May 16, 2026 · 1h 8m · 20vc
The Five Year Desert to Product Market Fit & a $5.3BN Valuation with Shiv Rao, Founder @ Abridge · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this in-depth interview, Abridge founder and CEO Shiv Rao discusses the company's journey to a $5.3 billion valuation, detailing the strategic pivots, fundraising challenges, technical choices, and wartime leadership principles that allowed them to scale a leading healthcare AI assistant.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 21.8% of the talking time here. How this is scored →
Harry as informed peer 4.0 Guest teaching 3.0 Guest disagreement 1.4 Harry pushing back 2.8
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:00:53 Premise rejection on San Francisco as a startup location
Shiv flatly rejects Harry's assertion that San Francisco is the worst place to start an AI company, arguing that the concentration of early-stage talent and osmotic energy in SF is unmatched.
Hardest push from Harry ▶ 41:28 Refusing framing on AI replacing doctorsHarry pushes back forcefully when Shiv describes AI suggesting diagnoses, arguing that if AI can provide diagnostic prompts, it should simply replace the doctor entirely.
Biggest teaching moment ▶ 11:26 Healthcare market segmentation masterclassShiv educates Harry on the structural nuance of the $5.3T US healthcare industry, explaining why staying downmarket is a trap and detailing how practicing physicians are concentrated across specific enterprise care systems.
Harry holds his own ▶ 38:19 Moral framing counterargument on healthcare dataHarry uses his domain expertise to challenge Shiv's refusal to monetize data, arguing that withholding patient data from frontier models presents an ethical dilemma by slowing down global healthcare advancement.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Meeting and Early Impressions | 2 | 1 | 0 | 0 | Harry opens with warm praise and an engaging binary question on founder motivation before asking about market timing lessons. Shiv responds collaboratively, reflecting on maintaining thesis resilience during Abridge's early wilderness period. | |
| Dying on the Hill: Unwavering Commitment to the Core Thesis | 3 | 2 | 1 | 2 | Harry challenges whether Shiv would have died on his thesis if the market hadn't caught up, pressing on early fundraising difficulties. Shiv articulates how he pivoted tactics while holding firm to his core thesis on healthcare conversations. | |
| The Differentiator of Taste in the Era of AI | 2 | 2 | 0 | 1 | Harry brings up taste as a human differentiator in AI, asking Shiv to reflect on company values and strong opinions loosely held. Shiv explains living at the edge of culture and updating priors as company scope expands. | |
| Ambition and the Impossibility of Founder Satisfaction | 4 | 4 | 1 | 2 | Harry uses his experience as an investor in Legora to frame questions about market adoption and GTM traps in enterprise healthcare. Shiv educates Harry on the fragmented $5.3T healthcare market structure and timing the move upmarket. | |
| The Timing of Being First and the Three Eras of AI-Native Companies | 5 | 4 | 2 | 3 | Harry asks how vertical AI companies survive foundation models building healthcare apps. Shiv responds forcefully that fighting foundation models is a losing battle and outlines Abridge's deep workflow integration and mid/post-training strategy. | |
| Proprietary Models vs. Riding the Frontier Model Wave | 5 | 4 | 2 | 4 | Harry pushes back on the decision to build in-house models given historical precedents like Cursor and fast-moving foundation models. Shiv explains that in-house models are necessary to optimize latency, cost, and workflow speed for doctors. | |
| Optimizing for Milliseconds and User Experience | 4 | 3 | 1 | 3 | Harry challenges Shiv on when cost consciousness becomes essential for funded startups, bringing up IPO timing. Shiv recounts a conversation with Henry Kravis to explain why serving the mission in private markets takes precedence over premature optimization. | |
| Data Cleanliness, Deployment Complexities, and the FDE Process | 4 | 4 | 1 | 2 | Harry questions whether an FDE model is mandatory for selling AI into enterprise healthcare. Shiv details how Abridge picked a universal spoken wedge (doctor-patient notes) to scale without needing massive forward deployment teams. | |
| Moving Closer to the Flow of Money and Threading the Stakeholder Needle | 5 | 4 | 2 | 4 | Harry pushes Shiv with feedback from an anonymous investor regarding moving closer to money flows, and asks how they compete with entrenched players like Epic. Shiv reframes Epic as a partner EMR while Abridge operates as the intelligence layer. | |
| Reflecting on Missteps: The Pivot from Patient-Facing to Enterprise | 4 | 3 | 1 | 3 | Harry directly probes whether building a consumer-facing app during the pandemic was a strategic mistake. Shiv acknowledges spending too many cycles on DTC monetization before shifting focus to enterprise healthcare revenue. | |
| Pricing Models and Competing with Microsoft Nuance | 4 | 3 | 1 | 3 | Harry asks how Abridge handles pricing mechanics against CFO expectations and bundling threats from Microsoft Nuance. Shiv describes counter-positioning against Microsoft's complex pricing structures and building category dominance. | |
| Focus, Counter-Positioning, and Competing with Giants | 5 | 4 | 3 | 5 | Harry presents a strong counterargument, asking if withholding healthcare data from frontier models like OpenAI is actually immoral given potential public health benefits. Shiv counters that trust and explicit health-system alignment are prerequisite to long-term impact. | |
| Talent Acquisition: Prioritizing People Over Early Model Access | 5 | 5 | 3 | 5 | Harry provocatively suggests that if AI is giving doctors prompts and diagnoses, AI should just replace them. Shiv firmly pushes back, explaining the sheer volume of healthcare tasks and Jevons paradox in medical demand. | |
| Inference Demand and Token Consumption Growth | 4 | 3 | 0 | 2 | Harry brings up Goldman Sachs token consumption research and explores executive hiring challenges. Shiv shares an anecdote about an late-night call from Jensen Huang on finding joy in all aspects of leadership. | |
| CEO Strengths, Weaknesses, and the Reality of HR | 4 | 2 | 2 | 3 | Harry challenges remote work and HR functions, citing pushback from his own tweets. Shiv defends a balanced 3-day in-office policy and delegating culture-carrying responsibilities to trusted executives as the company grows to 450 people. | |
| Defining Abridge's Culture: The Hiring Fire | 4 | 2 | 2 | 3 | Harry shares his 'Titanic rule' on response times and asks if Shiv is a wartime CEO. Shiv insists all high-growth startup CEOs must operate in wartime mode in fast-evolving markets. | |
| The Most Challenging Chapters of Growth | 4 | 3 | 1 | 3 | Harry asks how Shiv prevents loose spending after raising a massive $300M round. Shiv explains hiring principled finance leaders who maintain discipline while spending aggressively where it creates winning leverage. | |
| Navigating the Toughest Funding Round: The Series A-1 | 5 | 3 | 2 | 4 | Harry candidly reacts to Abridge having a 'Series A-1' round and probes the power transition between VCs and founders when a company becomes a fund-returner. Shiv explains navigating pre-inflection fundraising with supportive insiders. | |
| CEO Parenting and Personal Non-Negotiables | 3 | 2 | 0 | 1 | Harry and Shiv trade personal non-negotiables regarding family time while running high-demand careers. Shiv discusses accepting the explicit trade-offs of founder life while prioritizing weekly visits with his parents in Pittsburgh. | |
| The Case for San Francisco and its AI Hub Advantage | 5 | 4 | 4 | 5 | Harry explicitly challenges Shiv by stating San Francisco is the worst place to start a company due to intense talent competition. Shiv firmly rejects this premise, arguing that SF's sheer concentration of AI talent and osmotic learning make it an unmatched hub. | |
| Quick-Fire Round: Costco's Culture and the Reimagined Healthcare Business Model | 3 | 2 | 1 | 1 | In the quick-fire round, Harry asks rapid questions on organizational flattening, favorite investors, and underappreciated CEOs. Shiv highlights Costco's CEO and explains how AI will fundamentally rewrite healthcare business models over the next 3-5 years. |
Statements from this episode (60)
Healthcare system is unprepared for an impending surge in needs
“We're in for a tsunami of healthcare needs and the system is not prepared to deliver it.”
Great CEOs are driven by fear of losing or thrill of winning
“Great CEOs are motivated by one of two things I found. Either the fear of losing, or the thrill of winning.”
Clinical conversations can power new healthcare AI infrastructure
“It was this thesis that this idea, the secret hiding in plain sight, that this data set, this, like these conversations, it's the most human signal and it really powers healthcare. And you can build a whole new set of rails on top of this signal.”
Abridge raised a $5M seed round at a $15M pre-money valuation
“Our series seed was like, we raised five million dollars on a pre of 15.”
Doctors refuse to join hospital networks that do not provide Abridge
“There are doctors right now who we've heard who won't sign up to join a healthcare system, a hospital, if they don't have a bridge.”
800,000 US doctors actively practice, mostly in large delivery systems
“There's about a million doctors in the country, maybe 800,000 of them are actually practicing. The vast majority of them are concentrated in large care delivery systems.”
Healthcare founders trap themselves by staying down-market too long
“The trap that a lot of, I think healthcare founders fall into is that they stay down market. They don't figure out, they don't like Time their Yolo shot to go up market at the right moment.”
JAMA study showed 30% of US nurses planned to leave profession
“There was a study about nurses recently that was published in JAMA that suggested that 30% of nurses in the country Didn't want to be nurses in the next 12 months.”
Abridge was founded three months after the Transformer paper
“We started three months after the Transformer paper.”
Vertical AI startups cannot win by fighting foundation model labs
“If you are fighting against them, you've already lost. If you haven't figured out how you're going to win with them, how you're gonna not just coexist, but actually find ways to collaborate potentially, and if the tailwinds that they create are not yours to le…”
US healthcare sector is $5.3T and 18-19% of GDP
“Healthcare, 5.3 trillion dollar, you know, market. It's about 18, 19% of US GDP.”
Highest-upside vertical AI companies control full tech stack
“I think the vertical AI companies that have the most leverage that have the most upside are the ones that can reach farther down into the stack and own their destiny in a different differentiated way that can kind of control their P and L.”
Forty percent of Abridge's model outputs come from in-house models
“About 40% of our model outputs inside our product are generated by in-house models.”
Use frontier AI models when building features that require endless improvement
“If the market cares about you always improving on something that you don't think you'll ever be perfect on, and healthcare has a lot of use cases like this, it makes more sense to ride the frontier wave.”
Frontier models could not meet latency needs for key healthcare workflows
“But that kind of experience you need insane performance and latency And that's something that we weren't able to achieve in an easy, elegant way with Frontier model. So for certain use cases, there was really no choice.”
Abridge will remain private as long as private capital supports its mission
“I have a mission, and to serve that mission I will do what it takes, and if I can do it in the private markets, I'll raise money there. If I need to go public to do that, then I will go public.”
Data cleanliness will be AI's biggest enterprise barrier
“I think one of the biggest barriers to AI progression within large enterprise in the next few years will be data cleanliness, data structure within large enterprises.”
Healthcare AI startups do not need forward-deployed engineers
“No, you don't. Not in our case. You're going to get to that zero to one. You're going to get to that first product in all the ways, like people in, you know, in the health systems, in the clinics and listening to the doctors and the nurses and figuring it out.…”
Doctors are compensated for documented care, not care delivered
“So when you think about it in, in the United States, but really this is like true in many parts of the world, clinicians, doctors, they're not compensated for the care that they deliver. They're compensated for the care that they documented, that they deliver.”
Winning healthcare enterprise sales requires aligning CMIO, CIO, and CFO
“And so you got to take that responsibility very seriously because the enterprise motion in healthcare is complicated. You have a lot of different stakeholders. You have your end users, but then you have the decision makers and there's a CMIO, chief medical inf…”
Epic is an entrenched EMR product that nobody likes
“They're like the product that no one likes, but it's so entrenched. Like, so entrenched.”
Abridge will never become an electronic medical record system
“Like we don't ever want to be an electronic medical record.”
Epic was the best EMR software during clinical practice
“And I was at a health system that had multiple medical records, and I was always the happiest when I was using Epic compared to any others out there.”
Abridge is expanding back into patient-facing healthcare software
“Actually, we're getting back to the patient side of that slide now, again,”
Consumer-only healthcare startups rely on untrustworthy data-selling models
“The companies that you can build on top of just the consumer and healthcare, Oftentimes there's the kind of companies that, that paint themselves into a corner that can never be trustworthy enough to go enterprise. They're the kind of companies where the busin…”
Microsoft bought Nuance for $20B+ and competed with Abridge
“And the first big competitor for us was Microsoft. They had bought a company called Nuance for 20 some billion some years ago, and they had a product in our category and we still compete with them in the marketplace.”
Abridge rarely encounters competitors Nabla or Candidate in sales deals anymore
“We don't see either of those companies much anymore.”
Microsoft obfuscates pricing and overcomplicates enterprise software
“Especially like the Microsofts of the world just obfuscate things so much and make things so complicated.”
Make a competitor's side project your primary business focus
“I always want my main course to be someone else's dessert because like I'm going to care much more than they do.”
Top research talent is far more valuable than early model access
“B. B over A. Yeah, no question about it. No question about it, because ultimately it's always about the people, and if you've got incredible people, you're going to build your own models, your own primitives, and yes, the frontier models you're going to be abl…”
Internal medicine journal found doctors need 30 hours daily for tasks
“There was an American Journal of General Internal Medicine article that was published a few years ago that suggested that doctors need 30 hours a day to get all of their work done.”
AI will progressively automate low-stakes primary care workflows
“It's those types of workflows that are, will, you know, progressively get automated till we get to some approximation of maybe the most boring variant of primary care, you know, that nobody wants to practice.”
Declining AI inference costs will drive exponentially higher token usage
“I don't know about the specific number, but I think directionally, I think so. I think that you just can't get enough of this technology once you start to use it. And even as inference costs come down, you just end up using more and more and more of it.”
Founder mode means targeted tours of duty, not micromanagement
“Founder mode is just about, to me, it's about tours of duty depending on what's on fire inside the company or like what's the most important thing and what you're best at. You just go crush. You just have to go you know, kill whatever that challenge is. And it…”
NVIDIA CEO Jensen Huang maintains extremely fast response times to founders
“And so, and Jensen, NVIDIA is one of our investors, and Jensen told me this, he once called me at midnight, and I was like, on his way home, cold call, and I knew it was him and I picked it up, and it was awesome... He was calling me back Because of an email I…”
Work-from-home Fridays lead to lower productivity and are ineffective
“If you want to work from home on Friday, that's okay, but just accept that bluntly you're not going to work as hard, you're going to go to Pilates at 10, and work from home Fridays is bullshit, four days a week.”
Rapidly scaling companies benefit significantly from in-person work
“Things are moving so quickly that the more time that you have in person, the better. There's no question about it.”
Shiv Rao stopped interviewing every candidate around 100 employees
“Oh, it's been a long time. It's been a long time. I. A hundred? Even, probably even a little bit before that, honestly. It was around then.”
20VC enforces a three-hour WhatsApp response rule to drive urgency
“It's three hours long. Everyone should be able to go to a cinema with their families. But if you were asked a question on what's apps like, You need to respond within three fucking hours. It drives efficiency and urgency.”
An Abridge investor required Shiv Rao to undergo a psychological assessment
“We were talking about psychological assessments, I think, before we started, and how some VCs will do that, and I had one before we took money from one of our investors”
Harry Stebbings exits investments the minute founders step down as CEO
“The minute the founder transitions out of CEO role, I'm out.”
Operating without pressure is the hardest stage of startup growth
“You know, honestly, like the chapter of company that, that sucks the most is like the one where you don't feel pressure.”
Massive capital rounds are a brute instrument to win markets
“You raise this capital because it's a brute instrument now that is going to allow you to win. And so against certain things, you just don't want to even debate. And if you are, you're doing it wrong.”
Abridge's hardest funding round was its Series A-1
“The hardest round was a series A-one. You know, venture capital, you know, you obviously know it was a weak round. We were in a weak spot because there's like a digit at the end of it.”
Abridge attended approximately three investor LP meetings in 2023
“Yeah, I think in 2023, we did like three LP meetings. That's when I knew things were going well.”
Anyone who claims CEOs can have everything is lying
“There's trade-offs and there are sacrifices and the folks who say you can have everything are lying and you just have to be eyes wide open on what you're giving up.”
USV backed Abridge seed partly due to Duolingo precedent
“Also Union Square Ventures funded, and I'm sure that had something to do with USV being more comfortable with me, because I certainly didn't have a track record.”
San Francisco is the worst place to start a company
“I think San Francisco is the worst place to start a company because it is so impossible to acquire and retain talent”
Being in San Francisco is super important for AI startups
“As we were inflecting, we just needed to be, you know, in that moment, and the amount of ideas that you just osmotically absorb by being there, the kind of talent that you have access to, the kind of mentorship that you have access to, I think it's super impor…”
Abridge can operate as a flat company using AI tools
“And anymore, I'm in this very idealistic mode that we can be a very, very flat company in the near future. Just like less managers, like less, like a lot more ICs, super ICs taking on that much more, you know, everybody Everybody kind of with these tools able …”
Top VC kingmaking works via a psychological placebo effect on founders
“It's less the brand, it's more the placebo effect. It's more the psychology. It's when you feel like the more you can feel inevitable, the more you will be.”
Elad Gil is the single investor Shiv Rao would choose today
“If I start a new company right now, I would call Elad. Elad Gill. We'd go big, like single GP, you know, like all of the wisdom in his head all that I've learned from him, his ability to go every stage and be incredibly valuable.”
Corporate development requires binary focus, not split-role employees
“How you need to have people just fully focused. It needs to be their entire job. You can't split people's, you know, brains on it. If you're gonna do it, you gotta be binary.”
US healthcare stakeholders are fundamentally misaligned with patients
“And that's what healthcare is in the United States. It's providers, health systems, pointing guns and getting pointed at by insurance companies. And then there's also life sciences companies. And those three stakeholders in some way, shape or form are misalign…”