The rules investors use to build portfolios were written for a different era. Check sizes, reserve ratios, ownership targets, stage focus — these models were calibrated against a generation of B2B SaaS businesses with predictable cost structures and relatively linear growth curves.
Â
AI-native companies are breaking those assumptions. On one side you have dramatically lower marginal costs and faster revenue ramp, on the other you have massive infrastructure dependencies and winner-take-most dynamics.
Â
In this session, venture investor and Stanford lecturer, Trevor Loy, examines what this shift actually means for portfolio construction. How many bets should you make? How much should you reserve? When does concentration beat diversification? And does your fund size still fit your strategy?
This session is designed for active venture investors, angels, and emerging fund managers who are already making or evaluating investments and want to stress-test their portfolio construction assumptions against AI market dynamics. Dive into:Â
- Why the standard SaaS portfolio model doesn't translate to AI-native companies
- How AI's cost structure changes the risk/return math
- What winner-take-most dynamics mean for how many bets to make
- How to rethink reserve ratios when revenue ramp timelines compress
- When to concentrate versus diversify in an AI-native portfolio
- How to audit whether your fund size still fits your strategy
Trevor Loy is an Adjunct Lecturer in the Department of Management Science & Engineering at Stanford University, where he teaches entrepreneurial management, finance, and venture capital. He is affiliated with the Stanford Technology Ventures Program (STVP), the entrepreneurship education center at Stanford's School of Engineering that he helped create more than 25 years ago. He also teaches in executive education programs for Stanford's Graduate School of Business and Center for Professional Development, as well as a variety of programs for other global universities, governments, sovereign wealth funds, and corporate investors.
Â
Trevor is also the Managing Partner and Founder of Flywheel Ventures, a seed- and early-stage venture capital firm that has backed founders in emerging startup ecosystems for the past two decades. He is also a former director of the National Venture Capital Association (NVCA), the trade association for the U.S. venture capital industry, where he created the first-ever task force for emerging VC fund managers. In addition to Flywheel, Trevor founded Astria Semiconductor d/b/a MicroProbe (acquired by FormFactor) and ParkingNet (failed). He previously held management and/or technical roles at startups and large technology companies, including Gigabeat, Brooktree, Rockwell, Teradyne, and Intel. Trevor holds a BS and MS in Electrical Engineering with a focus on semiconductor physics and an MS in Management Science & Engineering, all from Stanford.
VC Unlocked: Silicon Valley is an educational program offered in collaboration with the Stanford Engineering Center for Global & Online Education. This exclusive, in-person, venture capital education program is tailor-made for emerging leaders eager to shape the future of the venture capital industry.
Experience a transformative 2-week program taught by the best in the industry. Engage with Stanford University faculty and 500 Global leaders in live, interactive sessions. This immersive approach ensures a dynamic learning environment that goes beyond theory, providing actionable strategies for real-world success.
Location: Stanford University
Dates:Â August 24 - September 4, 2026