No data centres, no AI champions

• AI ambition is high, but infrastructure awareness is low - Executives expect AI to transform business models and reshape jobs, but underestimate how far Europe and the Netherlands are falling behind in data centre capacity. 

• Data centres create a powerful competitive flywheel - Regions that invest in local computing power offer competitive advantages through lower costs, accelerating innovation and attracting talent and startups. 

• Closing the gap requires collective action - Europe has the talent to build AI champions, but scaling them demands coordinated investment, fewer internal barriers, and smarter government demand that kick-starts markets. 

AI applications come out front and centre in Opens in a new tabour recent 2025 CEO survey. Dutch executives view AI as the source of both the greatest opportunity as well as the strongest concern. A majority is already seeing a significant impact on business processes, and as many as 84% expect it to have a major influence on the company’s revenue model. They expect AI to have the biggest contributions in the support of R&D activities and in delivering tailor-made offerings through better customer segmentation. 25% are 'certain' that in the next few years AI will replace, on average, a quarter of current jobs. 

Respondents clearly identify several obstacles and challenges in putting AI to work. Financing the required investments is seen as an important issue, and so is attracting and retaining the talent needed to develop and implement AI applications.

However, Opens in a new tabthe survey also suggests that one obstacle is severely underestimated: only 21% are worried that the Netherlands is falling behind in the global AI and data center race. Only 2% of Dutch executives think the Netherlands isn’t investing enough in AI infrastructure. Meanwhile, 48% think investment levels are what they ought to be, while another 48% thinks we're doing enough but think we could do more. 

Our recent Opens in a new tabING Research publication 'The data centre divide: Why Europe's shortfall threatens future economic growth' paints a less optimistic picture. The paper forecasts that global data centre capacity will increase by roughly 15% per year until 2030. At the same time, it points out that this growth will be very unevenly distributed.  

It’ll be mostly driven by the US and China, who are well ahead of Europe already, by a factor of 3 and 1.5 respectively. As the report notes: 'The US will more than double its current capacity and China’s capacity will be roughly 2.5 times its current level. Europe, on the other hand, will increase its current capacity by around 75% until 2030. [...] In short, Europe is already behind the US and China, and the gap will widen in the years to come.' 

Wim Steenbakkers is certainly worried about this development. Like the survey respondents, he sees a pivotal role for AI. "Future competitive advantage will undoubtedly be fueled by AI, and AI, in turn, is fueled by data centres. So if the Netherlands, and by extension the EU, wants to build a competitive digital economy, they need to increase their efforts to build their own digital infrastructure, especially data centres."  

To further underscore how Europe is being left in the dust: the combined data centre capacity of London, Frankfurt, Dublin, Paris and Amsterdam — Europe's core markets — is already dwarfed by the capacity located in the US state of Virginia alone. Planned investments there, together with the Amsterdam moratorium on data centres, will only widen the gap.  

But why should it matter to European companies whether their AI applications run in a Virginian data centre or one in the Eemshaven? For a host of reasons, it turns out, ranging from legal and regulatory obligations to data security and to technical performance and continuity risks. 

But what's important from a competitiveness perspective, Steenbakkers says, is the flywheel effect that can originate from data centres. "Take, for example, the fact that Amsterdam is still home to companies like Booking and Adyen. That cannot be viewed separately from the city’s long-standing position as a frontrunner in digital infrastructure. The Amsterdam Internet Exchange, globally one of the largest internet hubs and perhaps the city's best kept secret, is an essential part of this."  

Data centres play a similar key role in providing the digital services that we increasingly use in our daily lives, making them a significant part of our economy. Their potential flywheel effect occurs because the more data centre capacity there is, the more this spurs innovation, which ultimately leads to more demand for data centre capacity.  

To quote from the Opens in a new tabaforementioned paper again: 'In countries with abundant data centre capacity, startups that build new AI applications and products can more easily access computing power thanks to better availability and lower marginal costs. This means that the tech successes of the future depend on strong digital infrastructure and the computing power that data centres provide. Therefore, countries with substantial data centre capacity have a better chance of achieving AI breakthroughs.' 

Of course, setting this flywheel in motion means facing real hurdles. If Europe wants to keep up in the global AI race, immense investments in data centres and power supply are needed. A different kind of hurdle is the 'not-in-my-backyard' attitude that Steenbakkers dislikes: "We need to lose the idea that data centres are just these ugly, energy-guzzling boxes that spoil the landscape and pollute the environment", says Steenbakkers. "That image merely rouses public indignation but misses the key point. First and foremost, data centres form the core of the future-proof digital economy that we need if we’re to sustain our prosperity and well-being." 

Timing is of the essence, Steenbakkers emphasises. "The advances made in recent years in training AI models have been spectacular. And there's not a company that hasn't thought about how AI could benefit its business. Companies are now at the point where they need to start developing and implementing AI applications that will bring about such benefits." 

And while Europe may be behind in data centre capacity, Steenbakkers is optimistic about European prospects in this area: "I'm convinced European companies are capable of developing excellent AI applications. Take Opens in a new tabDeepMind, for example, one of the AI pioneers responsible for a number of breakthroughs in the field. Its early investors were American, eventually Google bought it and they've done an excellent job scaling and marketing it. But DeepMind was founded in the UK by British entrepreneurs. There's a powerful lesson in that: Europe has to step up and nurture and raise future AI champions itself."

What can be done to increase European investments in the AI infrastructure needed to support such champions? It’s a matter of risk allocation, says Steenbakkers. "There's obviously a market for AI applications, and where there's a market, governments can play a limited role. But collective effort is required. The EU and national governments should formulate a general strategy and remove internal market barriers. And there’s a clear government demand for AI applications too. Offering government contracts that incentivise competition in developing such applications is much more effective than handing out subsidies to set the flywheel in motion."