Digital sovereignty: now what?

These are interesting times for our field. While the fundamental impact of AI continues to boom, we must now also take geopolitical developments into account. In short, time to put some thoughts in order. And while AI feels like an opportunity, geopolitical developments seem to be mainly a threat.

But before we delve further into that, how did we actually get here?
‍

Our digital journey so far

If you look at the international (real) economy, 'globalization' is perhaps the first thing that comes to mind: a movement where often complex production chains are optimized, and this is not limited to national borders. We now find it quite normal that a lot of manufacturing has moved to so-called low-wage countries (first Eastern Europe, then China) simply because the costs are lower. And I know I'm oversimplifying a bit, but this movement has been primarily driven by efficiency (= costs). As consumers, we all thought that was fine: it has allowed us to buy cheap goods.

The first time we collectively scratched our heads was perhaps during the corona crisis. Do you remember the lack of face masks? And the almost non-existent production capacity in Europe for vaccines?

But what about the IT side? Although the internet, as we know it today, has its origins in the extensive international academic community, it was mainly the American business community that took the lead when the internet became available to the 'average man and woman'. Logical in itself: Microsoft already had a monopoly on operating systems (Windows) by then and was therefore also 'front seat' (after some initial hesitation) during the introduction of the internet. Does anyone still remember the so-called 'browser wars'?

Back to today. We all have an iPhone or an Android, listen to music at home via our Sonos, use Strava while running, drive to our holiday destination via Google Maps, and are addicted to one or more social media platforms. What do they have in common? They are all American companies. And these are just a few: if you start making a list, you might be shocked by how long that list is...

And behind these solutions lies complex software ("platforms") which ultimately also needs to run on infrastructure ("the cloud"). And this deeply technical infrastructure is also largely an American affair: we are primarily talking about the so-called hyper-scalers.

And what we are slowly discovering: it's not just about typical consumer products, but also solutions like iDeal, DigiD, payment transactions in general, and healthcare: all dependent on these American companies.

And finally, you can also add the current AI revolution to that.

Have there been no critical voices about these dependencies in recent decades? Certainly, but this criticism mainly came from the security/privacy perspective: not necessarily the most exciting topic at a birthday party. And in the business world, the convenience and cost aspects dominated, so in summary, we haven't done much with this criticism.
There was also no political interest or urgency to do anything about this: the Atlantic friendship is deep and long, after all. And the Netherlands in particular is (or was?) particularly pro-American: think here, for example, of our culture (music, film, TV, etc.), and in business we always look with some admiration and respect at American entrepreneurship: "bold and fearless."

And when I look at myself: as someone born in 1974, I belong to the generation that grew up with the idea that America is like a beloved distant relative: that uncle you always looked up to and happily listened to when he told a story. And for me, America's role in the Second World War, passed down from grandparents, was also a major influence.
‍

In short

No reason to worry about this or even act on it at all... Until 2025. With the election of Trump for his second term, it quickly became clear that a fundamental change in the long-term relationship was underway: perhaps the love we thought was mutual was a little more one-sided than we thought. And, looking back at the beginning of this year (Greenland, remember?), perhaps even hostile.

‍

For me, an important moment in this realization was one of the actions of the American government in the summer of 2025: as an opponent of the ICC (International Criminal Court) in The Hague, the American government imposed personal sanctions against a number of judges. And in doing so, Microsoft was forced (!?) to restrict these individuals' access to email. In short, technology is being used as a tool of power. The ICC has since switched to a European supplier.

While politicians are still recovering from the shock and struggling to accept the breakup, you can see that the business community is already adapting to varying degrees: part of doing business is estimating risks and anticipating/acting where necessary. And for those who think: "But Trump will be gone in 3 years, won't he?" That's true, but the breakup was not initiated by Trump: this process has been going on for some time. So 'going back to the way things were' will no longer happen, is my firm conviction.

In addition to making an internal inventory of our own dependencies, we are also receiving the first questions from customers. These are often at an abstract level: "Can I get off the American cloud?". And as much as I would like to help a fellow entrepreneur with a simple "yes": it's not that simple.

And a bit more nuanced: it's very good for a company to think about it and conduct an inventory. Insight into the services, along with a rough risk assessment ("what can we do if this service is no longer available?"), is already very valuable. And for the entrepreneurs/executives with less knowledge of software engineering: this is precisely where the dependencies are often the greatest. I understand that it's annoying if your email isn't working, but if your development team can't work, the impact might be greater. In short: start with this inventory. Make distinctions: office software, tooling for engineers, infrastructure. And involve the entire organization: it won't be the first time that there is absolutely no clear insight at the management level into all the external IT services being used. Here too: the list will probably be longer than expected...

Then take the inventory one step further: investigate whether there is also an alternative for a specific service. And again: investigating does not mean actually executing. For our office environment (mail, chat, documents, etc.) we use a solution from one of the hyper-scalers. We have conducted an investigation into the possible alternatives, but we have not (yet) decided to actually implement this. But we are prepared!

Perhaps there is also 'low-hanging fruit': a very specific cloud service for which there is also a European alternative, with a 1-to-1 match in terms of functionality and price. Maybe you can make such a transition already. What does this yield? Not necessarily much, but it does provide experience and feedback from users, for example. All very valuable.

If you work in an organization that also develops (and hosts) its own software, this inventory may be even more important. Modern software engineers are to a large extent 'power users,' and their daily work may depend on a whole palette of cloud services. And migrating to another (European) provider is more than just 'getting used to the new software'; it also directly impacts their daily work and may necessitate changes in their own software - which, of course, again impacts the business. In short, perhaps not the most exciting topic to deal with, but given the potential risks, it is good to think about and create insight.

Additionally, it is precisely on this side where perhaps the most concrete, relatively small, steps can be taken. So, you don't necessarily have to leave a hyper-scaler tomorrow to go back to local iron (the old "server closet in the basement"), but you can prepare the software for a move. Furthermore, it is also sensible to steadily reduce the dependence on a specific hyper-scaler: this at least offers the possibility to 'move' to another hyper-scaler in a relatively short time, with relatively little effort. Have discussions about 'cloud native' vs 'stack native'. In short, on the engineering side, we see all kinds of possibilities to lower the various risks, and especially to be 'future fit' for what may await us.

What the inventory also taught me is that a starting point of "it has to work just as well and it shouldn't be more expensive" quite quickly leads to a hard blockade: the advantages of the hyper-scaler in terms of price and functionality are simply not always found 1-to-1 in a European alternative. Be prepared for this and keep it in mind, because otherwise, you don't even need to start the inventory.
‍

Conclusion

I advise every entrepreneur/CEO/CTO/CPO to start working on this. Hoping that it will all turn out fine and well doesn't sound like a solid strategy. If the past year and the beginning of 2026 have shown anything, it's that "oh well, it's not so bad" might not be the wisest attitude.
‍

Final note

Although the reason is concerning and we live in uncertain times, viewed more abstractly, it is also an enormous opportunity to give our high-quality European technological sector a much-needed push to determine our own future. And for that, we as Europe must stop "playing Calimero": Forward Europe!

‍