European Digital Sovereignty Is a Stack Problem

I'll be blunt: I think European digital sovereignty, as currently discussed, is mostly a mirage. The motivation is obvious — of course you'd want it. But the gap between wanting it and having a credible path to it is enormous, and I don't think most of the conversation is honest about that.

The reason is structural. To see it, you have to look at the full stack.

Layer by layer

Start at the top: the app layer. Most of the applications people actually use — whether it's TikTok, Slack, or whatever SaaS product runs your company — are American or Chinese. This is the layer where Europe has the best shot at competing. Entry barriers are real but not insurmountable. You could, in theory, build sovereign alternatives here under certain circumstances.

One layer down: the browser. Most people run Chrome, Edge, or Safari. Building a browser isn't actually that hard — you can base it on Chromium, distribution is manageable, and we do in fact have a European browser company. This layer is doable.

One more layer down: the operating system. This is where it gets tough. Both Windows and macOS are proprietary systems made by giant American corporations. Android is trying to push into desktop territory, but the jury is still out. The other option is Linux, which is open source and European in spirit — made by a Finnish guy. But as a desktop OS, Linux is not up to the job, as the usage statistics trivially demonstrate. It's not that it can't be done. It's that it requires an enormous amount of work that nobody is currently doing.

Below that: the hardware itself. The computers people use are generally manufactured by American or Chinese companies. There's no European player here that seems positioned to change that.

On the server side, things look slightly better. Linux is the default operating system for servers almost everywhere, so that's one place Europe is genuinely strong. But those servers need to run somewhere, which means cloud infrastructure — and here the big three are all American. The European cloud alternatives that exist are, quite frankly, terrible by comparison. Standing up a hyperscaler is a gargantuan task.

Then there's specialized hardware — GPUs, specifically. Nvidia holds what amounts to a monopoly, with a distant second that's also American, and some up-and-coming Chinese options. Europe is nowhere here.

And finally, frontier AI models. The Americans and Chinese are leading. The European alternative is not visible.

Adding up the score

If you tally this up honestly:

• Apps: Americans lead, Chinese competitive. Europe has a chance here.
• Browser: Americans lead. But entry is reasonably easy.
• OS: Americans dominate. Changing this does not seem remotely reasonable in the near term.
• Hardware: Americans and Chinese. Not feasible to create a competitor at this point.
• Cloud: Americans dominate. Creating a competitor is a gargantuan task.
• Specialized hardware: Nvidia dominates. Creating a competitor is a gargantuan task.
• AI models: Americans and Chinese. Europe is absent.

That's a lot of gargantuan tasks stacked on top of each other. Digital sovereignty means all of these layers, not just the convenient ones.

What could actually be done

I don't think any of this means Europe should stop requiring data residency or enforcing regulations on how American companies handle European data. That's reasonable and should continue.

But the broader ambition — asking people to use European products over American ones — requires reckoning with the fact that at most layers of the stack, European alternatives are meaningfully inferior. Pushing people toward worse products makes them less competitive at their actual work, which doesn't help anyone.

If Europe decided, as a serious multi-decade political project, that this had to change, I think parts of it could be addressed. Governments could mandate European operating systems for public sector use and fund the commercialization effort to make Linux viable for consumers. Economic incentives could be structured to favor European cloud deployment. The browser layer is already within reach.

But this would need to be a coordinated political push on a scale that's been essentially unseen in European history — maybe not as large as the post-war reconstruction, but that's the kind of scale we're talking about. And that's before touching specialized hardware or AI models, which seem like exceptionally difficult problems even with political will.

The uncomfortable conclusion

My honest read is that most current talk of European digital sovereignty is doing more harm than good. It creates the appearance of a strategy without the substance of one, and in the meantime it incentivizes choices that make European companies less competitive. I think naming the constraint clearly is more useful than pretending the aspiration is close to achievable. The stack is deep, the gaps are wide at nearly every layer, and closing them would require a level of coordination and investment that I don't currently see on the horizon.

Maybe that changes. I'd like it to. But I'd rather be honest about the distance than optimistic about the direction.