Are network effects a moat against AI?
Yes, if they come from human interaction AI can't simulate: each real participant makes the product better for the others, and a copier can't fake the crowd.
Holds against AI
Updated November 19, 2026.
What is it?
Each user makes the product more valuable to other users: buyers attract sellers, members attract members, answers attract questions.
Does it hold against AI?
Real human networks hold. Bustamante lists network effects as holding under LLMs, and Hockenmaier counts them among the four compounding moats.
Two cautions. First, if the value of the network is content AI can now generate, the network is exposed: Stack Overflow's Q&A loop weakened when AI answered in the code editor. Second, a network only protects you if people keep coming to you.
How do you know if you have it?
Would your product be worse with half the users? Is the value the people themselves (trust, transactions, relationships), or content they produce that a model could imitate?
How do you deepen it?
Weight the network toward interactions AI can't stand in for: transactions, reputation, introductions, accountability. Make leaving costly by tying identity and reputation to your network.
Questions people ask
- Did AI break Stack Overflow's network effect?
- It weakened it. AI assistants answered questions inside the editor, so fewer questions reached the site and the loop slowed.
Sources
- How to tell if a startup has a moat (Lenny's Newsletter): https://www.lennysnewsletter.com/p/how-to-tell-if-a-startup-has-a-moat
- Ten vertical-software moats under LLMs (Bustamante, summary): https://hq-raydata-co.pages.dev/vault/06-reference/2026-02-16-vertical-software-selloff/
- AI disruption risk assessment (Reforge): https://www.reforge.com/blog/ai-disruption-risk-assessment
Does your business have a moat?
Five AI attackers, live web research and a published rubric. Free, about five minutes, no account. Or talk it through with Matt on a strategy call.