Why Would They Buy You?

A founder showed me a product recently that solved a real problem for a real customer. Clean design, working software, paying users. Two years ago I’d have spent most of my diligence on the technology — whether the team could actually build what they were describing, whether the CTO had the depth to scale it. That question barely matters anymore. The product was already real, and it already worked.

What I couldn’t stop thinking about was a different question: why wouldn’t their biggest competitor just build this next quarter?

The Load-Bearing Wall

Venture capital runs on exits. You invest early, the company grows, and eventually someone writes a check large enough to return the fund — either a public offering or, far more commonly, an acquisition. IPOs get the headlines, but acquisitions are the engine. Most successful venture outcomes end with a bigger company deciding it’s faster, cheaper, and less risky to buy a startup than to build the capability themselves.

That’s the load-bearing wall of the entire model. Faster, cheaper, and less risky to buy than to build. It’s been true for decades. It may not be true much longer.

The reason acquirers bought was straightforward. Building software that worked after being deployed into a complex world was slow and expensive. Engineering talent was scarce. A startup with a working product, real customers, and a two-year head start represented something that would cost the acquirer tens of millions and a year and a half to replicate — by which point the startup would be further ahead. The math favored buying. So they bought.

I wrote recently that we’ve entered the Tommy Lee Jones era of startup formation — if you can say it, you can see it. The same AI tools that make it possible for a founding team to build a credible product in weeks also make it possible for a team inside Salesforce or Microsoft or Google to do the same thing. The startup’s two-year head start just became a two-month head start. The eighteen-month build timeline just collapsed to a quarter. The risk calculus tips.

When build beats buy, acquirers stop buying.

The Model

This isn’t a threat to individual startups. It’s a threat to the model.

Venture capital works because of a specific return profile. You lose money on most investments and make it back on a few big wins. Those wins require an exit at a meaningful multiple — someone paying significantly more than what went in. Compress the multiples, reduce the frequency of acquisitions, and the math that justifies a venture fund starts to come apart. Your LPs don’t get patient when the math gets harder. They get quieter, and then they stop returning your calls.

I’m the one making those calls on behalf of our fund sometimes. We’re long on AI, obviously. I wrote last week about how we’re all talking about the AI revolution in the third person, like it’s weather instead of something we’re actively financing. This is me trying not to do that. Take it for what it’s worth.

What’s Still Worth Buying

So what survives the new math? Not the code base. If AI can reproduce the product, the product isn’t the asset. What an acquirer can’t build — the things that justify writing the check — are the things that were always the hardest to build in the first place.

Network effects, where they exist. If the product gets better as more people use it, that’s structural, and AI doesn’t change the math.

Brand. A company that means something specific to its customers — that owns a point of view in their minds — can’t be replicated by a model. You can’t prompt-engineer trust. You can’t fine-tune loyalty. A brand that someone would miss if it disappeared is worth acquiring. A brand that’s interchangeable with three others isn’t.

Team. In a world where anyone can build anything, the people who matter are the ones who know what to build, and who actually build it. It’s very much in fashion right now to chase the one-person unicorn, but I think this misunderstands the inherent nature of entrepreneurship.

Customers. A real installed base generating real revenue with real switching costs. An acquirer can build your product. They can’t build the two years of relationships your sales team earned one at a time.

Culture. A team that works in a way the acquirer’s team doesn’t — faster, sharper, with a creative edge or a quality bar the big company can admire but can’t replicate. Culture is the one asset you genuinely cannot buy at any price, except by buying the company that has it, and not then fucking it up. Rare.

Everything else — the features, the UI, the integrations, the clever engineering — is increasingly reproducible. Which means an acquirer will pay less for it, or nothing at all.

For Founders

If your exit plan ends with “and then a strategic acquirer buys us because our product is great,” you need a new plan. Your product may well be great. A capable team at the acquiring company can now build something comparable in a fraction of the time it took you, in part because you’ve uncovered just what they need to build. Not a knock on your work, just the new math.

Build the thing a big company can’t replicate even with AI. It’s almost never the code. It’s the business around the code — the brand, the customers, the culture, the market position. Those are the assets you’ll actually be selling, when it’s time for us both to get paid.

For VCs

I’m talking to myself here.

We’ve spent years getting comfortable evaluating technology risk. That risk has largely evaporated. What’s left is business risk, market risk, and the hardest one to underwrite: the risk that nobody will care enough about what this team has built to pay a premium for it, in a world where building the same thing costs next to nothing.

The exit isn’t guaranteed by the product anymore. It has to be earned by the business built around it. The founders who understand that — who build companies worth acquiring for reasons a machine can’t replicate — are the ones who’ll still generate the returns the venture capital model depends on.

We just need to get better at telling the difference.