Companies across STATION F's ecosystem raised a record €2 billion in 2026, and the Paris campus has named its eighth Future 40 cohort, a list of 40 startups drawn from more than 1,000 applicants. The headline number is the money. The more telling detail sits in the cohort itself: six of the 40 companies went through Y Combinator and then returned to France to build.
That pattern, according to Tech.eu, reflects a growing willingness among founders to build global businesses in Europe after time in Silicon Valley. For years the default path was simple: get into a top US accelerator, stay in San Francisco, raise there. If that default is genuinely loosening, it changes what the next generation of startups looks like and where they are built. Here is what the numbers actually support, and what they do not.
What do the 2026 numbers actually show?
The €2 billion raised across the wider STATION F ecosystem in 2026 surpasses the more than €1.5 billion raised in 2025. Two alumni rounds did much of the work: Multiverse Computing raised €500 million and Alan raised €480 million. Strip those out and the year looks less like a surge and more like a steady climb, carried by many smaller rounds.
The Future 40 cohort itself is early-stage. Of the 40 selected companies, 31 are at pre-seed or seed stage and two are at Series A. Together they have raised €110 million to date, one of the better-funded editions. Across all eight editions, Future 40 companies have collectively raised €1.2 billion. The selection spans AI, deeptech, developer infrastructure, robotics and climate tech, with sovereign AI and physical AI named as notable themes. Notable rounds in the cohort include Arlequin AI at €28 million, CURA at $10 million, Cleavr at €8 million and Lemrock at €7 million.
Why are some founders skipping venture money entirely?
Five companies in this year's Future 40 are bootstrapped, and Camille Aubery, an investor at STATION F, ties that to a changed funding market. In 2021 and 2022, she told Tech.eu, startups could sometimes raise money with little more than an ambitious idea and a short pitch deck. That is no longer the case. But she also points to a second force: a growing conviction among some founders that they want to retain control of their businesses.
AI tools are the practical enabler. Oguste reached €400,000 in annual recurring revenue in under a year while remaining bootstrapped. Revox used an AI agent to book 250 calls with prospective customers in a single month, work that previously might have required several sales representatives. The mechanics are straightforward: if software does the outbound selling, a four-person company can produce the revenue motion that once needed a sales team. Founders can delay fundraising, or avoid it altogether. Two of the five bootstrapped companies are now raising, while others have chosen to stay independent.
The implication cuts both ways. Smaller teams reaching real revenue means fewer forced fundraises, which is healthy for founders. It also means the next cohort of breakout companies may be less visible in funding databases, because the signal arrives later.
What is physical AI, and why is it showing up in Paris?
Physical AI is the shorthand for robotics and hardware that combine physical products with AI models. This year's cohort includes Bleu Robotics, which develops AI models for industrial humanoid robots; Allude, which builds infrastructure for AI agents and robots in factories; and AICE, which develops low-cost subsea drone swarms for defence and critical infrastructure.
The presence of defence and industrial customers matters. Software startups can sell to anyone from a laptop; hardware startups need factories, certifications and buyers with budgets. STATION F supports this by bringing industry partners on campus. In healthtech, a partnership with Sanofi connects startups in AI-driven drug discovery, healthcare data and biological modelling with pharmaceutical R&D expertise and proof-of-concept opportunities. The Care Forward programme, launched with Doctolib, AP-HP and Roche, offers access to clinical environments, regulatory expertise and healthcare professionals.
The cohort's composition also signals a talent shift. Forty-seven per cent of the teams have at least one PhD, and 75 per cent are repeat founders. Aubery reads that as the profile of supported companies changing: more technically specialised founders and researchers on campus.
Are founders really coming back from Silicon Valley?
Six Future 40 companies did Y Combinator and returned to France. Aubery's framing is that founders now use the US ecosystem for networks and opportunities, then build at home. Her stated reasons: it is cheaper to build in France, and a global company does not require a US base.
The honest caveat is capital. France has a strong early-stage funding ecosystem, but late-stage capital remains thin, there and across much of Europe. Founders can start in Paris and still face a decision when a company needs a large growth round. The return pattern is real but early; six companies in one cohort is a signal, not a settled trend. Readers tracking European funding headlines can use our guide to France Startup News: How to Read the Headlines Like an Insider to separate round sizes from round significance. Readers following this should also see France Startup News: How to Read the Headlines Like an Insider.
How does STATION F pick the Future 40?
Aubery calls selection the one thing the campus does not want AI to do. The criteria are not fully objective: revenue, customers, fundraising, traction and milestones are considered, but the campus says it is assessing a startup's potential impact on its industry, the importance of its innovation, and whether the founders are the right people to build the business. It looks for founders with global ambitions rather than domestic-only plans. Every application is reviewed twice, because some founders present better than others. Startups typically stay on campus 18 months to two years. This connects to our earlier piece, Japan Startup News Today: How to Read the Headlines Without Getting Burned.
For a campus founded by Xavier Niel in 2017, now supporting more than 1,000 startups annually across 40 programmes, with around 9,000 startups having passed through and founders from 60 nationalities on site, the selection process is also a statement about what the campus values. The 40 chosen companies represent roughly four per cent of the 2026 community.
The open question is whether the record year converts into later-stage companies that stay in Europe. The early-stage money and the returning founders are documented. Whether the growth capital shows up to keep them is not yet.
