From steam engines to AI clusters: a historical perspective
Every technological revolution is supported by a specific infrastructure. Steam locomotives required extensive railway networks, electrification brought high-voltage grids, and the internet demanded fiber optic connections and data centers. Today, we are on the threshold of a new phase: large-scale artificial intelligence. AI is not a marginal technological innovation, but a fundamental productivity driver that is redefining industrial structures, value chains, and geopolitical power relations.
Historically, countries that invested early in the dominant infrastructure of their time achieved long-term economic and geopolitical advantages. In the 19th century, Great Britain built a railway network at record speed, the United States invested in a national highway system after World War II, and South Korea laid a digital backbone in the 1990s that now makes the country a world leader in semiconductors. Today, the focus is on large-scale, energy-intensive computing power.
AI runs on physical infrastructure: steel, silicon, and stable energy
Although AI is often associated with abstract software in “the cloud,” the technology relies on tangible, large-scale infrastructure. Behind every algorithm are high-density data centers with advanced AI chip clusters, connected by ultra-low latency networks and powered by stable, preferably sustainable, energy sources. Without this physical foundation, AI remains a theoretical concept rather than a catalyst for economic transformation.
The vast majority of this digital infrastructure is privately owned. Nevertheless, government control is essential for acceleration and coordination. Examples from abroad demonstrate this: France is investing billions in national AI supercomputers with long-term energy contracts; the United Kingdom is developing “AI Growth Zones” that combine computing power, research capacity, and business innovation; in the US, AI Compute Clusters are emerging in states such as Virginia and Texas; China is focusing on complete self-sufficiency, from chip production to strategic software, with an emphasis on direct applicability.
The Dutch starting position: strong assets, concrete risks
The Netherlands has significant strategic advantages: a high-tech industry, a strong position in European data traffic with international data connections by land and sea, a mature data center ecosystem, leading universities, and growing offshore wind capacity. However, AI places specific and new demands on infrastructure: extremely low latency, predictable and continuous energy supply, very high computing density, and a scale that far exceeds the current Dutch data center landscape.
If the Netherlands wants to continue to play a significant role, action must be taken now. Without an integrated strategy, we risk falling behind. In a context where licensing processes and policy discussions can take years, countries that invest now will soon determine the technical standards and market dynamics. These rules rarely change in favor of latecomers.
Towards a national AI infrastructure strategy
The urgency of the situation calls for a long-term strategy for AI infrastructure that integrates policy, spatial planning, energy supply, and talent development. Key questions include: Where do we establish new data center clusters? How do we integrate them into our energy system? How do we connect AI infrastructure to energy grids in a sustainable way? How do we ensure that highly qualified technical personnel are immediately available? How do we combine public and private resources on a large scale?
In addition, a strategy requires a realistic view of energy intensity. Training advanced AI models requires up to ten times more energy capacity than conventional IT workloads. This means investments in large-scale green energy, long-term security of supply, and load balancing strategies that utilize peaks in renewable generation.
The potential: from AI-ready campuses to digital sovereignty
With the right choices, the Netherlands can develop AI-ready campuses that bring together high-density computing power, universities, innovative startups, and energy hubs. Our geographical location offers opportunities to deliver low-latency AI services to the whole of Europe. In addition, developing an AI cloud infrastructure can increase our strategic autonomy and strengthen the anchoring of European values and ethics in AI systems.
Whoever possesses AI computing power possesses the key to innovation. The Netherlands must forge that key now.
Conclusion: breakthroughs are needed, time for coordinated action
The debate on AI infrastructure transcends political divides and sectoral interests. Just as railways, electricity grids, and the internet were strategic assets in previous eras, AI computing power is now. Investing in an international position and in technological capacity means investing in our strategic autonomy.
The choice is binary: do we start building AI infrastructure now and claim a leading position, or do we join the queue of digital followers tomorrow?