In January 2024 a rumour about PhotoRoom ran ahead of the company. TechCrunch reported that the Paris app was raising $50 to $60 million at a valuation of up to $600 million. The founders would not comment. A month later the real figure was out: $43 million at $500 million, a little lower than the whispers, confirmed by CEO Matthieu Rouif.
It was a striking moment for a company only about four years old, with roughly 50 employees.
It also sold access through an API and a web editor, which let other businesses bolt it onto their own tools.
Rouif is the CEO. Eliot Andres, his co-founder, is the CTO. Their schools, hometowns and earlier jobs are not in the interviews we found, so the story has to be told through what they built. What they built was a phone app that did one annoying job well: cut an object out of its background. People selling on online marketplaces, small shops and casual users all needed exactly that.
The product grew by being useful rather than clever. By early 2024 the app had passed 150 million downloads and, the company said, was processing about 5 billion images a year. It also sold access through an API and a web editor, which let other businesses bolt it onto their own tools.
The money followed. A Series A of about $19 million arrived in 2022, with early backers including Kima Ventures, FJ Labs and Meta. Balderton Capital led the 2024 Series B, with Aglaé Ventures joining and Y Combinator participating. Total funding stood near $64 million.
The part founders should notice is what they spent it on. Rather than rent someone else's image model, PhotoRoom said it trained its own from scratch, and claimed its custom design made generation up to 40% faster than rival visual AI tools. Launched alongside the round was Instant Diffusion, a tool for placing products in generated scenes. In May 2025 it bought a smaller company, Generate Banners, to add ad creation.
None of this is a story of collapse or courtroom. It is a story of narrow focus. Competitors are far larger: Adobe, Canva and others have the distribution. PhotoRoom's bet is that a tool built for one painful job can stay ahead of a suite built for everything.
The takeaway for a founder is plain. A single repeated task, done a million times a day, can fund the harder research later.