Strategy & ThesisApril 27, 2026ยท6 min read

Moats Are Dead. Long Live Canals.

Why the best companies in an AI-native world look less like fortresses and more like infrastructure.

TC
Trace Cohen
3x founder, 65+ investments, building Value Add VC

Quick Answer

In an AI-native, API-driven world, traditional moats โ€” proprietary data, lock-in, walled gardens โ€” are losing effectiveness. The new defensibility belongs to 'canal' companies: infrastructure that others route through by choice. Value now accrues to whoever facilitates movement of data, decisions, and transactions, not whoever walls them off.

For years, tech worshipped the moat. Defensibility meant keeping competitors out.

Proprietary data. Lock-in. High switching costs. Walled gardens.

The logic was medieval: survive by fortifying.

But in an AI-native, API-driven world, many moats are starting to look less like protection and more like prisons.

A moat often extracts. It traps users through friction. It holds data hostage. It makes leaving painful. It can preserve incumbency, but it doesn't always create flow.

A canal does the opposite.

A canal is engineered openness. It moves value.

Historically, moats protected castles. Canals built economies.

That distinction matters.

The best companies increasingly look less like fortresses and more like infrastructure. They don't win because no one can get in. They win because everyone wants to route through them.

A Moat Says:

  • โœ•Stay inside.
  • โœ•Defensive.
  • โœ•Blocks.
  • โœ•Scales through exclusion.

A Canal Says:

  • โœ“Build on top.
  • โœ“Generative.
  • โœ“Connects.
  • โœ“Scales through throughput.

The Modern Winners Are Canals

Think about the modern winners. Stripe wasn't built as a moat as much as a canal for internet commerce. Twilio became programmable plumbing. Amazon Web Services turned infrastructure into waterways.

The strongest platforms increasingly expose APIs, invite ecosystems, enable composability, and become routes others depend on.

That is canal thinking.

Value Accrues to Movement

In a world of agents, APIs, headless software, and interoperable models, value often accrues to whoever facilitates movement:

Moving data
Moving decisions
Moving transactions
Moving distribution
Moving trust

Not whoever walls them off.

The old question was: How do we make our moat wider?

The new question is: How do we become a canal others cannot afford to bypass?

That changes how you build.

Don't just ask whether your product has defensibility.

Ask whether it has flow.

The Flow Test

?
Can developers build through you?
?
Can partners distribute through you?
?
Can agents transact through you?
?
Can ecosystems coordinate through you?

The Deeper Shift

FROM

Protecting scarcity

TO

Orchestrating movement

FROM

Owning the castle

TO

Owning the route

And routes, historically, often outlast empires.

Moats protected kingdoms.

Canals created civilizations.

This essay reflects how I evaluate startups at Value Add VC. When I look at a company's defensibility, I'm increasingly asking not โ€œhow wide is the moat?โ€ but โ€œhow much flows through you?โ€ The companies in our AI Landscape tracker and AI Valuations dashboard that are winning the fastest are almost all canal companies โ€” they're infrastructure, not fortresses.

Frequently Asked Questions

Are startup moats still effective in an AI world?

Traditional moats built on proprietary data, switching costs, and walled gardens are becoming less effective as AI commoditizes many previously defensible capabilities. In an API-driven world, the most durable advantage often comes not from exclusion but from becoming infrastructure others depend on. Canal companies that orchestrate movement tend to outlast moat companies that restrict access.

What is the difference between a moat and a canal in startup strategy?

A moat protects a company by making it hard to compete against โ€” it scales through exclusion and relies on lock-in. A canal creates value by facilitating flow: developers build through it, partners distribute through it, and agents transact through it. Canal companies win by becoming routes others cannot afford to bypass.

What are examples of canal companies?

Stripe, Twilio, and Amazon Web Services are classic canal companies โ€” they became programmable infrastructure that other businesses route through by choice, not necessity. Their defensibility comes from throughput and ecosystem dependence, not switching costs or proprietary lock-in. The more that flows through them, the stronger their position becomes.

How can founders tell if their startup has 'flow'?

Ask four questions: Can developers build through you? Can partners distribute through you? Can agents transact through you? Can ecosystems coordinate through you? If the answer is yes to multiple, your startup has flow and is becoming a canal. If no, you may be building a moat that AI will eventually drain.

Explore 41+ free VC tools, dashboards, and recommended startup software.