Two years bootstrapped:
what we learned, what we wasted.

Drape turned two years old last month. We are now one hundred and eighty-four paying ateliers, twelve employees, and quietly profitable since month four. We have not raised a euro of external capital. We are not in any rush to.

I wrote this post for the founder reading our investor letter and wondering whether bootstrapping fashion-tech is actually possible. The honest answer is yes, but it is also the answer that comes attached to a long list of things we wasted money and time on, that we did not need to.

What we wasted

€18,000 on one rented GPU cluster decision. In month seven we ran the math on owning vs renting GPU. We talked ourselves into a 36-month reserved instance on AWS Frankfurt because the discount looked good and the spreadsheet said yes. Three months later, H100s became available on Hetzner at a fraction of the price, and we were locked into AWS reserved instances we could not exit. We paid out the contract and ate the loss. The lesson: never sign a 36-month commitment in a market where the underlying compute is halving in price every nine months.

€11,000 on a brand agency in Berlin. We hired them in month three to "develop our brand identity." Six weeks later they sent us a deck with three logo concepts, none of which we used, and a tone-of-voice document that suggested we should be "playful and disruptive." We are neither. We refunded ourselves by canceling halfway through and built the identity in-house in a weekend. The Drape mark you see today was drawn by Sirin on a flight to Lisbon.

€7,400 on a marketing automation suite. We did not need it. We had 23 customers at the time. The automation suite was sold to us with the implicit promise that it would scale us to 1,000. It would not have. Marketing automation at our scale is one Postgres table and a cron job. We use a Postgres table and a cron job today.

Most tools you buy in year one are not solving a problem you have. They are solving a problem you fear having in year three.

€4,200 on a fractional CFO. He was helpful for exactly one conversation — the one where he told us "you are profitable, stop hiring me." Honest, useful, exactly four billable hours.

€34,000 worth of equity to a "growth advisor" who gave us a deck of "growth tactics" we did not use and introduced us to one customer we lost in three months. We have since bought back the equity for €1. We learned that advisors with portfolios of forty companies are advisors of forty companies, which is to say, advisors of none.

What we got right

We charged from day one. The first version of Drape was barely walking — generated four frames per minute, no Shopify integration, no model library, no brand kits. We still charged €19/month for it. The first 12 customers paid happily. The act of charging from day one forced us to write the actual roadmap that customers wanted, instead of the roadmap we wanted.

We hired slowly and senior. Twelve employees in twenty-four months. Average tenure-in-industry at hire: 11 years. We did not hire a single junior in year one. Year two we hired our first associate-level engineer — and only because Daria, our head of platform, vouched for her personally. Slow, senior, deliberate hiring is the single largest contributor to our being able to ship at the velocity we ship.

We wrote the back office before we needed it. Our second engineering hire built an internal admin tool that handles refunds, customer-specific GPU priority bumps, model rights ledger lookups, and revenue dashboards. It took her three weeks. It has saved us hundreds of hours since. Most bootstrapped companies skip this step because it is unglamorous. It is the single best ROI you can get on engineering time in year one.

We refused investor meetings until we were profitable. Three different funds reached out in our first nine months. We declined all three. The reason: a profitable founder negotiating from strength gets a meaningfully better valuation, term sheet, and board composition than a desperate founder negotiating from runway. Whether or not we ever raise, the optionality is worth it.

Bootstrapping is not about avoiding capital. It is about being the one who decides when, from whom, and on what terms.

What I would do differently

Hire the customer engineer earlier. We waited until month sixteen to hire someone whose full-time job was the first thirty days of every new atelier. By the time we hired her, we had churned about €40,000 of ARR from ateliers who never quite got over the activation cliff. That €40k would have paid for ten months of her salary.

Spend more on the front door. We undertstaffed the marketing site for two years. The current version of drape.studio that you are reading went live in June 2026 — version five. The first four versions were each made in a single weekend by Andrei. They were fine. They were not great. The brand strength we built in year one would have compounded much faster with a great front door.

Open an office in London earlier. We waited eighteen months to open London. We should have done it in month four. The partnerships flywheel that London has unlocked — agencies, luxury houses, retailers — was always going to require boots-on-ground in a financial center. Milano is for the art. Paris is for the engineering. London is for the business. You need all three.

The numbers, anonymized

Month one to month four, we burned about €11,000 — mostly on the founders' minimum salaries and AWS credits running out. Month four we crossed €4,200 in monthly recurring revenue and have been profitable every month since.

Month twelve: €78,000 MRR, profitable, 6 people.

Month twenty-four (last month): €201,000 MRR, profitable, 12 people, no debt, six-month operating runway sitting in a Cypriot bank, no investor decks circulating.

Our compound monthly growth rate is 19.4%, which is not the kind of number a venture pitch shows on a hockey stick, but it is the kind of number that means we will be at €1 million MRR sometime in late 2027 if we keep doing what we are doing. Which we plan to.

A note to anyone bootstrapping in vertical AI right now

The temptation to raise — especially in AI, where every conversation in 2026 is about funding — is constant and loud. Ignore it for as long as you can. Build the smallest version of the product that solves a real problem for a real customer who will pay real money. Charge them from day one. Spend that money on building the back office and hiring senior. Refuse the meetings until you have the option to take them on your terms.

You will move slower for the first year. You will move faster forever after.