The AI Deltaplan is the foundation, now for the building blocks

The message from the startup scene is loud and clear: we embrace the Deltaplan, but we also see where it falls short. We need an ecosystem that not only cultivates talent, but gives it the opportunity to grow. And that requires more than just a plan.

The holy trinity: knowledge, money and space

Successful innovation rests on three pillars. The Deltaplan addresses knowledge with plans for AI literacy and talent development. That is welcome. But talent leaves if there are no challenging jobs to be had. And those jobs are created by scale-ups, which in turn depend on the other two pillars.

Money is the weakest link. The figures are well known. Dutch pension funds invest three times as much in American companies as in European ones. [2] Dutch investors account for only 15% of total venture capital raised domestically. [1] There is a 'forgotten valley' in the funding chain, the critical phase between €1 million and €10 million where many scale-ups get stuck. InvestNL is a fine instrument, but with ticket sizes of €10 to €30 million it is the A380 flying straight over that valley. [4]

Space is the third, often underestimated, pillar, both physical and regulatory. The current grid crisis is the most painful illustration of this. A company in the Brainport region that has to wait until 2033 for a grid connection is not being held back by the market, but by a lack of physical and policy space. [3]

The way forward: four concrete building blocks

The Deltaplan is the foundation. Now we need to supply the building blocks. Here are four concrete proposals that we consistently hear in the field.

1. Make risk capital cheaper and more accessible. The government can play a crucial role by reducing risk for private investors. Expand existing guarantee schemes such as the BMKB and the Groeifaciliteit, specifically targeting the 'forgotten valley' of scale-ups. [5] Have the government act more frequently as guarantor for banks, so that they are more willing to provide financing. This is not a subsidy, it is risk-sharing.

1. Reward success and encourage reinvestment. Make it fiscally more attractive to invest and to reward success. Rather than raising the tax on carried interest from 33% to 49.5%, a measure that directly damages our competitive position, we should be encouraging it. [6] Create a tax incentive for founders who, after a successful exit, reinvest their capital in the Dutch ecosystem. This creates a flywheel of capital and experience.

1. Put founders in the spotlight. We need a cultural shift. We are too modest about our successes. The government and media should present our successful entrepreneurs far more prominently as role models. The stories of the founders of Adyen, Mollie and MessageBird should become as well known as those of the CEOs of our multinationals. They are living proof that it can be done, and they inspire the next generation.

1. Communicate what we do well. We have one of the best digital infrastructures in the world, a highly educated population and an open economy. Let us tell that story more actively. A positive and confident business climate attracts international talent and capital.

From foundation to structure

The AI Deltaplan is the perfect starting point. It is an acknowledgement that as a country we need to take the initiative. But let us not stop at a foundation. Let us now provide the building blocks that our entrepreneurs need to build. By sharing risks, rewarding success and celebrating our champions. Only then will we build an ecosystem of world-class players on that foundation.

The National AI Deltaplan examined

On 23 November 2025, the National AI Deltaplan was presented in The Hague. It reads as an urgent call to action, a blueprint to transform the Netherlands from a passive AI consumer into a strategic European player. Written by more than fifty experts from academia, business and government, commissioned by outgoing minister Vincent Karremans (VVD), the plan maps out a route to reduce Dutch dependence on American and Chinese tech giants. This article analyses the plan's key points, reactions from the ecosystem, the feasibility of its ambitions and the critical choices facing the Netherlands.

The urgency of a Deltaplan

The comparison with the Deltaworks, the flood defences built after the 1953 North Sea flood disaster, is deliberate and strategic. The authors of the AI Deltaplan see a comparable existential threat, not from water, but from data and algorithms. "The Netherlands is barely participating in the development of AI. We are heavily dependent on American and Chinese AI companies", reads the alarming opening of the plan. [7] This dependence, the experts argue, threatens not only the Netherlands' earning capacity, but also its grip on the systems that shape society and democracy.

The plan describes a "historic moment of choice". The choice is between accepting the role of digital colony or investing in sovereignty, prosperity and resilience. "Doing nothing means choosing dependence; investing strategically means working towards our autonomy", the summary states. [7]

This sense of urgency did not emerge from nowhere. It is the product of growing unease within the Dutch tech community. Jelle Prins, co-initiator and founder of AI company Cradle, articulates this feeling aptly. "If we do nothing, our earning and competitive capacity will decline and we will become too dependent on Chinese and American companies that develop AI models. They control access and pricing. That makes us very vulnerable." [11]

How the plan came about

The seed for the Deltaplan was planted in early 2025 with an opinion piece in NRC by Jelle Prins and Michiel Bakker, a Dutch AI lecturer at the prestigious Massachusetts Institute of Technology (MIT). The piece, in which they expressed their concerns about the Netherlands' future in the AI era, gained considerable traction. "The opinion piece got a great deal of attention. Many organisations and institutions shared those concerns", Prins writes on LinkedIn. [8]

Things began to move when outgoing Minister of Economic Affairs Vincent Karremans recognised the urgency. He invited Prins and Bakker to lead the drafting of a more broadly supported plan. What followed was an intensive process of several months, involving more than sixty experts from all corners of society, from scientists and entrepreneurs to representatives from the media, education and the energy sector. "In the end, around sixty companies and experts were involved, from people in education and science to those in the energy and infrastructure sector, from ASML to startups", Prins said. [11]

The result is a document that presents not only a technical vision but also a societal one. It is an attempt to formulate a national, broadly supported response to one of the greatest challenges of our time.

The five pillars in detail

The Deltaplan is structured around five pillars, with a total of 52 concrete recommendations. Together, these pillars form an integrated strategy addressing both the technological foundations and the societal embedding of AI.

Pillar 1: Infrastructure The most concrete and perhaps most urgent recommendation is the need to invest in domestic computing capacity. The figures are sobering. The Netherlands' fastest supercomputer, Snellius, has 640 advanced GPUs. The new AI Factory in Groningen will soon add a further 2,500 GPUs. For comparison, OpenAI, the company behind ChatGPT, is estimated to own more than one million of these chips. [9] The plan therefore advocates accelerating permit procedures for data centres through dedicated "AI compute zones" and resolving the bottlenecks on the overloaded energy grid.

Pillar 2: AI adoption and literacy Technology alone is not enough. The plan emphasises that the Netherlands is lagging in AI adoption, driven by scepticism and a lack of knowledge. A large-scale upskilling drive is therefore proposed, targeting teachers, civil servants, politicians and employees in the private sector. The aim is to establish a baseline level of AI literacy across society.

Pillar 3: A competitive AI ecosystem The Netherlands trains a great deal of top talent, but loses a significant share of it abroad. The Deltaplan seeks to reverse this by improving the business environment. One of the most contentious proposals is relaxing employment protection legislation for jobs with very high salaries. This is intended to reduce the risk for startups when recruiting expensive international talent. [8] The plan also advocates "special economic zones" where innovations such as self-driving cars and delivery robots can be trialled.

Pillar 4: Democratic embedding To ensure that AI serves society, and not the other way around, the plan proposes establishing a National AI Impact Institute. This independent body would monitor the societal effects of AI and advise parliament. Citizens' assemblies are also proposed to give citizens a voice in AI policy.

Pillar 5: Flagships The most ambitious proposals are the "flagships". The establishment of a National Agency for Disruptive Innovation (NADI), inspired by the American DARPA, would fund technological breakthroughs. A Dutch ELLIS Institute would attract leading researchers. And a State Secretary for AI would provide political drive and coordination.

Reactions from the ecosystem

The reception of the Deltaplan is a mix of enthusiasm and caution. Minister Karremans responded positively, telling NRC that the plan contained "very good, concrete points that the government can act on". He expressed support for the idea of a State Secretary for AI. [9]

Endorsement has also come from the tech community. Erick Webbe of KickstartAI describes the plan as "essential if the Netherlands wants to become a frontrunner in AI". [10] Constantijn van Oranje, Special Envoy at Techleap, warns that the pace of AI development is moving faster than the Netherlands can keep up with, but also sees opportunities. [10]

There are, however, critical voices as well. AI expert Aaron Mirck is positive about the initiative but raises a number of concerns. "Work on autonomy is indeed being done through the development of GPT-NL, but this Dutch-language AI model is not mentioned anywhere in the document. I find that strange", he tells Businesswise. [11] Mirck also cautions against an overly technocratic perspective. "At times the document reads as though it was written primarily by technologists. Less bureaucracy can be more meaningful than applying AI." [11]

The most fundamental criticism concerns feasibility. "You can have all the plans you like for building data centres and other AI infrastructure, but if you are not permitted to build, plans remain just plans", says Mirck, referring to the nitrogen crisis and the congested electricity grid. [11]

The 2035 vision

The Deltaplan sketches an appealing perspective for 2035: a future in which AI systems detect cancer at an early stage, robots address staffing shortages in healthcare, and autonomous machines make sustainable agriculture possible. On the roads, self-driving cars could reduce the number of fatal accidents by ninety percent. [7]

This vision is optimistic, but according to the authors not unrealistic. It serves as a point on the horizon, an image of the societal and economic gains that can be achieved if the Netherlands makes the right choices now. It is an attempt to lift the debate about AI above the fear of job losses and disinformation, and to focus on the enormous potential for progress.

Critical analysis and feasibility

The Deltaplan's ambitions are considerable, but so are the challenges. The central question is financial and political feasibility. While the plan itself does not attach a concrete price tag, the AI Coalitie 4 NL, a collaboration of the Netherlands' largest AI organisations, presented its own investment agenda of €5 billion just days after the Deltaplan was released. [12] It remains unclear whether the political will exists to commit such sums, particularly given that AI barely featured in the most recent election campaign.

A comparison with neighbouring countries underlines the urgency.

The Netherlands will therefore need to invest substantially to keep pace. The biggest hurdle, however, is not financial but rather implementation capacity. The problems surrounding the electricity grid and the nitrogen crisis reveal a deeper issue: a country caught in procedural delays and political deadlock. The success of the Deltaplan depends on the government's ability to remove these obstacles.

Conclusion and outlook

The National AI Deltaplan is an impressive and necessary attempt to wake the Netherlands up. It puts its finger on the sore spot: the growing technological dependence and the absence of a coherent national strategy. The strength of the plan is that it does not focus solely on technology, but on the entire ecosystem, from education and regulation to societal embedding.

The coming months are critical. It is now up to the political establishment to take up the challenge and translate the recommendations into concrete policy. The initiators have done their work. The ball is now in the new cabinet's court. The stakes are high. This is not just about economic prosperity, but about whether the Netherlands wishes to remain a sovereign nation in the twenty-first century.