E20: Seth Berman, Cofounder of Susa and Kivu Ventures on History and State of Early Stage Venture Capital

Episode Summary

Executive Summary: Seth Berman traced Sousa Ventures’ evolution from an early seed fund to a later-stage investor as seed became overcrowded and competitive. He argued that founders should be problem-solvers with urgency, persistence, and domain insight, while AI is lowering barriers for creatives and non-technical builders. He also highlighted rising dilution, harder GTM, and the importance of timing, incubation, and selective high-value support.

Main Topics: Evolution of seed and pre-seed venture (Priority: 5/5): Berman explained how seed investing changed from a collaborative, small-check ecosystem into a crowded winner-take-all market with heavy competition from multi-stage firms and 900+ seed funds. Sousa Ventures’ strategy shift to Series A/B (Priority: 5/5): Because seed rounds became harder to win and later-stage companies offer clearer product-market fit, he moved focus from seed into Series A and B via Kivu. What makes a strong founder (Priority: 5/5): He emphasized founders who are problem-solvers, highly persistent, urgent, and resilient, especially those with lived experience or domain expertise. AI, physical intelligence, and democratized building (Priority: 4/5): The discussion covered vertical AI, physical intelligence, robotics, and how tools like Replit, Cursor, and Lovable let more people build products without deep technical backgrounds. Timing, dilution, and why stage matters (Priority: 4/5): Berman stressed that venture outcomes depend heavily on timing, with seed investors facing extreme dilution and long, uncertain paths to exit. Platform value, events, and founder support (Priority: 3/5): He argued that most VC platform services are overrated unless highly targeted, and that a few high-caliber events and practical help matter more than broad programming. Personal investing lessons and major misses (Priority: 3/5): Berman reflected on misses like Pinterest, Uber, and Anthropic, using them to show how outcome thinking and timing matter more than entry valuation alone.

Key Arguments: Seed and pre-seed are now crowded by 900+ seed funds and most multi-stage firms, making competition for great deals far more intense. Many startups now raise with little or no product if they have strong pedigree, which pushes experimentation earlier but also creates more noise. The best founders are not just people who want to start companies; they are people trying to solve a real problem and who will persist through setbacks. Immigrants often make exceptional founders because they feel they must succeed and are less likely to give up. AI is lowering the cost of building software and will expand access to entrepreneurship for creatives and non-technical people. Physical intelligence and robotics may become the next major wave, especially in manufacturing, agriculture, construction, and consumer wearables. Seed-stage dilution is so severe that later-stage, product-in-market investing can offer better risk-adjusted outcomes. VC platform services matter less than being a trusted, sector-savvy partner who can help founders raise the next round and solve real problems. Timing is as important as thesis: being too early can be as bad as being wrong, especially in deep tech and AI-adjacent sectors. Incubation can create better ownership, but it requires heavy bandwidth and talent-finding skill. The market has shifted so far that pre-seed and seed have largely converged, while Series A now resembles old seed. Better public market disclosure and AI-driven analysis may eventually make earlier IPOs more viable, increasing liquidity for employees and LPs.

Data Points: Years since Sousa Ventures was founded: 13 years - Berman described founding Sousa Ventures about 13 years ago, when seed funds were much rarer. Angel investing start year: 2010 - He said he began angel investing in 2010 and attended early YC demo days. YC demo day size then: ~12 founders and 12 investors - He contrasted early Y Combinator demo days with today’s much larger events. YC demo day size now: ~250 companies and thousands of investors - He noted the ecosystem has expanded dramatically. Angel check size: $5k–$25k - He said he initially wrote very small angel checks due to limited capital. Angels invested: 22 companies - He said he made 22 angel investments before starting Sousa Ventures. Quick exits from angel portfolio: 6 exits - He said six of those 22 companies sold relatively quickly. Example buyers: Google, Yahoo, BlackRock - He cited these acquirers as examples of early exits from his angel portfolio. First Sousa fund size: $25 million - He described Sousa’s first fund as a small early seed fund. Robinhood and Flexport: First investors in both - He said Sousa was fortunate to be among the earliest investors in these companies. Seed fund count today: 900+ - He cited the explosion of seed funds in the ecosystem. Typical seed valuation in 2010: $2M–$8M - He said those were common valuations when he began angel investing. Valuation level that felt extreme then: $10M+ - He recalled $10M valuations as unusually high at the time. Dilution from formation to exit: ~90% - He argued that repeated fundraising creates massive dilution for early investors. Potential check size at seed now: $50k–$250k - The conversation referenced that current pre-seed/seed checks often resemble former scout checks. AI science prediction window: 2–5 years - He said LLMs may reach “prime human” level intellect within this timeframe. Wearables prediction: 100 million pairs in 1.5 years - He predicted AI-connected eyewear could reach 100 million pairs sold across the ecosystem in about a year and a half. 2021–2022 vintage outlook: Likely weaker vintage - He said he does not think the 2021–22 venture vintage will perform well. Target LP return: 5x net - He said Sousa aims to deliver at least a 5x net return to LPs. IPO scale threshold: $100M–$150M revenue historically; potentially ~$1B later - He suggested public market requirements are rising and may demand much more revenue in the future. Robinhood IPO conviction: Kept 80% of stock - He said he retained most of his Robinhood holdings after the IPO. Anthropic reference valuation: $6.1 billion - He cited a missed opportunity to invest when Anthropic was valued around $6.1B. Anthropic revenue at that time: $5M current, $50M projected - He said Anthropic was doing about $5M revenue and projected to hit $50M that year when he evaluated it.

Pivotal Quotes: "I think there's two types of entrepreneurs. I think there's an entrepreneur that wants to start a company and there's an entrepreneur that wants to solve a problem." — Seth Berman: He used this distinction to explain the kind of founders Sousa wants to back. "The gates are down." — Seth Berman: He used this phrase to describe how much easier it is today for creatives and non-technical people to build products with AI tools. "A lot of companies die not because they, you know, run out of funding, but because founders lose optimism." — Seth Berman: He was explaining why persistence and founder mindset matter more than capital alone.

Implications: Venture is moving earlier, faster, and more competitively, but durable advantage is shifting toward clear problem selection, founder quality, and timing. AI will expand who can build, while later-stage investing, incubation, and selective high-touch support may offer better risk-adjusted returns.