Is the Netherlands building the infrastructure for the AI economy?

The Netherlands likes to present itself as a knowledge economy. Our economic strength lies in knowledge, technology, services, innovation, logistics, financial services, high-tech and data. In almost every government document on future growth, the same terms recur: artificial intelligence, digitalization and data-driven work. The ambition is clear. The Netherlands wants to count within the international AI economy.

But a knowledge economy of tomorrow is also a digital economy. And a digital economy rests on a physical layer that is less visible than an algorithm or a startup: energy, the electricity grid, data centers, fiber optics, digital infrastructure, buildings, workplaces and ultimately the user. That chain is one whole. The question in this article is whether the government treats that chain as a single economic issue as well.

This is not a political argument. It is an attempt to factually organize what the Netherlands is doing, what other countries are doing, and where the possible blind spots lie. And to make visible a layer that is missing from most AI discussions: the building where the user actually works.

Where the Dutch AI ambition places its emphasis

Anyone who lines up the publications of the Dutch national government, ministries and the House of Representatives on AI sees a consistent picture. Attention goes mainly to talent, research, responsible application, regulation, cybersecurity and stimulating AI adoption among businesses and public authorities. Chips, compute and European cooperation also recur, partly in light of the broader digital strategy of the European Commission.

That is understandable and valuable. Without talent, frameworks and trust, no digital economy gets off the ground. But it is striking how much of that ambition plays out at the digital front end: software, models, skills and policy. The physical conditions, in other words the question of whether there will be enough power, grid capacity, data connections and suitable buildings, receive relatively less integrated attention.

That raises a critical question. Do we mainly have an AI strategy for innovation and software, or also an infrastructure strategy for an economy that is going to digitalize exponentially more? Both are needed. The difference between the two is exactly what this article is looking for. In our knowledge base we go deeper into the technical side of that underlying layer.

The infrastructure paradox

The Netherlands wants more AI, more electrification, more electric mobility, more heat pumps, more sustainable industry, more data centers, more smart buildings and more economic growth, all at the same time. Each of these wishes makes sense on its own. Added together, they nearly all demand more from the same underlying facilities: electricity and data.

And that is exactly where it pinches. The Netherlands is dealing with grid congestion. In large parts of the country, businesses cannot simply obtain a new or heavier connection, and in some places this now also applies to feeding power back into the grid. Grid operators are investing heavily in expansion, but the waiting lists for connections and the lead times for grid reinforcement are considerable. Sources such as Netbeheer Nederland and the ACM sketch a picture in which demand grows faster than the grid can be expanded.

Data centers and digitalization add an extra dimension to this. They require not only space, but also substantial and reliable electricity. When AI applications are rolled out widely, that demand increases further. The central tension can be summarized as follows: can an economy digitalize exponentially when its physical infrastructure can mainly only be expanded linearly?

An economy cannot realize an exponential digital ambition on a physical layer that can only be expanded linearly.

AI moves faster than the foundation it rests on

Part of the problem lies in differing development speeds. An AI model can change fundamentally within months. A company can largely digitalize its processes within a few years. But a building is developed for decades. High-voltage infrastructure and large-scale energy production have development, permitting and construction trajectories that likewise span years to decades.

The tension can be summarized concisely: AI thinks in months, companies think in years, real estate thinks in decades, and infrastructure sometimes also thinks in decades. These speeds do not automatically align. And that raises an organizational question that often goes unanswered: who ensures that these four rhythms connect with each other?

For those working in real estate, this is not an abstraction. A choice about cabling, technical rooms or connectivity made today determines how usable a building is in an economy that will look very different ten years from now. That makes future-proof building not a luxury, but a way of preventing the slowest link from becoming the weakest one.

How other countries address the physical side

It is worth looking across the border, not at how much money countries put into AI, but at how they handle the physical conditions for it. On elements such as energy, grid capacity, data centers, national compute, connectivity, spatial planning and permitting, clear differences in approach emerge.

United Kingdom: AI literally gets room

The United Kingdom stands out with the concept of AI Growth Zones. Specific areas are designated where AI infrastructure can be developed on an accelerated basis. The distinctive feature is that this looks not only at technology companies, but explicitly at the combination of energy, land, water, planning, connectivity and data centers. The availability of large-scale energy capacity is a central criterion.

The United Kingdom thereby treats AI not solely as software policy, but also as a spatial and infrastructural issue. The question this raises for the Netherlands is simple: are we doing enough of that? Do we designate locations where energy, data and real estate deliberately come together, or does that mostly happen by chance?

France: is energy becoming a competitive advantage?

France has a relatively large share of nuclear energy in its electricity production. That gives the country a starting position to make itself attractive for data centers, hyperscalers and AI-intensive companies that require large amounts of reliable power. Substantial investment announcements around AI and data centers have been made in recent times. It is worth making a critical distinction here between projects that have actually been realized and plans that still need to materialize.

The underlying question reaches beyond France alone. Will access to large volumes of reliable electricity soon become just as decisive for location choices as tax climate, talent and accessibility? If so, competition between countries partly shifts toward their energy position.

Germany: the same European challenge?

Germany is a useful mirror for the Netherlands. It has a large knowledge economy, strong industry and high-quality technical infrastructure, but at the same time struggles with energy prices, grid expansion and the pace of digitalization. The comparison between the Netherlands and Germany is therefore not about who spends the most, but about who best connects AI, energy policy, industrial policy, data centers and digital infrastructure. Both countries likely run into partly the same structural European bottlenecks.

United States: AI becomes infrastructure politics

In the United States, investments in AI data centers, energy production, electricity grids, chips and hyperscale computing are becoming increasingly intertwined. Technology companies are making agreements with energy suppliers, considering private power supplies, and explicitly incorporating nuclear energy, gas and renewable sources into their plans. Amounts in the order of many billions are being mentioned, with tech companies, energy companies, infrastructure funds and financial institutions jointly looking at AI infrastructure.

That leads to an uncomfortable comparison. When other economies mobilize hundreds of billions for the physical infrastructure behind AI, can the Netherlands make do with mainly digital front-end policy? The answer does not have to be that we need to spend the same amount. It does mean that we need to take the physical side just as seriously as the software side.

An international scorecard

To compare the countries side by side, a simple comparison helps. Not with invented figures or grades, but with a qualitative assessment per topic. The categories are: well developed, developing, point of attention, and insufficiently mapped. This is about direction, not precision.

Two things stand out. The Netherlands does not score distinctly low on any topic, but has a point of attention on several physical elements and shows little visible policy on AI zones. And in the last row, digital infrastructure in buildings, virtually no country has systematic insight. That is exactly where the next section begins.

But there is still a layer missing

Suppose the Netherlands were to realize sufficient energy, data centers, fiber networks and compute. Then there is still one final physical link: the building where the user works. AI is not only used inside a data center. The computation can take place there, but the use happens from offices, business buildings, hospitals, universities, laboratories, government buildings, campuses, shops, logistics centers and millions of individual workplaces.

We can invest billions in the digital highway, but what good is that if the final digital kilometers to and inside the building are not sufficiently prepared? For office buildings and other commercial properties, this is not a theoretical matter but a daily reality for tenants and owners.

The digital last mile of our economy

Telecom companies know the last mile: the final stretch of the network to the user. Translated to commercial real estate, that is the final hundred meters to the workplace. Fiber optics may run through the street. A data center may be twenty kilometers away. Large cloud providers may offer virtually unlimited capacity. But ultimately, an employee on the sixth floor needs to be able to work without interruptions.

That requires a series of concrete facilities:

• good building connectivity and sufficient capacity;
• modern internal cabling and reliable WiFi coverage;
• redundancy in connections and power;
• secure technical rooms and network segmentation;
• cybersecurity arranged at the building level;
• scalability and a clear division of responsibility between landlord and tenant.

Perhaps one of the biggest blind spots in our national AI strategy is not the data center, but that last hundred meters. Anyone who studies network redundancy and fiber optics in buildings notices how often this link is overlooked in conversations that are otherwise highly technical and ambitious.

We measure energy, why not digital readiness?

The Netherlands has taken major steps in making the energy performance of real estate visible. We increasingly know how energy efficient buildings are, what sustainability improvements are needed, and where investments become necessary. The energy label has made that information part of decisions about renting, buying and managing property.

But do we also know how digitally future-proof our buildings are? How many Dutch offices have redundant fiber optics? How many buildings have modern data cabling? How many technical rooms meet current requirements? How many owners know exactly what digital delivery level they offer? As long as no reliable national data exists on this, that is itself an important conclusion: we simply do not yet know well how AI-ready our built environment is.

That insight is missing exactly where it is needed most, namely in investment decisions. Anyone wanting to know more about the relationship between the two measurement systems will find a detailed comparison in our article on the step from energy label to IT-label.

Who looks at energy, data, real estate and AI together?

The Netherlands makes policy for energy, climate, AI, cybersecurity, digitalization, real estate, spatial planning and telecommunications. Each domain has its own ministry, its own rules and its own pace. The question is not whether policy exists, but who brings these areas together.

Who looks in an integrated way at the combination of energy, data, real estate and AI? Is it Economic Affairs, Climate and Green Growth, Interior Affairs, Infrastructure, municipalities, provinces or the grid operators? Or does it ultimately remain nobody's domain, because it belongs a little bit everywhere and fully nowhere? The core policy question is therefore this: does the Netherlands actually have one party responsible for the physical infrastructure of our future digital economy?

From grid congestion to digital congestion

It is worth cautiously introducing a way of thinking: digital congestion. This is explicitly not the same as the formal congestion on the electricity grid. It is a way of thinking about capacity. The Netherlands has learned in recent years that economic growth can stall when physical energy infrastructure lacks sufficient capacity.

The question is whether we want to prevent having a similar discussion about digital capacity in buildings five or ten years from now. Without insight into the digital readiness of the building stock, there is a real chance that at that point we will have to conclude: we could have seen it coming. That is exactly the kind of surprise that can be avoided through timely measurement, as we also describe in our piece on digital congestion.

The IT-label as a measurement and translation layer

It is important to be precise here about what the IT-label is and is not. The IT-label does not solve grid congestion, does not build data centers, does not produce electricity and does not develop AI models. It is an independent knowledge collective and a classification methodology, not a certification with legal status.

What it can do is make one blind spot visible: how digitally future-proof is the commercial real estate in which our economy functions on a daily basis? The classification from PREMIUM to SHELL translates technical building characteristics into an understandable delivery level. This way, the label makes the difference between, for example, IT1+ PREMIUM and IT5 SHELL visible to people who know real estate but not necessarily IT.

In this way, the IT-label can function as a link between IT, real estate, users and government. A shared language in which digital building quality becomes structurally visible, exactly at the level where policy and daily practice meet.

From business location climate to digital location climate

The Netherlands traditionally assesses its international competitive position based on talent, taxes, accessibility, education, energy, regulation, the housing market and the labor market. A dimension can be added to that list that will weigh more heavily over the next ten years: the digital location climate. How easily can a digital company actually operate and grow here?

An AI company does not simply choose a country. It chooses an ecosystem of energy, compute, data, talent, connectivity and real estate. If one of those factors structurally lags behind, the investment shifts to a place where the whole picture fits. Digital real estate thereby becomes part of economic location policy, no longer a detail that only comes up upon delivery. For owners who want to understand what this means for the value of their real estate, this is a shift with direct consequences.

What property owners and policymakers can do now

If the Netherlands wants to become the first country to map not only its energy landscape but also the digital readiness of its commercial building stock, that starts with measurement. It then becomes visible where digitally high-quality buildings are located, where investments are needed, where fiber optics and redundancy are lacking, and which economic areas are digitally vulnerable.

The question for the government is ultimately not only how we ensure that Dutch companies start using AI. It is also whether we will actually have the energy, the infrastructure and the buildings to keep that AI economy in the Netherlands. The countries that win the AI economy may not be the countries with the best algorithms, but the countries that understand that a digital economy is built on physical infrastructure.

For property owners, the concrete next step is smaller and more direct: map out what digital delivery level your building currently offers. A good first step is to read why an IT-label makes that information visible, or to have your property's current status investigated via contact. Knowledge runs on data, data runs on infrastructure, and that infrastructure ultimately comes together in real estate. That is where a digital location climate begins.