We have the money. We have the people. We lack the will.

We have the money, we have the people. What we lack is the will. And that is costing my children their future.

AI is not hype. It is a paradigm shift, a rearrangement of how work, knowledge and power are organised. And Europe responds to it the way it responds to everything: regulate, consult, postpone. You fight fire with fire, not with one more directive.

The easy answer is that people don’t get it. That answer is lazy. Exponential development is cognitively unnatural for almost everyone; the citizen has no interface with the curve. The real complacency sits one layer higher, with those who are paid to look ahead.

The hypocrisy in one sentence

Bernardo Kastrup, philosopher, former ASML strategist and founder of chip company Euclyd, said it recently without any hedging about Europe’s position in AI:

“Frankly, we have the money too.”

Read that again. The talent is there. The capabilities are there. The capital is there. And yet that same European capital, from pension funds to institutional investors and family offices, flows en masse to American tech. We finance our own dependence and call it returns. We complain about Big Tech on Monday and buy the shares on Tuesday. That is the hypocrisy: not that we lack the money, but that we would rather hand it to American big capital than to our own builders. I saw that pattern up close with Solvinity: a Dutch cloud company that warned against American dependence and was then itself acquired by an American giant.

Kastrup adds something every executive should frame. Continuing to rely on foreign hardware while we deliberate over software frameworks is not caution, it is complacency. And his timeline is not thirty years. European AI sovereignty in strategic domains is achievable before 2030. If we choose.¹

The Wennink report: the diagnosis is in. Where is the treatment?

In December 2025 Peter Wennink delivered “The route to future prosperity” at the request of the cabinet, the Dutch Draghi. The conclusions are not ambiguous.

The Netherlands has the knowledge, the capital and the willingness to invest, but blocks itself with slow decision-making, grid congestion and procedures. We need 1.5 to 2 percent growth per year to pay for healthcare, defence, pensions and the energy transition; the forecasts remain stuck at 0.5 to 0.9 percent. Europe holds 5 percent of the world’s compute capacity. The United States 74 percent.² In the words of the report itself: “The Netherlands is willing, but it must also be allowed again.”³

That report has been on the table for a good six months now. I see symposia. I see reflection sessions. I see working groups. What I do not see: spades in the ground, mandates, capital allocation. In this country we have developed a culture in which drawing a conclusion is the finish line instead of the starting shot. Talking has become our product.

Groningen: symbolic politics with a price tag

Look at the AI factory in Groningen. Two hundred million euros in total, supercomputer at full power in 2027.⁴ The cabinet calls it “a choice for independence”.²

Industry experts from Nvidia and Vertiv calculated what a real AI factory of 100 megawatts costs: around six billion euros.⁵ We are building for a thirtieth of that and calling it sovereignty. By the time that machine is running, the frontier is three generations further on. This is not infrastructure policy. This is a ribbon to be cut. Symbolic politics gives the feeling of acting without the cost of acting. And the most dangerous part is that it removes the urgency: aren’t we already doing something? How big the Netherlands did dare to think, I described when we submitted SovereignAI Grid for the European Gigafactory tender.

The calculation nobody wants to make

This is where it gets personal, because this is about the earning power of the next generation. About my children, and yours.

In the Netherlands we like to hold debates about redistribution: inheritance tax, wealth tax, basic income. Legitimate debates. But they all rest on one assumption: that there is something to distribute. Whoever wants to pay for a basic income, whoever wants to keep the welfare state standing as the population ages, first needs earning power. And that earning power will be determined over the next ten years by whoever owns the AI infrastructure. Redistributing without building is eating the inheritance. A meritocracy that wants to tax the yield but not build the factory is not a meritocracy. It is living off the Marshall generation that did build it.

Everyone in the World Cup stands this summer understood the lesson that The Hague keeps failing to learn: teams that only defend lose to teams that attack. You cannot pass the ball around for ninety minutes and think you’ve won. Europe is playing catenaccio in a tournament decided on goal difference.

The governance culture that makes this possible

Why do we get away with this passivity? Because for a decade we have been conditioned to believe that “legally defensible” is the same as “right”. The culture of “no active recollection” has taught us that accountability is a procedural game you can win, not a duty you carry. Anyone following the hearings of the Dutch COVID inquiry sees the same pattern: carefully prepared self-justification in which the leading players sell their policy after the fact, while the questions that truly matter stay on the surface.

That is the same disease as AI complacency. A governance culture that legalises mistakes instead of acknowledging them cannot learn. And a system that cannot learn cannot build, because building means making mistakes. Making mistakes is not fraud. But in the Netherlands we treat entrepreneurial risk as suspect behaviour and administrative failure as a legal footnote. As long as that holds, the middle class is dismantled by exactly the layer that ought to protect it: the entrepreneur, the craftsman, the builder.

What needs to happen

Stop stacking reports. Wennink is in. Draghi is in. Kastrup has described the route. The diagnosis phase is over; every new exploration is postponement in disguise.

Allocate capital at scale, not at symbol. Not 200 million for a photo opportunity, but billions through pension funds and institutional capital towards European compute, chips and datacenters, with ownership models that keep control here.

Treat builders as allies. Permits, grid capacity, procurement rules: everything Wennink calls “deferred maintenance” is a choice. Reverse that choice.

Restore the norm that making mistakes is allowed and lying is not. Without that reversal, every transition agenda is dead paper.

The question

The question to every executive, every ministry, every pension fund is simple. If the paradigm is right, and the curve says it is, what part of your organisation still exists in 2030, and who then owns the infrastructure the rest runs on?

We have the money. We have the people. What we do not have is an excuse.

Comments welcome, especially from those who disagree. And from those who want to build instead of talk.