Elad Gil Explains Why 90% of Startups Fail in Every Tech Cycle
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
In this engaging episode of Sorcery, host Molly Wood sits down with renowned angel investor and entrepreneur Elad Gil for a wide-ranging conversation about his diverse career spanning Google, Twitter, and founding Color Genomics, to becoming one of the most successful AI investors. (02:37) Gil shares his philosophy of viewing technology as a force for good that has elevated millions out of poverty, while also discussing his unique monument-building project aimed at inspiring future generations. The conversation delves deep into his investment strategy, covering everything from his early days running out of money as an angel investor to now leading $500 million+ rounds, his thoughts on the current AI wave compared to the dot-com era, and predictions about which companies might remain "forever private." (26:27) Gil provides fascinating insights into founder psychology, the importance of total addressable market size for trillion-dollar companies, and why he believes we're simultaneously in an over-hyped and under-hyped moment in AI development. • **Main themes:** Strategic technology investing across waves, the psychology of successful founders, AI market dynamics, and the intersection of policy with technological progress
Speakers
Elad Gil
Elad Gil is a serial entrepreneur, investor, and author who has built an impressive career spanning multiple technology waves. He's held senior roles at major tech companies including Google and Twitter, founded Color Genomics (a healthcare genetics company), and authored "High Growth Handbook." As an angel investor and fund manager, he's invested in over 150-225 companies including Stripe, Airbnb, Figma, Instacart, Notion, Perplexity, and Harvey, often leading rounds worth hundreds of millions of dollars. Gil was an early investor in the crypto wave (2016-2017) and generative AI (2021-2022), demonstrating his ability to identify and capitalize on major technological shifts before they become mainstream.
Molly Wood
Molly Wood is the host of Sorcery, a podcast focused on technology investing and entrepreneurship. She brings a journalistic approach to exploring the intersection of technology, business, and society, conducting in-depth interviews with leading investors and founders.
Key Takeaways
Recognize and Act on Major Technology Waves Early
Gil emphasizes the importance of identifying massive technological shifts before they become obvious to everyone. (12:07) He cites his early investments in crypto (2016-2017) and generative AI (2021-2022) as examples of thesis-driven investing during sea changes. When he played with GPT-2 and GPT-3, he recognized the massive step function improvement and read scaling papers that indicated what was coming. This led him to proactively seek out founders working in these areas when very few people were looking. The key is balancing opportunistic investing (waiting for great companies to appear) with thesis-driven approaches during clear technological inflection points.
Start with "Easy" Problems in New Technology Waves
Gil's advice to AI founders doing complex, difficult work is counterintuitive: focus on the "dumb, easy" problems first. (17:55) He uses the metaphor of choosing a nice path around a mountain instead of climbing straight up it. Early in technology waves, there's abundant low-hanging fruit that may seem non-defensible but can actually build defensibility over time. The really hard, complex problems should be tackled 3-4 years later when the easy opportunities are exhausted. This approach allows founders to establish market presence and build resources before tackling the most challenging technical problems.
Total Addressable Market (TAM) Determines Trillion-Dollar Potential
When asked what it takes to become a trillion-dollar company, Gil's answer is simple: TAM - just a big market. (34:31) He points to examples like Google (giant search/advertising market), cloud businesses, productivity software combined with important operating systems, and selling phones to billions of people. The lesson for founders is that even with exceptional execution, you're constrained by the size of the market you're addressing. The most successful founders combine finding giant TAMs with building products people actually care about within those markets.
Know When to Sell Based on Four Key Factors
Gil outlines four primary reasons companies should consider selling, providing a framework for exit decisions. (29:42) First, founder fatigue - when you're burnt out, unhappy, or fighting with co-founders. Second, when an acquirer is paying so far ahead of realistic projections that the math clearly favors selling. Third, competitive dynamics that threaten the company's position. Fourth, recognizing that most companies should sell because most don't achieve massive scale - 50-80% of venture-backed companies fail outright, and even at Series B stage, 40-50% fail. The key insight is that only a handful of companies should "never sell" and keep going regardless of offers.
Navigate Conflicts of Interest Through Transparent Communication
Gil's approach to handling competitive investments demonstrates the importance of transparency and founder consent. (15:23) When Stripe approached him while he was already invested in Square, he asked both sides about potential conflicts. A Square executive said it was absolutely a conflict, but founder Jack Dorsey said to proceed. Gil invested in Stripe, which proved wise as the companies didn't directly compete. His method involves asking founders: "Do you view this as a conflict? Is it okay if I proceed?" This approach maintains trust while allowing for portfolio construction, recognizing that 90% of perceived competitive threats don't materialize as companies grow in different directions.
Statistics & Facts
- In February 1999 alone, approximately 450 companies went public, with another 450 in the first part of February, and potentially 1,000+ more internet companies going public before that period - totaling around 2,000 internet company IPOs. (26:17) Gil uses this statistic to illustrate the dot-com bubble's scale, noting that of those 2,000 companies, only 2-3 are highly relevant today (like Amazon), with about a dozen others remaining important, while the other 1,980+ companies died or were acquired.
- The US currently generates 18% of its electricity from nuclear power despite not opening a new nuclear power plant in 50 years, compared to France at 70% nuclear and Japan at 25%. (19:21) Gil presents this to argue that energy constraints in AI are policy choices rather than technological limitations, as nuclear provides a clean, abundant, and statistically very safe energy source.
- Azure reported approximately $28 billion in quarterly revenue with 10-15% attributed to AI spending, translating to several billion dollars per quarter in AI spend on Azure alone. (31:59) Gil uses this figure to demonstrate that massive real money is already being spent on AI infrastructure, yet there's relatively little actual impact visible, indicating we're still in the very early stages of AI adoption and transformation.
