Fractional CTO services for AI-first startups

My standard rate is $100 an hour. Until December 2026, or until I join a new venture full time, whichever comes first, I am holding it at $50, and the retainers are sized from that: 15 hours for $750, 50 hours for $2,500, 80 hours for $4,000. What those hours are not is a meter. For comparison, a full-time startup CTO in the US costs roughly $260,000 a year fully loaded and before equity. Most fractional engagements in this market run $4,000 to $15,000 a month, almost nobody publishes the number, and almost nobody tells you how many hours you are actually getting.

Because I have the capacity right now and I would rather spend it on companies doing interesting work than hold out for the standard number. It runs until December 2026, or until I join a new venture full time, whichever comes first. If you start inside the window, your rate is locked for the life of the engagement; it does not step up in January because the window closed. When it does close, new engagements go back to $100 an hour and the retainers go back to $1,500, $5,000, and $8,000.

It gets the hours. You will not get a surprise invoice from me for going over, and I have never sent one. If your engagement is running well past its tier month after month then we should talk about moving you up, but that is a conversation between us, not a line item that appears on a bill you did not expect.

Then it is a quiet month. The hours are a commitment I am making to you, not a balance you are spending down, and I am not going to invent work to hit a number. If quiet becomes the pattern rather than the exception, I will tell you to drop to a smaller tier before you think to ask.

I keep track for my own planning and I will happily walk you through where the time went. What I will not do is bill you in six-minute increments or turn our working relationship into an argument about a spreadsheet. If you are counting my hours, or I am counting them at you, something has already gone wrong.

Five areas, and I own all of them: architecture and the technical roadmap, engineering hiring and team structure, security and infrastructure posture, vendor and agency and code review, and the technical half of your investor and board conversations. The difference between this and consulting in the usual sense is accountability. A consultant hands you recommendations and leaves; I sit in your standups, my name is on the decisions, and if the roadmap slips it is my problem as much as yours.

Write down the decision that is actually stuck, in one paragraph, before you write a job description. Check the candidate has shipped what you are asking them to own and still writes code. Agree the hours, the overflow policy, and the first deliverable in writing. Then start with a paid architecture review rather than a long contract, because two weeks of real work tells you more than four interviews. With me that means a 30-minute call, a written assessment in week one, a three-month minimum, and month to month after that.

Same seniority, less of it, no equity dilution, and you can stop. A full-time CTO makes sense once technical leadership is a full-time job, usually post-Series A with a team of eight or more engineers. Before that you are paying a senior salary for a role that genuinely does not need forty hours a week.

A technical co-founder takes equity, commits for years, and shares the risk. A fractional CTO is paid, scoped, and reversible. I have been the technical co-founder six times, so I will give you a straight answer on which one your situation actually calls for, and sometimes it is neither yet.

Yes. Most of my clients are in the US. I hold overlapping hours with US Eastern and Pacific time, invoice in USD, and I have been doing distributed work across three continents for a decade. Time zones are a logistics problem, not a quality problem.

Both, and which one depends on the tier. At advisory level I am reviewing decisions. At Embedded CTO I am building the hardest parts myself. I have never been comfortable directing engineers on things I would not build, and I do not intend to start.

Three months. Anything shorter and I am still learning your codebase when it ends, which is not worth your money. After three months it runs month to month with 30 days notice on either side.

Yes, and it is a common starting point. I audit what exists and give you a written verdict: what is salvageable, what is risky, what needs rebuilding, and roughly what each costs. Sometimes the honest answer is that it is fine and you should keep going.

Partly, and it is negotiated case by case. The split runs from a small equity component up to roughly half equity and half cash, depending on your stage and how long we expect to work together. Early-stage teams tend to want more equity in the mix, which also means I am carrying some of the same risk you are. What I will not do is take equity and then tell you to build something you should not: if I think the plan is wrong, you get that opinion whichever way we are structured.