Founders & Startups
The doctor building AI and the Dutch quest for venture capital
Nov 3, 2025


In 2024, almost half of the Dutch healthcare workers will spend a quarter of their working time on administration. One in five will spend even half of their time on it [1]. This figure highlights the dissatisfaction with the increasing work pressure among medical specialists (30% in 2024) [2]. This is not an abstract issue. This is the reality of one of the best healthcare systems in the world, which is straining under its own weight.
Anesthesiologist and programmer Michel Abdel Malek has experienced this reality firsthand for many years. He witnessed the focus shifting from the patient to the system. From this frustration, his startup Delphyr was born. It's a story often told, but the angle here is different. This is not a tech entrepreneur spotting an opportunity in healthcare. This is a medical specialist building a technological solution for his own, deeply rooted problem.
The search function that reinvents the EHR
Delphyr is developing an AI assistant that supports doctors and nurses with generative search, a search function that not only finds information but also understands and summarizes it. The application listens in during a consultation, transforms the conversation into a structured note, and directly unlocks all relevant medical data from the often outdated and fragmented electronic health records (EHRs). This generative search acts as a springboard for future AI applications in healthcare. The promise is a drastic reduction in administrative burden, but behind this promise lies a series of strategic choices that make the story truly interesting.
The Sovereign Choice
The first choice is technological. Delphyr is building its own language model, specifically trained for Dutch healthcare. Even more important is the infrastructure. The company partners with Amsterdam-based Nebius as the primary cloud partner, but is in regular talks with both Microsoft and Google to ensure the best solution. This may appear to be a technical detail, but it is a fundamental strategic move. In a sector where data security and privacy are sacred, this flexible yet deliberate approach is a direct response to concerns about data sovereignty. By opting for a predominantly European stack, Delphyr is building a 'sovereign' solution. It is a selling point that carries significant weight with hospital executives, who bear ultimate responsibility for patient data.
Raising the Medical Bar
Where many healthcare startups choose the fast route, Delphyr has made a different strategic decision: pursuing rigorous medical certification. It is a deliberate choice that distinguishes the company. This certification is time-consuming and costly, but it provides hospitals with the assurance that the product meets the highest medical safety standards. For a sector that is rightly cautious with new technology, especially when it concerns patient safety, this is not a nice-to-have but a must-have.
Strategic Pillar | Delphyr's Approach |
Founder's Profile | An insider (doctor) solving the problem from within, not an outsider. |
Technology | A proprietary language model for generative search, specific to the complexity and nuances of Dutch healthcare. |
Infrastructure | Strategic partnerships with European and international cloud providers (Nebius, Microsoft, Google) for optimal data sovereignty. |
Certification | A deliberate choice for rigorous medical certification to build trust. |
Market Approach | Collaborates with existing EHR monopolists instead of trying to replace them. |
Navigating in a Land of Giants
The next strategic choice concerns the market. The Dutch hospital market for EHRs is a triopoly. Three companies, ChipSoft, Hix, and Nexus, together control 97% of the market [4]. A new player looking to replace these systems stands no chance. Delphyr is not attempting this. Instead, it positions itself as an intelligent layer that operates on top of these rigid systems. By collaborating with seven EHR suppliers, they aim not to storm the gatekeepers of health data, but to keep them as allies. It is a pragmatic approach that seems to be the only way forward in such a consolidated market.
The Paradox of Dutch Capital
Despite a promising start with €500,000 in seed capital and support from investors like Thomas Wolf (co-founder of Hugging Face), Abdel Malek exposes a painful paradox in the Dutch innovation climate.
"There is a saying among Dutch founders: we have enough capital, but what we lack is venture capital,"
according to Abdel Malek.
The figures confirm his observation. In 2024, investment in venture capital per capita in the Netherlands was €193, far behind the €546 in the United States [5]. Dutch pension funds, which manage vast amounts of capital, invest only 0.01% of it in domestic venture capital [5]. Investors are cautious in the early, high-risk phase where deep-tech companies like Delphyr find themselves. They wait for proven success, but it is precisely the lack of early funding that makes achieving success challenging.
What to Watch For
Delphyr's story is more than a startup narrative. It is a case study for the future of technological innovation in the Netherlands. The opportunity for Delphyr is clear: if they succeed in being among the first to lay a secure, sovereign, and effective AI layer over the existing healthcare infrastructure, the potential market is enormous. The threat is twofold: the power of EHR suppliers who can shut their doors, and the risk-averse investment climate that may force the company to temper its growth ambitions or seek expansion abroad.
We are thus particularly interested in the next steps. Will Delphyr succeed in closing the current funding round and attracting the necessary 'venture capital'? And perhaps even more importantly: will they manage to convince the first major hospitals that their sovereign AI solution not only eases administration but genuinely improves healthcare? The answers to those questions will not only determine Delphyr's future but also reveal something about whether the Netherlands can solve its own most complex problems with its own most advanced technology.
Willem Blom
Founder Dutchstartup.AI
References
[1] FNV Zorg & Welzijn (2024). FNV-onderzoek naar administratiedruk in Zorg & Welzijn. https://www.fnv.nl/nieuwsbericht/sectornieuws/zorg-welzijn/2024/06/pijn-meten-bij-patienten-zonder-pijn-en-alles-van
[2] Skipr (2024). Administratieve lasten leiden tot toenemende druk op specialisten. https://www.skipr.nl/nieuws/administratieve-lasten-leiden-tot-toenemende-druk-op-specialisten/
[3] Nebius (2025). Nebius introduces Nebius AI Cloud 3.0 “Aether”. https://nebius.com/newsroom/nebius-introduces-nebius-ai-cloud-3-0-aether-delivering-enterprise-grade-security-compliance-and-control-for-ai-deployment-at-scale
[4] Consultancy.nl (2024). EPD-markt voor ziekenhuizen in handen van drie partijen. https://www.consultancy.nl/nieuws/51326/epd-markt-voor-ziekenhuizen-in-handen-van-drie-partijen
[5] IO+ (2024). Dutch startups’ toughest hurdle: winning venture capital. https://ioplus.nl/en/posts/dutch-startups-toughest-hurdle-winning-venture-capital


Platform Launch: January 2026
Secure your place in the Dutch AI ecosystem now
Dutch AI
Built Different
An initiative by Willem Blom & Max Pinas | Powered by Studio Hyra
Dutch AI. Built Different 2025
Dutch AI
Built Different
An initiative by Willem Blom & Max Pinas
Powered by Studio Hyra
Dutch AI. Built Different 2025
Dutch AI
Built Different
An initiative by Willem Blom & Max Pinas | Powered by Studio Hyra
Dutch AI. Built Different 2025



