Achieving full technological independence would cost Europe about €3.6 trillion over ten years, versus roughly €300 billion if it opts for alliances with other countries, according to calculations by Nicklas Lundblad, a non-resident fellow at the Center for European Policy Analysis (CEPA). The author, who previously led public policy at Google DeepMind, argues in the piece “Digital Sovereignty: Can Europe Afford It?” that the European Union should replace the goal of self-sufficiency with a network of interdependencies with the United States, India, and Southeast Asia. His figures are his own estimates and, as he acknowledges, approximate.
The cost of European digital sovereignty in 20 seconds
- Lundblad puts Europe’s technological self-sufficiency at about €3.6 trillion over a decade.
- Alliances with the United States, India, and Southeast Asia would cost about €300 billion.
- These are rough calculations from a former Google DeepMind executive, with a margin of error of 40%.
The piece was published on October 29, 2025, in Bandwidth, CEPA’s technology-policy publication — a Washington-based think tank focused on transatlantic relations. CEPA notes that the views are the author’s own. A year later, the debate he raises is still very much alive: digital sovereignty is one of the European Commission’s priorities and comes up every week in announcements of sovereign clouds, from the project Telefónica and Indra are studying to AWS’s European sovereign cloud.
How You Get to €3.6 Trillion
Lundblad adds up six line items over a decade. The largest isn’t an investment at all but an opportunity cost: €1.7 trillion, derived from assuming that trying to catch up would shave 1% a year off the growth of a €17 trillion economy. Next come semiconductors, at €680 billion to build a dozen fabs and their supply chain, benchmarked against TSMC’s investment in Arizona; cloud and AI, at €500 billion, compared with the infrastructure spending of Meta, Amazon, or OpenAI; software, at €300 billion; the salary premium needed to attract talent, at €250 billion, calculated from 500,000 workers earning €50,000 more a year; and a services layer equivalent to Google or Meta, at €200 billion.
| Line item (10 years) | Author’s estimate |
|---|---|
| Opportunity cost (1% of GDP per year) | €1.7 trillion |
| Semiconductors | €680 billion |
| Cloud and AI | €500 billion |
| Software | €300 billion |
| Talent salary premium | €250 billion |
| Services layer | €200 billion |
| Total self-sufficiency | €3.6 trillion |
| Alliance path | €300 billion |
The author himself presents the figures as rough estimates and admits they could swing 40% in either direction. There are also some caveats worth noting. Nearly half of the total — the opportunity cost — rests on a single assumption, that 1% annual GDP figure, which the author backs with general references to the OECD and McKinsey rather than a detailed calculation. The talent line and the alliance-path figure rest mostly on the author’s own assumptions. And the gap between the two paths, which the piece frames as ten to one, actually works out to about twelve times by his own numbers. The €3.6 trillion is equivalent to a bit more than 20% of the EU’s annual GDP, spread over ten years.
The Alternative: Strategic Interdependence
Against a Europe locked into its own technology, Lundblad proposes spending about €300 billion over a decade on joint research centers, coordinated standards bodies, co-investment funds, and institutional capacity, with industrial open source as a centerpiece, to which he allocates between €30 billion and €50 billion. He also proposes opening public procurement to this model and having Europe lean on its role as a regulator and meeting point between tech ecosystems. His underlying argument is that the real risk isn’t depending on others, but depending on too few of them, and that diversifying partners builds more resilience than trying to do everything at home.
Europe today produces around 10% of the world’s chips, according to the piece, and the European Chips Act aims to reach 20%. For Lundblad, chasing that autonomy at any cost doesn’t pay off, because money alone isn’t enough to recreate a Silicon Valley. The view clashes with the path governments like Germany’s are following, which has started to turn digital sovereignty into contracts, GPUs, and European software, and with those who argue that an alliance with U.S. providers doesn’t remove the risk that American law applies to European data.
What the Analysis Doesn’t Settle
The piece measures the cost of one extreme: total autarky. Much of the political debate sits in a middle ground — which layers are worth controlling, such as critical data, government cloud, or defense chips, and where it’s enough to simply diversify suppliers. It also doesn’t calculate the cost of doing nothing if a partner changes the rules, a risk the diversification argument itself acknowledges. For many companies, sovereignty comes down to smaller decisions, like where to host data and with what technology — for instance, with private cloud and open tools like Proxmox.
Frequently Asked Questions
How much would full digital sovereignty cost Europe?
According to estimates by Nicklas Lundblad, published by CEPA, about €3.6 trillion over ten years, with a margin of error of 40%.
What alternative does the analysis propose?
A strategy of alliances with the United States, India, and Southeast Asia, built on joint research, common standards, co-investment, and open source, for about €300 billion over a decade.
Is this an official CEPA report?
No. It’s an opinion piece by a non-resident fellow, published in October 2025, and CEPA notes that the views are the author’s own.
Sources:
