What are the key takeaways from this a16z Podcast episode?

Key insights include: **Early-stage investment thesis:** Speedrun targets pre-incorporation teams of two to three people, often still employed full-time, investing up to $1 million before founders have even formed a legal entity. In the most recent 70-company batch, roughly a dozen teams had to quit jobs and incorporate from scratch before funds could be wired.; **Venture power law at seed stage:** Historically, approximately half of early-stage startups fail outright, another two to three out of ten return modest capital, and only the top decile generates the majority of returns. Speedrun accounts for this by tracking founders across multiple companies, re-funding promising individuals even after a first venture returns minimal capital.; **Location as a regulatory decision:** Two-to-three-person founding teams are highly mobile and actively choose where to incorporate based on cost, investor proximity, and regulatory climate. Founders building AI-native products who anticipate heavy state-level AI regulation factor that burden into their location decision before planting roots or hiring.

What did Andrew Chen and Matt Perault discuss on a16z Podcast?

Andrew Chen, General Partner at a16z, describes the Speedrun program — a 12-week accelerator investing up to $1 million in two-to-three-person startups — and explains how early-stage founders operating from kitchen tables face regulatory burdens designed for large companies, with zero time or resources to engage policymakers. Key topics include: **Early-stage investment thesis:** Speedrun targets pre-incorporation teams of two to three people, often still employed full-time, investing up to $1 million before founders have even formed a legal entity. In the most recent 70-company batch, roughly a dozen teams had to quit jobs and incorporate from scratch before funds could be wired.; **Venture power law at seed stage:** Historically, approximately half of early-stage startups fail outright, another two to three out of ten return modest capital, and only the top decile generates the majority of returns. Speedrun accounts for this by tracking founders across multiple companies, re-funding promising individuals even after a first venture returns minimal capital..

How long is this episode of a16z Podcast?

This episode is 38 minutes long. SignalCast provides an AI-generated summary so you can get the key insights in about 3 minutes.

a16z Podcast

What It Takes to Build a Startup | Andrew Chen & Matt Perault

September 11, 2026

38 min episode · 2 min read

Andrew Chen,Matt Perault

Episode

38 min

Read time

2 min

Topics

Career Growth, Relationships, Investing

AI-Generated Summary

Published Sep 11, 2026

Key Takeaways

✓Early-stage investment thesis: Speedrun targets pre-incorporation teams of two to three people, often still employed full-time, investing up to $1 million before founders have even formed a legal entity. In the most recent 70-company batch, roughly a dozen teams had to quit jobs and incorporate from scratch before funds could be wired.
✓Venture power law at seed stage: Historically, approximately half of early-stage startups fail outright, another two to three out of ten return modest capital, and only the top decile generates the majority of returns. Speedrun accounts for this by tracking founders across multiple companies, re-funding promising individuals even after a first venture returns minimal capital.
✓Location as a regulatory decision: Two-to-three-person founding teams are highly mobile and actively choose where to incorporate based on cost, investor proximity, and regulatory climate. Founders building AI-native products who anticipate heavy state-level AI regulation factor that burden into their location decision before planting roots or hiring.
✓Little tech's policy representation gap: Early-stage founders lack lobbyists, policy teams, and time to travel to Sacramento, Washington, or Brussels. Policymakers who consult only large tech companies receive a structurally skewed view of industry needs. Venture firms can bridge this gap by arranging direct conversations between officials and curated groups of five to ten two-person startups.
✓AI tools compress founding team size: The current Speedrun cohort relies heavily on AI-assisted coding, allowing a single technical co-founder to build and iterate on product without outsourcing or early engineering hires. This makes the two-person model — one technical co-founder building inward, one business co-founder selling outward — more viable and capital-efficient than in prior cycles.

What It Covers

Andrew Chen, General Partner at a16z, describes the Speedrun program — a 12-week accelerator investing up to $1 million in two-to-three-person startups — and explains how early-stage founders operating from kitchen tables face regulatory burdens designed for large companies, with zero time or resources to engage policymakers.

Key Questions Answered

•Early-stage investment thesis: Speedrun targets pre-incorporation teams of two to three people, often still employed full-time, investing up to $1 million before founders have even formed a legal entity. In the most recent 70-company batch, roughly a dozen teams had to quit jobs and incorporate from scratch before funds could be wired.
•Venture power law at seed stage: Historically, approximately half of early-stage startups fail outright, another two to three out of ten return modest capital, and only the top decile generates the majority of returns. Speedrun accounts for this by tracking founders across multiple companies, re-funding promising individuals even after a first venture returns minimal capital.
•Location as a regulatory decision: Two-to-three-person founding teams are highly mobile and actively choose where to incorporate based on cost, investor proximity, and regulatory climate. Founders building AI-native products who anticipate heavy state-level AI regulation factor that burden into their location decision before planting roots or hiring.
•Little tech's policy representation gap: Early-stage founders lack lobbyists, policy teams, and time to travel to Sacramento, Washington, or Brussels. Policymakers who consult only large tech companies receive a structurally skewed view of industry needs. Venture firms can bridge this gap by arranging direct conversations between officials and curated groups of five to ten two-person startups.
•AI tools compress founding team size: The current Speedrun cohort relies heavily on AI-assisted coding, allowing a single technical co-founder to build and iterate on product without outsourcing or early engineering hires. This makes the two-person model — one technical co-founder building inward, one business co-founder selling outward — more viable and capital-efficient than in prior cycles.

Notable Moment

Chen reveals that regulatory compliance stacks are cumulative, not incremental — a California AI startup must simultaneously navigate data provenance rules, existing privacy law, and newer AI-specific legislation from day one. Policymakers typically evaluate only the marginal burden of each new bill, missing the compounding weight founders absorb from the start.

Know someone who'd find this useful?

Episode Transcript

This is truly little tech. The average team is two to three people. They're running their companies not not in their office, not in the co working space. They are running it at the kitchen table. For these founders, they are so mission focused trying to survive as a business. They just don't have time to participate. They don't have lobbyists. They're not really represented. It's a choice whether or not each state or each city wants to have startups or not. What does it actually look like to build a startup from day one? In this episode, Matt Perault sits down with a sixteen z general partner and speedrun lead, Andrew Chen, to talk about little tech, the tiny teams at the very beginning of building a company. Andrew shares what life looks like for founders who are often just two or three people building from a kitchen table, focused on getting a product to work and finding their first customers. They rarely have lobbyists, policy teams, or even the time to participate in debates about rules that could directly affect them. Matt and Andrew discuss how regulation can influence where startups choose to build, what policymakers can learn by hearing directly from founders, and what it takes to create an environment where the next generation of companies can get started. Andrew, welcome to the a 16 z AI policy brief. Thank you for having me. So you you lead our speedrun program. So can you tell us a little bit about what speedrun is? Yeah. So, know, many folks in the audience will know that startups have to come from somewhere. And so, you know, our job is to try to find them on day one when the founders are just starting to think about their companies. And so, what's great is we we host a program, you know, based in the San Francisco office where we will announce to the broader Internet and through our marketing channels and podcasts and substack newsletters and everything else that we'll be investing up to a million dollars into brand new startups. And the ones that we really focus on are the ones that where folks are just just getting going. You know, folks that maybe have full time jobs and they're starting to come up with, you know, with with with a new idea with their best friend and they wanna go start something. It could be folks that are have already left their, you know, roles or or they just graduated from school, and they are are coming up with something new. And so what we do is we spend twelve weeks with them. They get the full force and and power of the firm and all of our relationships and all of that great stuff. And then after we invest, then we share them with the broader ecosystem. We host a a speed run demo day where we have over a thousand angel investors and seed funds and …

Get the full transcript (7,485 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all a16z Podcast transcripts →