AI startups outside the Randstad and Brainport: three regions compared

Outside the Randstad and the Brainport region around Eindhoven, several regional AI ecosystems with distinct identities are taking shape in the Netherlands. Twente, Arnhem/Nijmegen and Groningen are each, in their own way, laying the groundwork for AI-driven business activity, but they are doing so from very different starting positions and with unequal resources.

The three regions share a common point of departure: one or more universities or universities of applied sciences that act as magnets for technical talent and as breeding grounds for spinoffs. At the same time, each of the three lacks what Amsterdam and Eindhoven do have: scalable venture capital networks, a dense concentration of corporate partners and direct access to international markets. That tension defines the portrait of each of these ecosystems.

Twente: manufacturing industry as the foundation for AI applications

The University of Twente and the surrounding manufacturing industry form the two pillars on which the Twente ecosystem rests. Startups here typically build AI applications that connect directly to production processes, quality control and industrial automation. This is no coincidence: the region has had a strong association with textile manufacturing for decades, and later with precision technology.

A clear example of that orientation is Rapid Engineering, from the nearby town of Urk. Founded in 2019, the company develops vision software for quality control in food production and raised five million euros in early 2025. It illustrates how AI in this region is applied in a concrete, sector-specific way, with a direct link to physical production environments.

What Twente lacks is a critical mass of later-stage funding rounds. Seed capital is available through regional funds and European subsidy schemes, but Series A and beyond almost always requires founders to connect with investors in Amsterdam or further afield. This pulls the growth phase geographically away from the region.

Arnhem/Nijmegen: life sciences as a specialisation

The region around Arnhem and Nijmegen positions itself more distinctly at the intersection of healthcare and data analysis. Radboud University and Radboudumc are the defining forces here. They attract researchers with expertise in medical imaging, genomics and clinical data science, and that expertise permeates the startup scene.

Saddle Point Science Europe, based in Nijmegen and founded in 2022, works on Bayesian data analysis for more precise treatment outcomes in clinical practice. The company received 17.5 million euros in grants. That grant model is characteristic of the region: the life sciences sector here relies relatively heavily on public funding via ZonMw, NWO or European health programmes, and less on private venture capital.

This has both advantages and disadvantages. Grants offer stability and allow for longer research timelines, but they impose different requirements around accountability and timelines than private investors do. Startups seeking to scale into commercial markets outside healthcare must take an extra step to recalibrate their governance structure and go-to-market approach. Compared with Amsterdam, where Castor EDC raised 45 million euros in private financing as an eClinical platform, the contrast in ecosystem maturity is noticeable.

Groningen: agri-tech and energy as a niche

Groningen operates on a different playing field. The province is explicitly aligning itself with the energy transition and, to a lesser extent, with the agrifood sector that has academic roots through Wageningen. The University of Groningen and Hanze University of Applied Sciences produce a steady stream of graduates, but the translation into a vibrant startup scene remains more tentative than in the other two regions.

Notably, Wageningen, as a knowledge centre for food and agri-tech, radiates influence into surrounding regions. Demeter42, based in Wageningen itself, builds data-driven decision support for greenhouse management and raised four million euros. Although Wageningen is not geographically part of Groningen, the model illustrates how sector knowledge and academic embedding can together create a market position. Groningen is seeking a comparable anchor in the energy sector, but the commercial AI applications flowing from it are still largely in their infancy.

The support structure in Groningen consists primarily of regional development agencies such as NOM and a limited number of incubators. Risk capital from outside the region finds its way to Groningen AI startups less readily than to comparable companies in the Randstad, partly due to geographical distance and the relatively smaller network of business angels.

What connects the three regions, and what sets them apart

All three regions are building their ecosystems on academic knowledge, but their sector profiles differ substantially. Twente is oriented towards manufacturing and production automation, Nijmegen towards medical data analysis and clinical research, and Groningen towards energy and agri-tech. That specialisation is a strength: startups find sector expertise, customers and testing environments there that are harder to organise elsewhere.

The shared vulnerability lies in the funding infrastructure. None of the three regions has its own cluster of growth funds comparable to what is available in Amsterdam or Eindhoven. Seed funding is achievable through regional funds, European programmes or NWO grants, but Series A and above almost always requires a move towards the Randstad or abroad. This means founders must think early about building a network outside their own region, even when their customers and partners are located there.

A second difference from Amsterdam and Eindhoven is the density of corporate partners. Large technology companies and multinationals are more concentrated in the Randstad and Brainport, making pilots, collaborations and procurement processes easier to initiate. Regional startups sometimes bridge that gap through industry associations or via the university as an intermediary, but this costs time and energy that is not needed in more direct ecosystems elsewhere.

For investors and policymakers seeking to strengthen the regional AI scene, the challenge is clear: not so much building more incubators, but bridging the gap to growth financing and corporate networks that sit outside the region. The sector specialisation is already there; the scaling infrastructure is not yet.