Bootstrapping is the new default. Building alone is not.

The last ten ventures I have worked on all opened the same way. The founder was excited, and the excitement was always about cost. With AI, we can build this for a fraction of what it used to take. No need for a big team. No need for a big round.

They were right. That part of the argument is settled. What interested me was what happened next in every one of those conversations: they still wanted to work with me, and they still paid my rate.

That is the detail worth sitting with. If AI had genuinely removed the need for experienced people, the logical next move would have been to stop hiring them. Instead the pitch simply changed shape. It stopped being “here is a team to build your product” and became “here is how to use AI properly instead of handing an agency a six-figure fee.” The value did not disappear. It moved.

Fewer people, not more

At BSD Education, the platform I spent years helping build into a global EdTech business, we reduced the size of the engineering team. Not because of a downturn, and not to cut costs for their own sake. The benefit of adding people had simply stopped being there. Every additional head was contributing less to actual output than it was adding in management overhead. More people meant more coordination, more context to keep in sync, more of my week spent on the org rather than on the product.

What worked was the opposite move: a smaller number of genuinely talented people, paid properly, equipped with the right tools. That combination made an enormous difference, and it is the part founders arriving today have correct. Headcount stopped being the proxy for capability.

Then come the rescues

The other half of my year has been spent going into companies that took the same insight and drew the wrong conclusion from it.

A founder builds the entire product line by vibe coding it. It works, in the sense that there is something on the screen. Underneath, nobody made a single decision about how the data is structured. Somewhere along the way the tooling suggested wiring in a third-party service for authentication, and the founder accepted it, because why not. That recommendation was not wrong in the abstract. It was wrong for their business, and the model suggesting it had no way of knowing that, because context is everything and it did not have any.

The free tier is the same story in miniature. Free is generous right up until the product starts working. Then usage crosses a line, and suddenly you are paying real money for infrastructure you never chose, wired into a product you cannot easily unwire it from. The cheap decision became the expensive one at precisely the moment things started going well.

The mistake underneath all of it

Every one of those rescues had the same root cause. When the founder decided to get help, they hired for code. They brought in someone whose job was to produce software, on the theory that the way to take advantage of AI is to point it at a person who writes a lot of it.

That is the misread. If throughput is the cheap part now, then hiring for throughput is buying the thing you already have. What none of those companies had was someone making architectural decisions with the whole business in view: how the data should be shaped, what should be owned versus rented, which cheap decision becomes expensive at scale, and whether the thing being built is the thing anyone will pay for.

So the shift I am seeing is not really from big teams to small teams. It is from paying someone to code something, to paying someone to be accountable for the whole business idea. The first is now close to commoditised. The second never was.

Bootstrap or raise

When a founder asks me which one they should be doing, I do not start with the market or the size of the round. I ask how far they are from their first dollar.

It is the fastest way to find out whether capital is actually the constraint. If the honest answer is that revenue is close, and the only thing in the way is building and asking, then money is not what is missing. If the answer is that revenue is a long way off for a structural reason, that is a different conversation and sometimes a genuine reason to raise.

Getting to the first dollar

The trick I use with bootstrapping founders is simple. Find someone who is genuinely suffering from the problem. Offer them a cheap deal. Then make absolutely sure they pay you, whether that is swiping a card or sending the money across.

Doing it for free is useless. A free pilot tells you someone was willing to accept a gift, which is not information. The point is not the amount. The point is that a payment is the only signal that survives contact with reality.

The fear I hear most often is anchoring: if I price low now, I can never raise it. So say the quiet part out loud. Tell them you are still discovering price, here is what you are comfortable doing it for today, and here is what you believe the natural price will be later. Founders are astonished by how well that lands, because it is simply true.

It is harder with subscription tiers, and it is worth admitting that. If a company has to restructure around your pricing, telling them the price is unsettled is a real cost to them. That is exactly why your first few customers are beta customers, MVP customers. Treat them as such, give them a discount that reflects it, and be clear about what they are getting in exchange. That is a fair trade, not a concession.

When raising is still the right answer

None of this makes raising wrong. It makes it specific. Capital is the correct instrument when something structural sits between you and revenue: inventory or hardware you have to buy before you can sell, regulation you have to clear before you can operate, a market with real winner-takes-most dynamics where someone funded is already moving, or a network that is worth nothing until it hits a threshold.

Those are all reasons the money buys something you cannot reach any other way. Wanting to hire ten people is not one of them, and for most of the founders I meet, it is the only reason on the list.

What the work actually is now

Ten conversations in, the pattern is clear enough to say plainly. Bootstrapping has become the sensible default, not because building got easy, but because the expensive part moved somewhere else. It moved into architecture, into judgement, into knowing which of the model’s confident suggestions to ignore.

That is the work founders are still paying for, and it is the work I do alongside them through Collective Global: building and operating, taking real risk in the ventures we help start, rather than writing a cheque and waiting to see what happens.

I wrote the decision framework behind this piece as a working paper: what a team actually cost then versus now, how far you can climb before you need capital, a runway worksheet you can fill in, and the six situations where raising is genuinely the right tool. Get the Bootstrap-or-Raise Framework →