Summary
Investors described how AI startup evaluation is shifting from hype to durable revenue, token COGS, defensibility against model labs, and whether products survive API changes. The strongest opportunities discussed were upstream infrastructure bottlenecks, AI-native enterprise software that compounds through doing work, and founders with deep technical skill, agility, and human-centered judgment.
Key Topics
5 key topics from George Hammond, Jai Das, Katelin Holloway and Quentin Clark at HumanX 2026 USA. Thicker branches were mentioned more often in the talk.
Are AI Startups Building Real Value or Just Chasing Hype?
Click a topic to open it
Key Takeaways
- Test whether products survive API and model changes before investing.
- Underwrite token COGS and durable revenue, not headline growth alone.
- Favor upstream infrastructure where model labs depend on external bottlenecks.
- Build operational flywheels that improve as AI software performs work.
- Assess founders for deep domain skill, agility, and human-centered taste.
“if your product roadmap disappears when the API changes, you're probably not a company that we're gonna look at further.”
Share
Up Next
SF
WatchNext in agenda
How Google Is Scaling Enterprise AI Adoption with Trust
Sharon Goldman, Francis deSouza
enterprise aigoogle cloud
KHK
WatchAlso on ai startups
What Makes a Founder Stand Out in the AI Gold Rush
Kate Clark, Hans Tung, Kerry Wang, Sangeen Zeb
ai startupsventure capital
NM
WatchAlso on enterprise ai
How WRITER Won the Fortune 500
Natasha Mascarenhas, May Habib
enterprise aiai agents