David Graus, affiliated with the Institute for Logic, Language and Computation at UvA, argues that Europe remains at a substantial distance from the US and China in scale and speed. At the same time, he identifies a potential advantage in Europe’s stricter approach to privacy, bias, and societal risk. In this view, regulatory discipline can translate into AI systems with stronger trust and quality characteristics. Rather than trying to mirror Big Tech, he points to “social AI” as a realistic path where technical performance and public values are integrated from the start.

Michiel Bakker, a Dutch AI scientist at MIT, is more optimistic about Europe’s ability to compete if investment choices become sharper. He argues for a coordinated national strategy that combines talent, compute capacity, energy infrastructure, entrepreneurship, and faster AI adoption across public and private sectors. He also highlights strategic infrastructure such as chips and data centers. His main warning is that the key risk is not only underinvestment, but misallocation, for example scaling compute without upgrading the power grid.

Both perspectives converge on the same tension: AI investment is accelerating globally, but market structure and deployment constraints are becoming more visible. Capital flows between major technology firms are highly interconnected, while limits around data center capacity, energy, and available training data are increasingly relevant. This makes targeted investment more important than investment volume alone.

For the Netherlands, that implies combining public research strength with selective ecosystem building. Initiatives such as the AI Factory in Groningen show how infrastructure and research investments can support both scientific positioning and economic impact. Read more at NEMO Kennislink.