Venture capital is shifting away from certain AI SaaS startups as investors grow wary of companies that lack true differentiation. Aaron Holiday of 645 Ventures notes that generic horizontal tools, thin workflow layers, and surface-level analytics are no longer attracting funding. Investors now favor startups building AI-native infrastructure, vertical SaaS with proprietary data, and systems that help users complete tasks rather than just manage them.

Abdul Abdirahman of F-Prime warns that vertical software without proprietary data moats is falling out of favor, while Igor Ryabenkiy of AltaIR Capital emphasizes that UI and automation alone no longer constitute a competitive advantage. The barrier to entry has dropped significantly, making it harder for companies to build sustainable moats. Ryabenkiy points out that massive codebases are no longer an advantage; instead, speed, focus, and adaptability matter more.

Jake Saper of Emergence Capital highlights a crucial shift: products focused on human workflow stickiness may struggle as AI agents take over tasks. He cites the difference between Cursor and Claude Code as illustrative - one owns the developer workflow while the other simply executes tasks. As Anthropic's model context protocol makes integrations easier, being a connector is becoming a utility rather than a moat. The message is clear: investors are reallocating capital toward businesses that own workflows, data, and domain expertise, while moving away from products that can be easily replicated.