She Knows the 250 People Building AI. Here's What They Actually Believe.

Watch on YouTube ↗  |  September 01, 2026 at 12:37  |  1:10:53  |  ILTB Podcast
Speakers
Sarah Guo — Founder, Conviction

Summary

Sarah Guo discusses how she navigates the AI boom as an early-stage investor, emphasizing frontier researchers, compute and energy constraints, open-source models, robotics, semiconductors, and AI-enabled biology. She explains her framework for building conviction without backtesting and where Conviction is investing across AI infrastructure and applications.

  • Sarah highlights the extreme pace and competitiveness of frontier AI labs and the recursive self-improvement debate.
  • She sees compute and energy supply chains, including nuclear and data center buildout, as key constraints and investment areas.
  • She argues open-source models will democratize AI and that restricting them would disadvantage US businesses.
  • She is bullish on robotics reaching homes soon, with Sunday Robotics as an example.
  • She believes AI can create large value in biology, citing Chai Discovery's pharma partnerships.
  • She expects AI agents and applications to produce broad productivity gains, not just in software engineering.
Ideas
Sarah Guo Founder, Conviction 8:49
Legal AI like Harvey leverages structured language.
Law is structured language, making it a strong fit for large language models: legal work requires reading many documents, retrieval, and text generation, with abundant precedent text. Sarah backed Harvey because its founders believed AI would transform law practice, from answering landlord-tenant questions to doing 85% of M&A work.
Harvey Legal AI HIGH
Sarah Guo Founder, Conviction 23:00
Semi-humanoid home robots are coming soon.
Sarah backed Sunday Robotics because Tony Zhao and Changyi have contributed many key robotics AI ideas and could solve robot generalization through clever, cheap real-world data collection. She believes general semi-humanoid robots will enter homes in beta at the end of this year or next, moving robotics from a question of if to when.
Sunday Robotics ROBO HIGH
Sarah Guo Founder, Conviction 40:06
Open-source AI models enable broad adoption.
Sarah believes open-source AI models are already competitive and unstoppable; they democratize capability and are preferred where frontier APIs are too expensive, sensitive, or slow. Businesses want open-source models to control their own destiny and economics, and restricting US use would only handicap law-abiding American businesses while doing little to adversaries.
Open-source AI models HIGH
Sarah Guo Founder, Conviction 49:22
AI compute independence requires infrastructure buildout.
Sarah frames AI compute as requiring energy independence: data centers need abundant power, cooling, and supply diversity. She is investing in nuclear energy, alternative chip architectures, and the data center labor gap, while also looking at data center builders and solar/battery installers because capacity buildout is the binding constraint.
URA Data center builders Alternative chip architectures Solar and battery installers HIGH
Sarah Guo Founder, Conviction 52:18
Semiconductor startups now have changed risk profile.
Sarah says venture investing in semiconductor companies used to be a terrible business, but the market has changed because there is now consolidated at-scale demand for accelerators and supply-chain independence from big buyers that want alternatives to relying on TSMC. That changes the risk equation and makes semiconductor startups and venture-backed semi companies more attractive.
SMH AI accelerators MED
Sarah Guo Founder, Conviction 54:06
AI biology models create enormous value.
Sarah has moved strongly to the view that AI models can create and capture enormous value in biology. Chai Discovery, her first check in the space, works with top-10 pharma companies to accelerate R&D, and she expects AI to produce a massive acceleration in cures, with young companies capturing part of the value creation rather than only traditional biotech asset-centric models.
AI biology Chai Discovery HIGH
Sarah Guo Founder, Conviction 64:19
AI agents will diffuse across every domain.
Sarah expects Jevons paradox in practice: AI agents and products will do more mundane work across all domains, analogous to the transformation already seen in software engineering. She sees autonomous AI departments and broad agent adoption increasing productivity and expanding investment opportunities in AI applications.
AI Application Software AI agents MED
Up Next
57:31
Let's Be Wary of an "Entire Sector Rise"! Why You Get Stuck When You Chase Surging Stocks Late | Myeong Min-jun, Kim Gi-eun, Hwang Yu-hyeon [Stock Beginner Rescue Squad]
3PRO TV (삼프로TV) · Sep 01
31:30
Property Play: Coldwell Banker’s CEO on how AI and SpaceX money will hit home prices
CNBC · Sep 01
6:15
'Stay Calm': Legendary investor David Booth on the importance of staying invested
CNBC · Sep 01
40:54
HBM Shortage Goes Until '2028'! The Real Reason Why the AI Bubble Theory is Ridiculed | Kim Jang-yeol, Head of Research Center, Unistory Asset Management [Focus Today's Stock]
3PRO TV (삼프로TV) · Sep 01
4:31
We're going back to a stock pickers' market, says DCLA's Sarat Sethi
CNBC · Sep 01

This ILTB Podcast video, published September 01, 2026, features Sarah Guo discussing Harvey, Legal AI, Sunday Robotics, ROBO, Open-source AI models, URA, Data center builders, Alternative chip architectures, Solar and battery installers, SMH, AI accelerators, AI biology, Chai Discovery, AI Application Software, AI agents. 7 trade ideas extracted by AI with direction and confidence scoring.

Speakers: Sarah Guo  · Tickers: Harvey, Legal AI, Sunday Robotics, ROBO, Open-source AI models, URA, Data center builders, Alternative chip architectures, Solar and battery installers, SMH, AI accelerators, AI biology, Chai Discovery, AI Application Software, AI agents