The Next Biggest Moat in AI: Code Can Be Copied, Products Can Be Copied, But There's One Thing Nobody Can Steal

The Next Biggest Moat in AI: Code Can Be Copied, Products Can Be Copied, But There's One Thing Nobody Can Steal

Products can be copied. Categories can be rebranded. Technology advantages can evaporate in months. The one thing nobody can steal in the AI era? Your company's organizational form.

This bold claim comes from Jaya Gupta, Partner at Foundation Capital. And when her article "The Next Biggest Moat in AI" landed on X, it ignited a firestorm — 3.2 million views and counting.

People couldn't stop sharing it. "This connects so many dots I couldn't articulate," one commenter wrote. Another simply called it "a damn good piece."

Gupta should know what she's talking about. Earlier this year, her co-authored piece "Context Graphs: AI's Trillion-Dollar Opportunity" (with Ashu Garg) became a blueprint for the next wave of Agent-native startups. That essay argued that decision traces — not models, not data — are the real moat for next-gen Agent companies.

Now she's turned the same lens inward, on the companies themselves.

Everything Is Converging. What's Left?

It's impossible to miss: everything in AI is collapsing toward the middle.

A year ago, you couldn't have predicted which companies would be direct competitors today. Apps are sinking into infrastructure. Infrastructure companies are crawling up into workflows. Every startup has rebranded itself as some kind of revolution.

Buzzwords rotate every few months: Context Graphs, Action Systems, Organizational World Models. A new category gets named, every website refreshes overnight, and within weeks the market is flooded with companies claiming to be the inevitable platform for the future of work.

When models iterate fast, interfaces converge, and product velocity becomes commoditized, the visible parts of company-building are the easiest to copy.

What's much harder to replicate is what's underneath: how a company attracts extraordinary people, organizes their ambition, concentrates judgment, distributes power, and turns work into a compounding system that no other company can duplicate.

The best companies have always known this: Talent isn't an input to the company. Talent IS the company.

But in AI, this truth cuts sharper than anywhere else, because everything else is moving so fast.

If products can be copied, categories can be renamed, and technical advantages can vaporize in months — then the only remaining question is: what kind of organization did you build around the people who can make this happen?

Company form is the moat.

Great Companies Are Organizational Inventions

The most important companies are, at their core, organizational inventions. They build a new kind of organization around a new kind of work — and in doing so, they make a new kind of person possible.

OpenAI isn't like a university lab. It isn't like a corporate R&D division. It isn't like a traditional software company. It puts frontier model training at the gravitational center, with safety, policy, product, infrastructure, and deployment all orbiting around that core.

This structure enables a new kind of researcher — someone who wants to stand at the intersection of science, product, geopolitics, and civilization-level risk, all at once.

These companies don't fit in any existing category. Neither do the people who build them.

Great companies aren't just places talented people go. They're places where people become who they were meant to be.

Form Determines Who Can Become Themselves Here

The best companies in the world compete not just on market or compensation. They compete on identity.

Ambitious people desperately want a few things: to feel exceptional, to be close to power, to become undeniable, to keep their options open, to belong to a mission, to stand at a hinge of history. They often can't articulate exactly what they're chasing.

That's why the strongest institutions start recruiting absurdly early. They target freshmen at top universities — before someone's self-concept has hardened, before they've figured out what they want to be famous for, before they can tell the difference between what they're good at and who they want to become.

A great company gives you a language for your ambition. It tells you: that thing you've been circling in your head but couldn't quite name? You can do it here.

You can be the person who pushes the Mars timeline forward. The person in the room where the frontier moves. The person who makes work happen inside broken institutions. The person whose work becomes undeniable.

This is why great institutions are just shells for a certain kind of person.

Many companies choose to compete on cash. This is the most boring strategy in the book.

Cash can buy a person. It can rarely make one.

Top talent stays loyal only when the company offers something more concrete than money: a path to becoming who they already want — or haven't yet realized they want — to be.

Every Emotional Commitment Needs a Structural Commitment

This is where most companies fail.

If you say customer closeness is your moat, but the roles closest to customers have the lowest status in your company — your promise is a lie.

If you say ownership matters, but all decision rights are concentrated at the top — your promise is a lie.

If you say mission matters, but your mission doesn't offend anyone, filter anyone, or cost anyone anything — it's not a mission.

So what do people actually want?

• People want to feel exceptional. Rare. Seen. Irreplaceable. The pitch: only you can do this. You're unique enough to build it here. This targets the quiet insecurity at the heart of most high-performers: the fear that their excellence is fragile, that someone else could do their job, that they've never actually been truly seen. This only works in a form small enough that one person can genuinely change the company's trajectory.

• People want to feel destined. Like their life has been heading toward this inevitable place. Anthropic is the best current example. They're one of two or three companies deciding how this technology gets safely deployed — and the people in that room are the ones actually doing it. This feeling only holds in a form that is genuinely one of the top two or three.

• People want to feel like they're not missing out. That they're in the room where compounding happens. Look at how many legendary CTOs Anthropic has poached this quarter alone.

• People need to prove something. The brilliant investment banker with a lifetime of credentials who quietly suspects their work means nothing.

• People want optionality. McKinsey mastered this: generalist placement, two-year analyst cycles, the ability to try different industries at 21 when you have no idea what you want.

• People want proximity to power and status.

• People want to sacrifice for something bigger than a paycheck. Most companies call this mission. What really works is a cult around what this group genuinely believes.

The new wave of AI labs has sharper value propositions than the mission statements of the last era. Because every one forces you to pick a side.

Open source puts you against closed-source labs.

The strongest mission is necessarily one that makes some people NOT want to join — which is exactly the same thing as making the right people desperately want to join.

People are people. The best companies just pick one or two of these emotions that a specific candidate is intensely craving — and build a form around that person.

The Founder's Real Question

For founders, the question isn't "how do we tell our story better?"

The question is: what kind of person can only become themselves here?

Most companies tell their literal story: we're building models. We're building rockets. We're building a CRM for X. We're automating Y.

Accurate. Honest. And in today's world, accuracy is no longer enough to hire the best talent.

The best companies operate on a higher plane. They describe what change their existence makes possible: which industry gets resurrected, which institution gets rebuilt, which civilization-level bet gets won, which kind of human labor becomes possible for the first time.

Some people think this is just marketing. A different narrative from fundraising.

But a story's scale must match the company's form. A big story with a small form sounds like bragging. A small story with a big form makes the best people feel it's not worth joining.

Candidates evaluate the alignment between these two things — even if they can't quite articulate it themselves.

If you believe customer closeness is your moat, then customer-facing work must be the highest-status work in the company.

If you believe speed is your moat, then decision rights must live at the edge.

If you believe talent density is your moat, then mediocre people cannot define your company's tempo.

If you believe deployment is your moat, then the people closest to reality must hold power — not just responsibility.

For Those Making the Choice

And for those on the other side of the table — the ones choosing where to go — the signal to listen for is different.

Don't listen to the emotional promises. Look at the structural commitments.

Being selected is different from being seen. Being selected is emotional: you're special, you belong here. Being seen is structural: this is your scope, your compensation, and what the company will change for you when you succeed.

The most dangerous person in any organization is the one whose emotions have been filled — but whose structure is still that of an employee.

The most dangerous promises are priced in time. Over time, work deepens. Scope expands. Structure catches up. But time doesn't tap you on the shoulder. You'll reach the next stage of your life and realize the company's promise was never actually kept.

---

Building a company with the organizational form to win in AI? TokenSmind provides the infrastructure layer that lets you focus on what matters — unified access to 200+ models, granular permission management, full billing analytics, and enterprise-grade stability. So you can build the extraordinary organization — and leave model infrastructure to us.

Original source: Foundation Capital, Jaya Gupta's "The Next Biggest Moat in AI"