Europe's Quietest Startup Incubator Is 736 Years Old and Has Already Built Three Unicorns.
By Creatives Takeover · July 4, 2026
Europe's oldest startup ecosystem.
In 1991, the University of Coimbra spun up a technology transfer organization called Instituto Pedro Nunes, or IPN. No press tour. No hype cycle. Just a mandate to move research out of university labs and into companies.
IPN works across three fronts at once: applied research labs, business incubation and acceleration, and specialized training in entrepreneurship and technology management. Each research lab is led scientifically by a University of Coimbra professor, while IPN's own team runs daily operations. That structure means a startup inside IPN gets access to business support and to the university's engineering and research infrastructure, under one roof.
Over more than three decades, IPN has supported the creation and development of over 500 startups. Three of them reached unicorn status. More than 30 became scaleups. One internal tally from 2021 found the incubator's companies returned eleven euros in taxes for every euro the region spent on R&D subsidies.
Three unicorns, one address
Feedzai is the clearest case. It started inside the University of Coimbra's Department of Informatics Engineering, where Pedro Bizarro was a professor and Nuno Sebastião was his student. Sebastião brought simulation experience from the European Space Agency, Bizarro brought a doctorate in real-time big data processing, and Paulo Marques brought distributed systems expertise. In 2011 the three of them founded Feedzai out of that hallway, incubated inside IPN. By March 2021 it had crossed a $1 billion valuation. By October 2025, the European Central Bank selected Feedzai as the fraud and risk management provider for the digital euro. In 2024 alone, the company processed roughly $8 trillion in transactions.
Talkdesk took a different path in. Tiago Paiva and Cristina Fonseca, both engineers from Instituto Superior Técnico in Lisbon, won a Twilio hackathon in 2011 and built a cloud call center product from it. They went through the US accelerator 500 Startups, not IPN. But as the company scaled, it built its TDX research labs in Coimbra and Aveiro and partnered with the University of Coimbra's engineering department to launch the TDX-UC Data Science Academy, 240 hours of applied training taught by university researchers and Talkdesk engineers together. Talkdesk hit unicorn status in October 2018 at a $1.2 billion valuation. Today it's valued at roughly $10 billion.
Sword Health rounds out the three. IPN's own account describes its role as supporting the "birth and/or development" of Feedzai, Talkdesk, and Sword Health, three companies that took different routes into the same regional ecosystem before all reaching billion dollar valuations.
The pattern is older than any single unicorn
This didn't start in 2011. In 1998, Critical Software was incubated at IPN and landed NASA as one of its first clients, twenty years before Feedzai. In 2013, the European Space Agency invited IPN to co-run an incubation program applying space technology to problems on Earth, agriculture, water systems, industrial monitoring, medical devices. That program has since supported more than 60 companies.
The pattern is still active today. Two current IPN startups, Sensing Future and FiberSight, are applying the same model to balance rehabilitation devices and fiber optic infrastructure monitoring. Neither is chasing a headline. Both are working the same way Feedzai's founders did: research problem first, company second.
What founders should actually take from this
Most founder advice about "ecosystems" focuses on the wrong variable. People copy the format, twelve week cohorts, demo days, pitch competitions, and skip the part that actually compounds.
Here's what's different about Coimbra's model, and what it means if you're building outside a major hub.
Proximity beats geography. Feedzai's founders didn't fly to a conference to find cofounders. One was teaching the other down the hall. If your best collaborators are already two doors away inside a lab, a university department, or a former employer, that's a stronger signal than a founder-matching event ever will be.
Depth beats breadth in a small market. IPN never tried to become Lisbon or Porto. It built a narrow, deep specialization around applied engineering, aerospace, and data science, and let that specialization compound for three decades. A regional platform doesn't need to cover every vertical to produce outsized outcomes. It needs one thing it's genuinely better at than anywhere else.
Infrastructure outlasts hype cycles. Twelve week accelerator formats optimize for a good demo day. IPN optimized for a permanent connection between where research happens and where companies get built, one that kept producing unicorns across three different decades and three completely different industries: fraud detection, cloud communications, and digital health.
The lesson isn't "go build a university." It's that the founders who win long term are usually the ones who built something durable before it was fashionable, not the ones who moved fastest once everyone else noticed.