Overview of 20VC: Lessons from Jensen Huang on "Founder Mode" with Shiv Rao, Founder & CEO of Abridge
Harry Stebbings sits down with Shiv Rao, the founder and CEO of Abridge, the AI healthcare company now valued at $5.3B after a $300M Series E. The conversation focuses on Abridge’s long road from a 2018 founding and years of uncertainty to category leadership in healthcare AI, as well as Shiv’s views on market timing, founder mode, vertical AI, enterprise GTM, model strategy, and what it takes to build in one of the most complex industries in the world.
Abridge’s Long Path to Product-Market Fit
The “five-year wilderness”
- Abridge was founded in 2018, but the company didn’t hit real momentum until years later.
- Shiv describes the early years as a period of resilience, where the key was to “stay standing” until the market and technology caught up.
- He says he would have died on the hill of the core thesis: healthcare is built on human conversations, and those conversations are the most valuable signal.
Market timing matters, but conviction matters more
- The company’s thesis was right before the market was ready.
- Shiv emphasizes that founders need a true north and should be willing to pivot on product or go-to-market, but not on the core mission.
- His view: being early is hard, but with enough persistence, you can survive until the market opens.
Fundraising, Investors, and Founder-Partner Fit
Early fundraising was relationship-driven
- Abridge raised its seed round in a different era: $5M on a $15M pre-money valuation.
- Shiv credits not just founder-market fit, but founder-partner fit—chemistry with investors matters.
- He specifically recounts how he “stalked” Union Square Ventures for years before meeting them.
Why USV fit
- He admired USV’s ability to pattern-match across seemingly unrelated domains, including music.
- That abstraction ability signaled they might understand healthcare in a nontraditional way.
- His takeaway: the best investors are not just capital providers; they become long-term strategic partners.
Vertical AI, OpenAI, and Why Healthcare Is a Massive Opportunity
Why vertical AI is still wide open
- Shiv sees the recent moves by OpenAI and Anthropic—including forward-deployed engineers and enterprise partnerships—as proof that vertical AI has a huge future.
- He argues that vertical AI is especially powerful in regulated, high-stakes industries with proprietary data and complex workflows.
Healthcare is not one market
- He stresses that U.S. healthcare is actually many markets, not one giant one.
- To win, companies must understand the fragmentation across:
- integrated delivery networks
- academic medical centers
- payer-provider organizations
- clinicians, CIOs, CMIOs, and CFOs
Why Abridge won
- Abridge started with a wedge that existed everywhere: the doctor-patient conversation.
- From that signal, they expanded into:
- notes
- orders
- billing
- revenue-cycle-adjacent workflows
- The product became more than a note-taking tool; doctors began using “Abridge” as a verb.
Product Strategy: Models, Latency, and User Experience
Build your own models when it matters
-
About 40% of Abridge outputs are generated by in-house models, though that mix changes over time.
-
Shiv’s rule: if a task is:
- binary,
- latency-sensitive,
- workflow-critical,
then owning the model can be worth it.
Frontier vs. in-house models
- He is pragmatic, not ideological:
- use frontier models where they’re best,
- use open source or in-house models where performance, cost, or control is better.
- He repeatedly returns to the same principle: the only thing that matters is the end user.
Why latency matters in healthcare
- Doctors have little patience for clunky tools.
- In workflow, milliseconds matter.
- Abridge aims to feel like “good air conditioning”: present, reliable, and invisible when working well.
Enterprise GTM in Healthcare
Why enterprise is the real prize
- Shiv says enterprise healthcare is where the real scale, data, and impact live.
- Selling into healthcare requires earning trust across multiple stakeholders:
- clinicians
- IT
- compliance
- finance
The workflow wedge
- Abridge intentionally focused on workflows that are:
- common across healthcare
- painful
- high leverage
- Examples include documentation, orders, and billing—jobs clinicians hate doing.
Why bundling isn’t the main threat
- Shiv doesn’t see competitors like Microsoft/Nuance or Epic as fatal threats.
- His stance: Abridge is building the intelligence layer, not trying to become an EMR.
- The wedge is the conversation, not the note.
Trust, Data, and the Economics of Healthcare AI
Trust is the moat
- Shiv says healthcare “moves at the speed of trust.”
- Abridge does not sell data because trust is foundational.
- When they use data to improve models or build features, they do so with partner approval—“earn the right.”
Data complexity is a major barrier
- Healthcare data is messy, sensitive, and highly regulated.
- Different types of data must be handled differently:
- behavioral health
- primary care
- internal medicine
- This complexity is why enterprise AI in healthcare is hard—and why there’s so much opportunity.
Moving closer to the flow of money
- Abridge is intentionally getting closer to billing and reimbursement.
- Shiv notes that clinicians are often paid based on what they document, not just what they do.
- That makes documentation AI commercially meaningful, not just operationally useful.
Leadership, Founder Mode, and Company Culture
Jensen Huang’s lesson
- Shiv shares a story about Jensen Huang calling him late at night.
- His takeaway: fall in love with the job and adapt to what the role requires.
Founder mode, interpreted
- For Shiv, founder mode is about tour of duty, not micromanagement.
- The founder should get closest to the most important fires and solve the hardest problems directly.
What he values in executives
- The hardest hires are high-judgment executives who can operate in a fast-changing environment.
- He wants people who can:
- bring patterns and priors,
- but also update them quickly.
Culture at Abridge
- The company is built around:
- mission
- urgency
- taste
- high standards
- One memorable rule: the “Titanic rule” — respond to Slack/WhatsApp within three hours.
Ambition, Family, and Personal Tradeoffs
The reality of being a founder-parent
- Shiv is candid that you cannot “have everything.”
- He travels heavily and acknowledges sacrifices, but says he makes peace with them because the mission matters.
What he won’t miss
- Despite the demands of running Abridge, family is non-negotiable.
- He highlights weekly time with aging parents and his children as a core source of meaning.
The Bigger Future of Healthcare and AI
The next decade
- Shiv believes AI will not just automate tasks—it will help change the business model of healthcare.
- He expects:
- more automation of low-stakes, high-frequency tasks
- better support for clinicians
- more preventive, aligned care models
What Abridge is really trying to do
- Save time for clinicians
- Save money for the system
- Save lives
- Make clinicians feel like superheroes
Key Takeaways
- Core thesis matters more than short-term timing.
- Trust is the moat in healthcare, not just technology.
- Vertical AI wins by owning workflows, not just layering on model outputs.
- In healthcare, the conversation is the wedge.
- Founder mode is about doing the hard thing at the right time, not controlling everything.
- Building a great company in a complex market requires taste, patience, and relentless execution.
Memorable Lines
- “You have to taste good things to have good taste.”
- “The more you can feel inevitable, the more you will be.”
- “If you are fighting against the foundation models, you’ve already lost.”
- “The industry moves at the speed of trust.”
- “Pressure makes diamonds.”