AI was made for bootstrappers

VC and big tech pays. Bootstrappers eat for cheap.

The current AI boom is being built and funded by giants. OpenAI, Anthropic, Google, Microsoft, NVIDIA — hundreds of billions of dollars of capex and venture money, more compute than the entire internet ran on a decade ago, all aimed at pushing the frontier of what these systems can do. The story being told is that AI is the great new technology of our era, ushered in by the biggest companies in history, for the benefit of… whom, exactly?

The most counter-intuitive answer — and I think the correct one — is: the smallest. AI in 2026 is a bootstrapper's technology wearing a hyperscaler's coat. It's been built at trillion-dollar scale, but its real value lands hardest in the hands of someone running a company of one or two or five.

Why AI struggles inside a big company

I spend most of my working life in and around enterprise tech, so I get to see how AI is actually being adopted across organisations of every size. The pattern is remarkably consistent: the bigger the company, the worse the fit.

Not because the technology is hard — it isn't. Because the organisation is. Large enterprises run on politics, silos, remits and departmental boundaries. There's a procurement process for everything. There's a security team, a compliance team, a data governance committee, and an architecture review board. The CIO has to answer to a board, who have to answer to regulators. And on top of all of that, there's a fundamental tension at the heart of every modern AI system: nobody, including its makers, can fully explain what it's going to say next. For a regulated business, that's not a feature.

So in big enterprise, AI gets caged. It gets pushed into narrow, low-stakes use cases. It gets passed through compliance layers that strip away most of what made it interesting. It gets deployed as a glorified search bar inside a permission-checked sandbox. The output is real but modest. The effort to get there is enormous.

Why AI thrives in a bootstrapped one

Now picture a small bootstrapped team. There's no silo. There's no committee. The remit is the whole business. The same person is doing sales, product, support and finance across a single afternoon. And the thing AI is uniquely good at — being a fluent generalist that can pivot from a customer email to a SQL query to a contract review in seconds — is exactly the kind of work that small team has always had to do anyway. There's a near-perfect match between what AI is and what a bootstrapper needs.

A bootstrapper doesn't have to enterprise-procure an AI strategy. They sign up, they pay $20 or $100 a month for a serious-grade model, and they start. No security review. No vendor risk assessment. No steering committee. The friction between "this might be useful" and "this is now part of how I work" is approximately zero.

The asymmetry

The same tool is producing wildly different amounts of leverage depending on who's holding it. A 200-person enterprise spends a year and seven figures to deploy AI into a fraction of one workflow. A two-person bootstrapper spends a Tuesday afternoon and £40 to deploy it across their entire company. The big company gets a small percentage uplift on the bits AI was permitted to touch. The small company effectively gains a third, fourth and fifth team member overnight.

And it gets worse for the big company over time, not better. The organisational features that make it too slow to deploy AI well — the committees, the layers, the audit trails, the regulated workflows — are the same features that get harder to dismantle the bigger you get. This isn't a temporary state of affairs while the giants "catch up". It's structural.

Eating the lunch of the well-funded

The interesting consequence is that AI may end up doing what it was always meant to do — but for everyone except the people who paid for it. The classic asymmetry between a small team and a scaled-up rival has always been roughly: the small team has speed and focus; the big team has people, money and tooling. AI flattens the second half of that. For the first time, a tiny team can buy access to "people" by the API call, and the tooling cost has dropped from millions to tens of pounds.

What's left is the things bootstrappers were always good at — speed, focus, and not being beholden to anyone else's idea of what to build — now sitting on top of a brand-new layer of generalist labour that scales with their ambition rather than their headcount. The heavily-funded scale-up, meanwhile, has all the cost and politics of being big without yet having the scale to absorb either. It's the worst seat in the house.

VC and big tech are paying for it. Bootstrappers are eating for cheap.

There has never been a better time to bootstrap a company

I write that sentence at the risk of sounding like every other person who's ever blogged about starting a business. So let me be more precise: it has never been cheaper, faster or easier for a small focused team to do the work that previously required tens of millions of dollars of headcount and tooling. The marginal cost of generalist labour has collapsed. The marginal cost of running a bureaucracy has not.

If you've been waiting for the right moment to start something on your own terms, this is it. The biggest companies in the world have spent the last three years building the tool you needed. They were the ones who paid for it. You're the one who gets to use it.