsoft/Nuance, Abridge, and Suki) that plug into Epic to auto-transcribe and structure physician-patient encounters represent Epic's clearest near-term AI upside, with one hospital CIO speculating the EHR interface itself could eventually fade into an 'AI operating system.'

Because Epic already functions as the industry's choke point deciding which third-party software reaches hospitals, it can capture value from ambient-AI partnerships through revenue share or equity warrants regardless of which vendor wins, while also having the option to build a competing in-house product later.

ai-disruption-potential

An ongoing antitrust lawsuit from the startup Particle Health, alleging Sherman Act violations over how Epic restricts third-party access to patient data, is the single biggest identified risk to Epic's long-term dominance.

The hosts note Epic has aggressively defended its IP and enforced strict data-sharing terms for good HIPAA-compliance and durability reasons, but that posture is now the basis of a live antitrust case; a material adverse ruling is characterized as a 'company-changing' event, distinct from smaller regulatory cracks like the Cures Act's information-blocking rules that legalized screen-scraping workarounds.

regulatory-tailwinds

Media referenced

Seven Powers - paper - Hamilton Helmer's framework (counter-positioning, scale economies, switching costs, network economies, process power, branding, cornered resource) used in the Power segment to explain Epic's moat.
Death by a Thousand Clicks - article - KFF investigative piece quoted on how the HITECH Act's $36B incentive pot pulled the EHR vendor industry into compliance despite griping about the requirements.
An Epic Dystopia - article - The American Prospect article (acknowledged as one-sided) quoted for doctor complaints about billing-code pressure inside Epic's software.
Health API Guy (Substack, Brendan Keeler) - article - Cited repeatedly as some of the best independent writing on Epic; source for the Philips partnership failure and quotes on Epic's integrated-architecture advantage.
Worldly Partners' Epic write-up (Arvind Navaratnam) - article - Companion research piece linked in show notes; source for the R&D-spend-as-percent-of-opex comparison across Epic, Athenahealth, Oracle, Apple, Amazon, and Google.
Startups = Growth (Paul Graham, 2012 essay) - article - Ben quotes Graham's hill-climbing framing of startup growth as an analogue to Judy Faulkner's own description of her strategy: focus on the next hill, not the whole mountain.
Standard Oil (Acquired episode) - podcast - Referenced as the origin of the Rockefeller Foundation funding that helped professionalize American medical schools and standardize patient records in the 1920s.
Microsoft (Acquired episodes) - podcast - Recurring comparison throughout: Epic's founding era, no-VC capital structure, campus strategy, bundling playbook, and Judy-as-Bill-Gates parallel.
Lockheed (Acquired episode) - podcast - Referenced when explaining how defense-contractor subcontracting economics (paid more for failed projects) shaped the disastrous DoD/VA EHR contracts.
Nintendo (Acquired episodes) - podcast - David's carve-out ties back to prior Nintendo coverage when discussing anticipation for the Switch 2.
IPL / Indian Premier League (Acquired episode) - podcast - Referenced by name when framing the 'mega bull case' speculation about EHRs fading into an AI operating system.

Companies

Epic Systems - Subject of the episode; privately-held Verona, Wisconsin EHR company, $5.7B revenue in 2024, founder-owned, never raised VC, never acquired anyone, never lost a customer in 47 years.
Cerner (now Oracle Health) - Epic's historical main rival, built via 30+ acquisitions; lost the pivotal 2003 Kaiser Permanente deal, won the disastrous DoD and VA EHR contracts, bought by Oracle for $28B in 2021 and now run as a reported 'headwind' to Oracle's growth.
Meditech - The original EMR company (Neal Pappalardo, MIT-trained, out of Massachusetts General Hospital); mentored Judy Faulkner on how to run a company and remains the #3 EMR player behind Epic and Cerner.
Kaiser Permanente - Won by Epic in a 2003 RFP over Cerner despite Cerner offering Kaiser 10% equity; the deal roughly doubled Epic's revenue overnight and established it as the industry's gold standard.
Google, Microsoft, GE - Named as companies that have unsuccessfully tried to acquire Epic over the years; Judy's trust structure ensures the company can never be sold.
Philips - Early-2000s co-development partnership (a Philips-branded version of Epic for radiology) that failed within a year or two and left Epic with lasting distrust of partnerships.
Veeva Systems - Named as the largest public vertical-market healthcare software company, but at roughly $20B market cap, far smaller than the hosts' Epic valuation estimate.
Constellation Software - Named as the other plausible large vertical-market software comparison, since several of its biggest holdings are healthcare IT roll-ups.
Microsoft / Nuance, Abridge, Suki - Ambient-AI scribe companies partnered with Epic that transcribe and structure physician-patient conversations directly into the EHR.
Particle Health - Health-data startup suing Epic; an ongoing antitrust case alleging Sherman Act violations, flagged as the biggest single company-risk in the bear case.
United Healthcare - Used as a revenue-scale comparison: does about $400B a year and $35B in EBITDA, six times Epic's entire revenue, illustrating how little of total healthcare spend Epic actually captures.

Techniques and frameworks

Single unified database architecture (Chronicles) - Every Epic application (EpicCare, Resolute billing, Cadence scheduling, Cosmos, MyChart) reads and writes to one core database Judy Faulkner designed in the 1970s, eliminating the data-sync failures that plague competitors' bolted-together systems.
Seven Powers framework - Used to diagnose Epic's moat as primarily switching costs, scale economies, network economies (Care Everywhere and shared training/hiring pools), branding, and process power, with counter-positioning and cornered resource largely absent during the slow takeoff phase.
Vertical market software strategy: build it all yourself and listen hard to customers - David's central 'quintessence' argument: unlike horizontal software (where over-indexing on customer requests produces bad products), vertical software in a high-stakes, compliance-heavy field rewards going deep into one customer type's exact workflow and never modularizing out core functions.
No sales and marketing, CIO-only enterprise sales - Epic employs about eight inbound-only 'salespeople' pulled from implementation staff; HIPAA business-associate-agreement requirements make bottom-up product-led growth structurally much harder than in typical B2B SaaS.
Standardized implementation as a pricing lever - Epic gives discounts mainly to customers who adopt the standard package rather than customizing, which keeps updates simple to push and interoperability consistent across the install base.

Summary

Ben Gilbert and David Rosenthal trace Epic Systems from its 1970s origins as a side project by University of Wisconsin programmer Judy Faulkner to its current position as the dominant electronic health record vendor in American healthcare, with 607 hospital-system customers, $5.7 billion in 2024 revenue, and a claim to never having lost a customer in 47 years. Faulkner's core insight, formalized while she was working on scheduling and records software for the Wisconsin Medical Center, was to build one single database (Chronicles) that every clinical and billing application would read from and write to, rather than stitching together separate systems the way nearly every competitor eventually did through acquisitions. The company, financed with just $70,000 of equity and a $70,000 bank loan and never touched by venture capital, grew slowly for its first two decades: only 24 customers and $1.5 million in revenue after ten years. Faulkner's early mentorship from Meditech founder Neal Pappalardo shaped an explicitly programmer-run company culture, hardened processes, college-only hiring, no traditional "business guys" that persists essentially unchanged today at Epic's famously whimsical Verona, Wisconsin campus.

The episode's pivotal business story is the 2003 Kaiser Permanente contract, where Epic, then a roughly $50 million company, beat the billion-dollar, publicly traded Cerner for the largest health system in the country by refusing to split the deal between vendors and refusing an equity stake Cerner had offered, then winning a technical bake-off by modeling Kaiser's actual transaction volume overnight rather than presenting theoretically. The deal roughly doubled Epic's revenue and established it as the industry's gold standard, a position later cemented by the 2009 HITECH Act, which distributed $36 billion in incentive payments for EHR adoption. Because the legislation rewarded reliability over innovation and effectively made price irrelevant to hospitals, it disproportionately benefited Epic even though the hosts conclude the company was likely headed toward market dominance anyway; the incentive mainly pulled forward growth that would have happened over a longer horizon, while also closing the door on any newer, theoretically better EHR emerging in the meantime. A useful counterpoint is the Department of Defense and VA EHR contracts, won by Cerner as a subcontractor in 2015 and 2017: the DoD system took nine years to go fully live, and the VA project's current target is "as early as 2031," a fiasco Epic's own customers are reportedly relieved the company didn't win.

The hosts spend significant time on Epic's unusual internal culture: a "software factory" ethos with a 10-commandments list posted in every bathroom (do not go public, do not acquire or be acquired, software must work), an aggressive up-or-out hiring pipeline built on programming-aptitude tests rather than interviews, mandatory clinical "immersion trips" for every employee, and a policy that developers fix their own bugs to keep defect rates near zero given that lives and federal billing compliance are on the line. Faulkner's succession plan transfers her voting control at death into a purpose trust jointly run by family, five longtime senior managers, and three customer CEOs, with an ironclad rule against ever selling or going public, and a requirement that future CEOs be longtime Epic software developers. In the analysis segment, applying Hamilton Helmer's Seven Powers framework, the hosts identify switching costs (the highest of any software category, given life-or-death stakes), scale economies, network economies (via the Care Everywhere data-sharing feature and shared physician training), branding, and process power as Epic's durable advantages, and argue the company is essentially a case study in how vertical-market software inverts the usual horizontal-software wisdom: going deep and listening intensely to one type of customer (hospital CIOs and CFOs) is the correct strategy precisely because HIPAA compliance and patient safety make a single integrated system the only workable architecture.

Looking forward, the hosts frame the bear case around the live Particle Health antitrust lawsuit (a Sherman Act claim they call the single biggest company-specific risk), international market limits, and looser regulatory cracks like the Cures Act's information-blocking rules that now legalize third-party data scraping. The bull case centers on Epic's expansion beyond hospitals into payers and pharma (starting with automating prior authorizations), continued cross-selling of new modules into its existing base, and the potential of ambient AI scribe products (from Microsoft/Nuance, Abridge, and Suki) to reduce physician administrative burden, with one hospital CIO speculating that the EHR interface itself could eventually fade into an "AI operating system" for the entire billing and clinical workflow. The hosts value Epic, conservatively, somewhere in the $50-100 billion range using Cerner's acquisition multiple and comparable public software multiples, which would make Judy Faulkner likely the wealthiest self-made female entrepreneur in history, a valuation that, by design, will never be tested because the trust structure makes a sale or IPO permanently impossible.

Notable Quotes

"The sun was shining, I was dis-attentive, I was just sitting there, and suddenly it all came to me. Here's how you build it. The integrated system. I remember running to the kitchen, grabbing a pad of paper, and just writing code, code, code, code." - Judy Faulkner (quoted by David Rosenthal, on the origin of Chronicles)

"You get what you pay for 100% of the time, despite Epic being 'not cheap.'" - Jeff Gautney, CIO of Rush University System for Health (quoted by Ben Gilbert)

"I think they want it to be an irrational and irresponsible decision for a healthcare system not to use Epic." - David Rosenthal

"This is not uncommon, especially think about ERP, the number of times you've heard oh failed ERP implementation... This thing does happen in the hospital world, too, and Epic has bet the whole company on having a reputation for we don't have failed implementations. And that wins deals." - Ben Gilbert

"Meaningful use in the HITECH Act wildly succeeded at digitization of the industry. It did absolutely nothing on the digital transformation of the industry. We digitized, but we didn't transform in the way that I think people were optimistically hoping for." - unnamed hospital CIO (quoted by David Rosenthal)