The Seven Powers: How AI Startups Build Moats

In this episode of Y Combinator's Lightcone podcast, hosts Garry Tan, Jared Friedman, Diana Hu and Harj Taggar revisit Hamilton Helmer's 'Seven Powers' framework and reboot it for the AI era. Prompted by anxious college founders who fear their AI-agent companies are just easily-cloned 'ChatGPT wrappers,' the group argues that durable moats absolutely exist in AI, they're just not obvious. They walk through process power, cornered resources, switching costs, counter positioning, branding, network economies and scale economies, adding an eighth power the book leaves out: speed. Along the way they use concrete, current examples like Cursor, Windsurf, Casca, Greenlight, Plaid, Happy Robot, Salient, Legora vs Harvey, Giga ML, Speak vs Duolingo, Avoca, Exa and OpenAI vs Google to show how founders should think about defensibility, and when.