In the AI Era, Your Moat Is Not the Model
Every founder has been hit with the one question that makes your chest tighten a little:
"What's your moat? What stops a competitor from stealing your customers?"
In the AI era, the familiar answers — "my technology," "my model" — have lost their magic. Models are strong today and stronger tomorrow. Anyone can call the OpenAI or Anthropic API. A clever feature gets copied next week.
So what's left to defend with? The three pieces below — drawn from Aravind Srinivas (CEO of Perplexity) and Anthropic's new founder's playbook — paint a counterintuitive but very down-to-earth answer.
1. The trap called "moat"
Srinivas said it bluntly in a recent interview: in the first 1–2 years, hunting for a moat is a trap.
It sounds backwards, but it holds up. When the company is still tiny, a real moat doesn't exist yet. You sit around inventing a theoretical "fortress wall" — patents, network effects, economies of scale — just to reassure yourself. And that exercise of "designing a moat in your head" actually slows you down, at the exact moment when speed is the only thing you've got.
At this stage, your only temporary moat is this: moving faster than your competitors.
2. Moving fast is a form of humility
This is Srinivas's best line:
"Moving fast is the ultimate expression of humility — because you constantly collide with reality."
We usually think moving fast means being aggressive, overconfident. Srinivas flips it completely. Picture two kinds of founders:
The slow kind (actually the arrogant one): "I already know what customers want." Spends 6 months making it perfect, ships it — and finds out nobody uses it. Half a year wasted just to learn they were wrong.
The fast kind (the humble one): "I'm guessing customers want this, but I'm not sure." Builds a rough version in 2 weeks, pushes it to real users, gets criticized, fixes it, ships again. Every two weeks they learn whether they were right or wrong.
The fast founder is more humble, because they don't trust the story in their own head — they're willing to let reality slap them in the face and correct them. You go slow out of ego (so sure you're right you skip verification); you go fast out of humility (knowing you might be wrong, so you rush to find out).
Speed, here, isn't sloppiness. It's the discipline to let real data override your ego.
3. So where's the real moat? — In your users
If the model isn't the moat, and features aren't the moat, then what is? The answer is short:
Your moat is your users.
What a competitor can't copy doesn't live in your codebase — it lives in the relationship between you and your users:
- Habit. They're used to it, too lazy to switch. (You're so used to typing into Google that even when something better shows up, you can't be bothered to move.)
- Data. The more they use it, the more the product understands them, the more it personalizes. A brand-new competitor has none of that data, so it serves them worse.
- Trust. They open your product by default, before even thinking about an alternative.
- Network. Their friends are on it too — leaving means losing the connection.
Stitch this together with the two points above and you get a clean loop: you move fast (humility) to win users and make them stay before competitors can react. Those users are the moat. Speed isn't the destination — it's how you dig the moat.
4. The last piece: the AI-era founder is an orchestrator
The next question is very practical: where does that speed come from when the team is just a handful of people?
This is where Anthropic's new "Founder's Playbook" offers an answer. The founder's role shifts: from doing the work to orchestrating a team of AI agents (orchestrator of agents).
For every task, the first question becomes: "Can I hand this to an agent first?" Market research, drafting content, scaffolding code, writing sales material — push the first pass to AI. The founder keeps only the four things machines can't do for you: judgment, taste, trade-offs, and trust.
The consequence: the team stays genuinely small, you "reach revenue before scaling headcount," and — most importantly — your rate of collision with reality skyrockets. A good orchestrator of a fleet of agents can ship and learn faster than a ten-person team doing it all by hand.
Wrapping it into a formula
The three scattered ideas are really one chain:
- Don't waste the first 1–2 years building an imaginary moat.
- Instead, move fast — because moving fast is humility, the willingness to let reality correct you.
- Move fast to win users, because users are the real moat.
- Where does that speed come from? From being a founder-as-orchestrator, letting agents carry the repetitive load.
In an era where everyone has equally great models, the winner isn't the one with the coolest technology. It's the one humble enough to move fast, and fast enough to make users fall in love before a competitor ever shows up.