Distribution Is the Last Moat

(Was this newsletter forwarded to you? Sign up here.)

On two podcasts in late April, Evan Spiegel said where Snap is moving its money: out of engineering, into distribution.

The facts, because they carry the whole argument:

• More than two-thirds of Snap's new code is now written by AI.

• A service that used to take a team now takes "half a person's time," in his words.

• His conclusion, on Cheeky Pint with Stripe's John Collison: "I do think there's a real premium on distribution. I think people are going to reallocate more resources away from things like software engineering, for example, to distribution in order to grow faster."

He said a version of the same thing on Lenny's Podcast the day before, where the episode was built around one line: distribution is the most important moat now.

My read

This is the most CMO-friendly thing a public-company CEO has said in a while, and almost nobody framed it that way.

Strip it down. When AI makes building cheap, the product stops being the advantage. A competitor's intern can clone your feature before lunch. What they can't clone by Friday is an audience that opens your email, a brand a buyer already trusts, a reason to be the answer when someone asks. That is distribution. That is your job.

What "reallocating to distribution" actually looks like inside a company is more concrete than it sounds. It's the eleventh engineer you don't hire, funding instead a content team, a community, a partnerships lead, a brand budget you'd have defended line by line a year ago. It's treating the owned audience as infrastructure instead of a nice-to-have. Spiegel isn't speaking in metaphor. He's describing a headcount decision, and he's making it in public.

For a decade, marketing has argued it's an investment, not a cost center. Spiegel made that argument for you, from the CEO chair, with money instead of a slide.

The obvious objection

Snap sells distribution. Of course, its CEO says distribution is the moat. He's talking his book.

Fair. Take the discount. The logic still holds even if you don't trust the messenger. The barrier to building has genuinely collapsed, and not only at Snap. When the supply of "product" explodes, the price of the product falls, and the price of attention rises. That isn't a Snap talking point, it's what happens to any market when one input goes abundant, and its complement stays scarce. Spiegel is early to name it because he's standing on the scarce side. That makes him biased. It also makes him a useful leading indicator.

The part that should keep you up

Distribution is the new moat, and it's also getting eaten.

The same AI wave that made building cheap is rebuilding distribution underneath you. Agents are starting to do the buying. AI answers are sliding in between you and the customer. The reach you rent, and the search clicks you optimize are depreciating while you read this.

So "distribution is the moat" isn't a victory lap. It's a question: which kind do you own? Rented reach isn't a moat, it's a bill that climbs with demand. The moat is the distribution a platform can't tax away from you. An audience that's actually yours. Authority that gets you cited inside the AI answer instead of buried under it. Relationships that don't reset every quarter. That second fight, the one inside agentic commerce, is its own story.

What I'd do with it

If you're building the 2026 plan, this is ammunition. The CFO conversation changes when you can point to a public tech CEO pulling budget out of his own engineering org to fund distribution. Use the quote by name.

Then audit what you actually own. The test is one question: if a competitor rebuilt your product over a weekend, what would still be yours? Owned audience, brand demand, earned authority, and first-party data pass. Last-click ROAS and rented reach don't. Most budgets are pointed at the things that fail that test, because those are the things that are easy to measure. The stuff that passes is hard to attribute and hard to fake, which is exactly why it's a moat.

The uncomfortable follow-on: if your team's whole value is buying and optimizing rented reach, AI is coming for that work the way it came for Snap's engineers. The seat that survives is the one that builds owned demand, not the one that rents it.

Spiegel already moved his money. The only open question is whether you can prove yours is in the right place before someone moves it for you.e.