Scot Wingo shared live at RDSW a framework he developed while fundraising: the 12 competitive moats that matter now and why most startups get this wrong. The core idea: You’re not building a product anymore, you’re building a fortress.
Back in Feb of 2025, we put this post out→
In there we said the second ‘what you should do’ is “Analyze and Cover your AI Risk”. Think of Scot’s talk and this post as the ‘going deeper’ on that.
Competitive_Moats_In_The_Ai_Era_Final_Sm.pdf
1 MB
Sources of Information on AI Competitive Moats
Here are the 8 core sources used for the deck if you want to go deeper. You can go REALLY REALLY deep and I recommend it to understand the nuances of each moat.
Y Combinator (Hamilton Helmer & The Lightcone Team) (aca YC) The 7 Most Powerful Moats For AI Startups This is the gold standard. They take Helmer’s classic “Seven Powers” and apply it directly to the AI era. Process power and counter-positioning are the standouts here. How AI Companies Will Build Real Defensibility Flint introduces the “Motte-and-Bailey” strategy. Speed and distribution are your temporary wooden fort (the bailey), buying you time to build network effects and deep embedding (the stone motte). Bessemer Venture Partners (BVP) Building Vertical AI: An Early Stage Playbook for Founders BVP drops a massive 10-principle playbook. They argue the best technical moats come from multimodality and solving hyper-specific, nuanced workflows. Andreessen Horowitz (a16z) a16z is pushing “services-led growth.” The idea? Use forward-deployed engineers to replace human workflows, capture the data, and eventually become the System of Record. I also love this Alex Rampell content and Defensibility in the Age of AI: Five Attributes Every Vertical Startup Needs Perez synthesized multiple VC frameworks into five core attributes. Her take on “Workflow Ownership”—owning the authoring layer where work actually begins—is crucial. Generative AI and Competitive Advantage: Where the Real Moat Is (and Isn’t) Azati does a great job separating the mirages (like prompt engineering) from the real moats (like domain expertise encoding and proprietary training data). Euclid outlines why moats matter even more in Vertical AI, focusing heavily on data loops and long-term lock-in strategies from day one. Insignia Ventures Partners In the Age of AI, Moats Matter More Than Ever The paradox: AI makes it easier to build but harder to defend. Insignia focuses on how data flywheels and switching costs are the ultimate survival tools. How to Watch/Listen?
Here’s the YouTube:
Or check out the audio for all your favorite podcast links, or listen right from our Substack: