How Hamilton Helmer's 7 Powers Apply to AI Startups

What happens when you map Hamilton Helmer's 7 Powers to AI startups: counter-positioning and switching costs win; speed alone is not a moat.

Garry TanhostHarj TaggarhostDiana HuhostJared Friedmanhost

Oct 2, 202545mWatch on YouTube ↗

WHAT IT’S REALLY ABOUT

Seven Enduring Moats Every Ambitious AI Startup Must Master Fast

The episode reframes Hamilton Helmer’s “7 Powers” as seven types of moats that still apply in the AI era, despite originating in a pre-AI, Web 2.0 world.
The hosts argue that early-stage founders should obsess over speed and solving painful, concrete problems first, then worry about moats only after they’ve built something valuable to defend.
They walk through each moat—process power, cornered resources, switching costs, counter-positioning, branding, network effects, and scale economies—and illustrate how AI startups are already using them today.
Throughout, they emphasize that AI doesn’t eliminate moats; it changes where they appear (e.g., in workflows, data, evals, and pricing models) and often amplifies advantages for fast-moving, deeply embedded application-layer startups.

IDEAS WORTH REMEMBERING

5 ideas

Solve a painful problem first; design moats later.

Founders shouldn’t reject ideas because they can’t see a long-term moat on day one. The priority is finding a truly painful, urgent customer problem and shipping fast—moats emerge as you build, integrate, and learn.

Speed is the only real moat at the very beginning.

Early-stage AI startups like Cursor won by shipping useful features on one-day sprints, outpacing large companies bogged down by process. Relentless execution speed lets you discover valuable verticals and iterate before incumbents react.

Process power comes from hard, unglamorous engineering that’s tough to copy.

Mission-critical AI agents for banks, legal workflows, or KYC require years of edge-case handling, evals, and reliability work—far beyond a weekend demo. That accumulated complexity and operational know-how becomes a powerful moat.

Own unique data, workflows, or access to create cornered resources.

Deep relationships with regulated customers, private workflow data, custom evals, and even proprietary fine-tuned models become assets others can’t easily replicate. Forward-deployed engineering that codifies real-world processes is a key way to build this.

Exploit and create switching costs through deep integration and memory.

AI startups that embed themselves into a customer’s bespoke workflows—or accumulate rich user memories and personalization—make it extremely painful to rip them out, even if an alternative is slightly better or cheaper.

WORDS WORTH SAVING

5 quotes

A moat is inherently a defensive thing, and you have to have something to defend.

— Harj Taggar

The early stages at the beginning, the only moat that startups have is really just speed.

— Varun Mohan (quoted, Windsurf)

The version you build in a hackathon isn’t useful to anyone.

— Jared

The cornered resource doesn’t have to be a diamond mine. It could be the diamond mine in your customers’ heads.

— Garry Tan

Don’t use these frameworks to count yourself out prematurely.

— Jared

Why AI founders are suddenly obsessed with moats and “ChatGPT wrapper” fearsThe primacy of speed as an unofficial eighth power for AI startupsAdapting Hamilton Helmer’s 7 Powers to AI applications and agentsConcrete examples of process power and deeply-engineered AI workflowsCornered resources in AI: data, models, regulatory access, and customer workflowsNew and old forms of switching costs in AI-era enterprise and consumer productsCounter-positioning, branding, network effects, and scale economies in AI markets

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.