You have an idea you have actually tested, maybe a few hundred users on something held together with tape, and no way to build the real thing. Ask three people what to do and you get three answers: find a technical cofounder, hire an agency, or keep going with AI tools. Each answer arrives with total confidence, and each comes from someone whose own interests point that way.
So here is the version with the arithmetic in it. A technical cofounder costs equity and buys you a partner. An agency costs cash and buys you a delivery. Which is right depends far less on your budget than on whether the hardest problem in your business is technical. If it is, no amount of paid delivery replaces someone who owns the outcome. If it is not, giving away a third of your company to get a first version built is the most expensive way to buy something you could have bought with money.
Technical cofounder, agency, or neither: the three routes and what each really costs
| Technical cofounder | Agency or studio | Paid developer or fractional CTO | |
|---|---|---|---|
| What you give up | Equity, and a share of every decision. Typically 40 to 50 per cent if they join before there is a product, 15 to 30 per cent if they join after an MVP exists. | Cash. Illustratively $25,000 to $150,000 for a first version, depending on how much has to exist behind it. | Cash, usually less: an hourly or monthly rate, sometimes with a small slice of equity, commonly 1 to 5 per cent for a fractional CTO. |
| What you get | A partner who owns the outcome, makes technical decisions at eleven at night without a change request, and can hire the next engineer. | A team assembled this month: design, backend, mobile and testing, working to an agreed scope with a knowable cost and date. | Hands on the work and, with a fractional CTO, judgement about what to build and what to refuse. No partner, no marriage. |
| Fits when | The product is the technology, the problem is genuinely hard, and you need someone for four years rather than four months. | You can describe what to build, the hard part is execution and speed, and you would rather keep the company. | You know roughly what you need, the scope is modest, and you want to keep both the equity and the flexibility. |
| The risk | The wrong person, vested. This is the most expensive mistake on the list and the hardest to undo. | You own code and documents, not judgement. When they leave, the thinking leaves with them unless you insisted otherwise. | Continuity. One person gets ill, takes a job, or moves on, and there is no bench behind them. |
Most founders treat this as a binary between the first two columns. The third is the one that fits more situations than either, and the one nobody on the internet argues about.
One more thing about that third column, because it is the option founders discover last. A great deal of what people imagine needs a cofounder is a defined build: a portal with bookings and payments, a first version of a web app or SaaS product, an internal tool that replaces a spreadsheet. Those are scoped, quoted and finished. Reaching for a partner because the work feels large, when the work is merely specified, is how a quarter of a company gets spent on something an invoice would have covered.
The arithmetic almost nobody does
Equity is the only money you can spend before you have any, which is exactly why it gets spent carelessly. Do this calculation before you make an offer.
Suppose you give a technical cofounder 25 per cent to build the first version. If the company later raises at a ten million dollar valuation, that quarter is worth two and a half million dollars of ownership. Building the same first version with cash might have cost you somewhere between twenty-five and sixty thousand. Those two numbers are not close, and founders who have been through it once tend to be visibly careful the second time.
That comparison is also unfair, and it is worth saying how. A cofounder is not buying you one build; they are buying you years of technical leadership, the ability to hire engineers who will not join a company with no technical founder, and someone whose worst night is also your worst night. Some investors will not fund a solo non-technical founder at all. If any of that describes your situation, the equity is not overpriced, it is the price.
The mistake is not choosing equity. The mistake is paying partner prices for contractor work: handing over a quarter of the company for what turns out, in hindsight, to have been three months of building and then a slow fade.
What a cofounder gives you that an agency cannot
- Ownership of the outcome, not the deliverable. An agency is accountable for what was agreed. A cofounder is accountable for whether it works, which is a different and much larger thing.
- Decisions without a change request. The hundred small technical judgements a week that nobody would think to write into a scope, made by someone who is not billing for them.
- Judgement about what not to build. The most valuable thing a good technical partner does is refuse things, and refusal is structurally harder for a supplier being paid to build.
- Recruiting. Engineers join engineers. Your first three technical hires are dramatically easier with one, and dramatically harder without.
- Standing with investors. Some funds will simply not lead a round with no technical founder. If your plan requires raising, find out early whether that applies to the funds you are targeting.
What an agency gives you that a cofounder cannot
- Speed, starting now. Finding the right cofounder commonly takes six months or more, and the wrong one costs a year. A studio starts in a fortnight.
- A whole bench. Design, backend, mobile, infrastructure and testing, in the proportions your project needs this month rather than the ones a single person happens to have.
- A cost you can know in advance, tied to milestones you can stop at. Equity has no ceiling; a contract does.
- All of your company. You keep every share, which matters most in exactly the case founders underweight: the company that does fine rather than the company that is enormous.
- No breakup. Ending an agency relationship is an email and a handover. Ending a cofounder relationship is a lawyer, a renegotiation, and often the end of a friendship.
The situation almost nobody planned for: you already built it
A large share of founders now arrive at this decision with something already working, or nearly working, built with an AI tool over a few intense months. That changes the question. You are no longer asking who will build it. You are asking who will finish it, secure it and take responsibility for it, which is a smaller job on paper and a more specific skill in practice.
Be careful with the estimate in your head. The parts an AI tool builds fastest are the visible ones, and the parts it leaves out are accounts, permissions, data structure, security and everything that happens when something fails. That is why these projects so often feel eighty per cent finished and stay there. Our guide to making an AI-built app production-ready sets out what is usually missing and what it costs to close.
Keeping ownership, whichever route you take
- Assign the intellectual property in writing, to the company. Paying an invoice does not by itself transfer copyright, and neither does a handshake with a cofounder. Both need a signed assignment, and both need it before there is anything worth arguing about.
- Accounts in the company's name from day one. The domain, the repository, the hosting, the cloud account, the app store listings. Add people as collaborators you can remove.
- The repository lives in your organisation, not in a supplier's or a cofounder's personal account, with you as owner.
- Pay agencies against milestones, with the final one tied to a documented handover rather than to a date.
- Watch for the tell. If a candidate cofounder is asking the questions an employee would ask, about hours, salary, scope and holiday, they are applying for a job. That is perfectly reasonable, and the right response is to price it as a job rather than as a partnership.
A decision test you can run this week
- Is your hardest problem technical, or is it distribution? Be honest. Most software businesses fail at the second and hire for the first.
- Could you write down what to build, well enough for someone to quote it? If yes, you have a project. If no, you have a research problem, and research problems want a partner rather than a supplier.
- Do you need someone for four years or four months? Say the answer out loud before you say it to a candidate.
- Do you have cash? Ten to sixty thousand changes this decision completely. Nothing changes it back.
- Would you hire this person if they asked for a salary instead? If the answer is no, do not give them equity either.
- If they walked in six months, what would you be left with? Answer for each route. The answers differ more than you expect, and they are the whole decision.
Not for you if
The decision you are actually making
You are choosing between a partner and a purchase, and the honest version is that most founders reach for a partner because a purchase feels frightening when the money is real and the equity is not. Run the arithmetic. Run the six questions. Then choose deliberately, knowing that a good agency will hand you code and documents, a good cofounder will hand you years, and the wrong choice in either direction costs about the same: a year, which is the one thing neither of them can sell you.