20VC: Lessons from Jensen Huang on "Founder…

20VC with Harry Stebbings | Guest: Shiv Rao, Founder & CEO of Abridge

1. Key Themes

The "True North" Thesis as Survival Mechanism During Long Market Timing Gaps

Abridge was founded in 2018 but only became "hot" around 2023 — a five-year wilderness walk. Shiv frames this not as patience but as conviction-backed resilience: you don't need to know when the market opens, just that it will.

"You live every day in this anticipation that the sky is going to open up because you have this thesis about the market. And you might not necessarily know exactly when it's going to come true, but you know it's going to come true. And when you feel that in your bones, then I think you can draw a lot of resilience." 00:05:42

"We were pre-selling the market and preparing the market for what we could bring. And then in 2023, when the sky opened, we ran right in." 00:15:11

Vertical AI's Moat Is Data Depth, Workflow Integration, and Regulatory Complexity — Not the Model

Shiv makes a compelling case that the real defensibility in vertical AI is not the underlying model but the layers built on top: proprietary data, post-training from user edits, workflow specificity, compliance architecture, and trust built over years. OpenAI and Anthropic deploying forward engineers to enterprises validates this, not threatens it.

"It's not easy to go into one of these enterprises and figure stuff out, to get access to the data, to clean the data, to organize the data, to be able to integrate into specific workflows, to do it in a compliant way... There's just so much to it, and it takes a lot of effort to build a machine that's scalable." 00:24:21

"People, I think, in 2023 thought that was like last mile. That's actually most of it is that part. The model piece is much less." 00:17:52

Healthcare AI's Real Prize Is Business Model Transformation, Not Feature Deployment

The most ambitious insight in the episode is Shiv's framing of what AI can actually do in healthcare: not automate tasks at the margins, but fundamentally realign the broken incentive structures between providers, insurers, and patients. He sees AI as the mechanism to shift from sick care to preventive care.

"The opportunity that we've got right now with AI is not to deploy the latest model and make science fiction happen in a clinic. It's to change the business model. And this technology is going to do it." 01:00:44

"We're going to see new models that actually align folks and incent prevention. Prevention meaning like care and not like healthcare, not sick care." 01:01:10

2. Contrarian Perspectives

Forward Deployed Engineers from Foundation Model Companies Are a Bullish Signal for Vertical AI, Not a Threat

Most founders fear that OpenAI and Anthropic building enterprise deployment capabilities signals competitive encroachment into vertical AI. Shiv reads it as the opposite — it confirms the enormous complexity and opportunity that only specialists can capture.

"When you saw the announcements from OpenAI and Anthropic about forward deployed engineers and these partnerships with big private equity groups to go help their portfolios, if that wasn't a sign that there is absolutely an incredible opportunity for the foreseeable future for vertical AI, I don't know what is." 00:23:58

"Start Down Market, Swim Upstream" Is Dangerous Advice for Healthcare Founders

The conventional VC wisdom of starting small, finding PMF, then scaling up-market is actually a trap in healthcare because the market is highly segmented and the real money is concentrated in large integrated delivery networks — not small practices.

"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." 00:14:13

"Healthcare specifically in the United States, it's not one $5.3 trillion market. It's a bunch of different markets. And depending on who you're trying to serve, you have to be really careful about how you segment." 00:13:16

You Don't Need to Go Public — and You Shouldn't Default to Wanting To

Against the conventional startup narrative of IPO as validation and goal, Shiv (prompted by Henry Kravis) argues that going public should only be a funding mechanism triggered by mission need, not an aspiration.

"One of his questions was, do you want to go public? And I remember starting to like clumsily navigate an answer on the fly. And he interrupted me and said, no, you don't... The right answer was, no, I don't need to. I have a mission. And to serve that mission, I will do what it takes." 00:23:04

Taste Is a Learnable, Cultivable Company Value — Not an Innate Gift

Against the popular notion that taste is a rare, unteachable trait belonging only to exceptional individuals, Shiv codified it as a company value with a concrete operational definition: expose yourself to the edges of culture, ML research, and UX to develop the pattern recognition needed.

"You have to taste good things to have good taste... We should be reading the latest archive papers about some new type of machine learning model... We should be thinking about the latest UI, UX patterns... We should be just sort of like living at the edge of culture if we also want to create it." 00:09:35 / 00:10:26

Being Early Is Recoverable — But Only If You Evolve Your AI Stack Generation by Generation

Shiv reframes the "being early is being wrong" problem: early-stage AI companies can survive and win, but only if they aggressively upgrade their product to each new model generation (transformer → LLM → agent). Failure to do so — not the earliness itself — kills companies.

"Anymore, depending on your vintage, you have to make sure that you become the latest variant as fast as you possibly can. And that might mean that your product evolves pretty significantly. It might also mean that the way you organize your company and you operate your company evolves pretty significantly." 00:15:40

3. Companies Identified

Abridge AI-powered clinical documentation and workflow company serving major health systems. Started 2018, hit breakout in 2023, now valued at $5.3B after $300M Series E. Captures doctor-patient conversations and automates notes, orders, billing, and more.

"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 Abridge. It's one of those moments where all of a sudden, almost overnight, over the course of like three to four years, something just became table stakes." 00:12:48

Union Square Ventures (USV) Early-stage VC firm known for internet infrastructure and network effects investing. Led Abridge's seed and Series A. Shiv identified them as the right partner not through traditional channels but by studying their public intellectual patterns — specifically their cross-genre music taste on Twitter — as a proxy for abstract pattern-matching capability across domains.

"They had this ritual where they would talk about music. And I was always really impressed that they seemed to pattern match across totally different genres... And to me, that's kind of how I enjoy music as well. And that if they could abstract at that level, my thesis was they could be the right tech investor to think about healthcare in a new, different way." 00:08:06

NVIDIA Chip and AI infrastructure company. NVIDIA is an investor in Abridge. Jensen Huang personally engages with portfolio companies, including late-night calls, with extremely fast response SLAs.

"Jensen told me this. He once called me at midnight... He responds that fast. It's really wild." 00:43:04

Duolingo (referenced via Luis von Ahn) Language learning app founded by Luis von Ahn, also a Pittsburgh native and USV-backed company. Referenced as a model for mission-driven founder commitment articulated at the IPO stage.

"I remember in his S1... he has that one line in the S1 about how the only thing you need to know is that I'm dedicating my life to this." 00:53:42

Costco Membership-based retail giant. Cited as an underappreciated company for its long-tenured CEO, operational culture continuity, and growth despite being unglamorous.

"He carried the torch of culture. And he's managed to continue to grow a business that obviously like he inherited or he was already awesome. But they've only had what, three or four CEOs. That's like one of those businesses that you don't read about enough." 00:59:24

4. People Identified

Andy Weissman Partner at Union Square Ventures. Led Abridge's seed round. Described as a foundational partner who invested pre-traction and pre-track-record. Known for cross-domain pattern recognition and low ego ("coffee's coffee, man").

"People betting, you know, on me early, you know, when I had nothing really to show for it." 00:56:56 "He was just like, coffee's coffee, man. I'll have any coffee... he'll say, I like all the sounds." 00:08:52

Elad Gil Solo GP and prolific angel/growth-stage investor. Shiv names him as the one investor he'd take with him if starting a new company. Valued for multi-stage involvement, deep operating wisdom, and specific expertise in corporate development from his days at Twitter.

"I would call Elad Gil. We'd go big, like single GP, you know, like all of the wisdom in his head... his ability to go every stage and be incredibly valuable." 00:58:05 "His lessons learned from his days at Twitter and like, how do you do that effectively and how to think about CorpDev and how aggressive you should really be... If you're going to do it, you got to be binary." 00:58:31

Jensen Huang Founder and CEO of NVIDIA. Investor in Abridge. Referenced for his extreme responsiveness and the lesson he imparted about falling in love with whatever the job demands — including the unglamorous parts like travel.

"Your job is to fall in love with whatever the job is. That is something you can do. You can convince yourself. You have to find a way to bend your DNA." 00:43:33

Henry Kravis Co-founder of KKR, investor in Abridge. Provided a pivotal reframe on the IPO question — pushing Shiv to anchor on mission over liquidity event as the guiding principle.

"He interrupted me and said, no, you don't. Like, you don't need to. Why are you even thinking about it?" 00:23:04

Joshua Wolfe (referenced as "Joshua from Lux") Co-founder of Lux Capital. Referenced for his "chips on shoulders makes chips in pockets" aphorism — used by Shiv as a hiring heuristic for identifying people with deep-seated drive.

"Chips on shoulders makes chips in pockets. I think that's a hard thing to kind of be able to assess in a person. But there are just certain people who have insane slope. And that slope is coming from something very, very deep and core to who they are." 00:47:24

Ali Ghodsi CEO of Databricks. Named by Shiv as an idol for strategic market navigation — specifically praised for how he "plays chess" in a highly competitive market.

"I have people like Ali Godzi, who I think is an absolute legend in the way he plays chess and navigates his market." 00:59:24

5. Operating Insights

Counter-Positioning Against Large Incumbents Is a Deliberate Architectural Choice, Not Just Messaging

Rather than competing head-on with Microsoft (which acquired Nuance for $20B+), Abridge built from the conversation signal up into revenue cycle — a direction Microsoft can't follow without cannibalizing its own EHR/enterprise positioning. This is Hamilton Helmer's "Counter-Positioning" power applied tactically.

"Big companies oftentimes have a lot of soft spots that you can counter position against where they just have to watch you... When you're building in a way where the competitor couldn't build because it would impact their current business, it would hurt them, it really pushes them to think hard and probably prohibits them from directly competing against you." 00:34:49 / 00:35:19

The "Earn the Right" Data Principle as Both Ethics and Competitive Moat

Before building any feature that uses patient data, Abridge goes back to health system partners, discloses the roadmap in advance, and gets contractual blessing. This isn't just compliance — it's a trust-building flywheel that lets them move into sensitive areas competitors cannot access.

"We always go back to our health system partners. And oftentimes, they know about our roadmap in advance. And they've already blessed it. And we've already papered this into the contract that we can build XYZ with the insights that we get." 00:36:42

In-House Models for Binary/Deterministic Tasks; Frontier Models for Continuous Improvement Tasks

Abridge has developed a principled framework for when to build in-house vs. ride frontier models: if a task has a clear "done" state (discrete data entry into a field), own it with a fine-tuned model for latency and cost. If a task requires perpetual improvement and the market values that trajectory, ride the frontier wave.

"Those sorts of tasks, being able to just crush it with an in-house model makes total sense... But then there are certain challenges in your product suite that you're never going to be perfect on. But every week, every month, you want to be able to look back and say you're less imperfect than you were before... it makes more sense to ride the frontier wave." 00:19:31 / 00:20:00

The Three-Stakeholder Threading Model for Enterprise Healthcare Sales

To win in large health systems, you must simultaneously resonate with three distinct buyer personas who have completely different lenses: the CMIO (clinical quality), the CIO (technical integration), and the CFO (financial return). Failure to thread all three loses the deal.

"You have a lot of different stakeholders... There's a CMIO, chief medical information officer. There is a CIO, a chief information officer, and then there is a CFO. And they all have a different lens. And being able to thread the needle, being able to resonate with all three is really how you win the day." 00:27:15

6. Overlooked Insights

USV's Music Taste Was the Actual Investment Thesis Signal — and It Returned ~$1B

This is mentioned almost in passing as a quirky anecdote, but it reveals something profound about how the best founder-investor matches are made: Shiv didn't pitch USV because they were prestigious. He stalked their intellectual pattern-matching behavior on Twitter for years, identified that their ability to find signal across wildly different musical genres mapped to the cognitive flexibility needed to see healthcare differently than traditional health tech investors. The title of the episode even references "USV liking music made them $1BN on an investment." This is a fully replicable diligence methodology for founders choosing investors — study how a VC thinks across unrelated domains as a proxy for their ability to pattern-match in your domain.

"They had this ritual where they would talk about music. And I was always really impressed that they seemed to pattern match across totally different genres. The country western on Monday and then indie hip hop on Wednesday and then like Swedish death metal on Friday... And to me, that's kind of how I enjoy music as well. And that if they could abstract at that level, my thesis was they could be the right tech investor to think about healthcare in a new, different way." 00:08:06

The Consumer Healthcare App Was Both a Strategic Mistake and the Future — Simultaneously

Shiv mentions almost in passing that building the direct-to-consumer patient app was "a strategic mistake" because the business models would require selling sensitive data and painted them into a corner. But then he immediately says "we'll get back to that... we're getting back to the patient side of that slide now again." This is a hugely significant signal: Abridge at $5.3B valuation is now re-entering the patient side of the market — the second half of that original seed pitch slide. That represents an entirely new product surface area and potential revenue stream that has not been publicly discussed, and it comes armed now with enterprise trust, data scale, and model capabilities that make the consumer play viable in a way it wasn't in 2020.

"It was a strategic mistake then to build that patient side... But it was those types of workflows... we'll get back to that. Actually, we're getting back to the patient side of that slide now again." 00:30:30 / 00:30:56