The Sovereignty Illusion

Investigative report Across Europe, technology is marketed as sovereign. This report follows the money, the machines and the law to find out how much of it really is. Digital sovereignty means control over who can reach your data, who can switch your systems off, and who writes the rules, without foreign interference. This report tests that standard company by company, following ownership, infrastructure, technology, capital, and law to show where Europe's claims hold and where dependence remains. 50 companies scored 219 companies profiled 1,320 sources Five kinds of dependence Average score 4.5 out of 10 What the numbers show Fifty companies often presented as European champions, scored across five kinds of dependence. Three findings stand out; the distance between the claim and the measured result is the one with consequences. 4.5 /10 The average The typical champion sits below half-sovereign, closer to dependence than to independence. 7.4 The ceiling Only five of the fifty clear six and not one reaches eight. The genuinely sovereign players exist, but they are smaller and far less celebrated. +1.2 The gap Claimed sovereignty runs about 1.2 points ahead of the real thing, widest where the marketing is loudest, in artificial intelligence. See the full scorecard and all five dimensions Europe is everyone's preferred customer. A customer is not a power, and a customer can be cut off. From The triangle , on Europe caught between the United States, which owns the stack, and China, which is building its own. Choose your route You do not need a background in technology, finance or law. Pick the route that fits the time you have. In a hurry The ten minute version Read the prologue , then the scorecard. Together they give you the claim, the test and the result without the supporting detail. P 5 Best for Anyone who wants the claim, the test and the verdict in one sitting Time About ten minutes Go to the scorecard The full case Cover to cover Start at chapter 1 and read through. The argument is built in order: how the illusion is sold, who owns what, which machines we depend on, and why the law seals it. 1 2 3 4 5 6 7 8 9 E Best for Readers who want the complete case, built in order Time A few evenings Start at the beginning Looking for solutions What would fix it If you mainly want to know what real sovereignty would take, and what it would cost, begin with the final chapter, then read the epilogue. 9 E Best for People who care most about the remedy and what it costs Time About twenty minutes Read the blueprint The full report From the prologue through nine chapters and the dossiers to the epilogue, with the appendices and the full bibliography. P Prologue: the morning you realised The question almost nobody asks out loud, and the moment a company lawyer, under oath, could not promise that European data was beyond American reach. The words that matter and the two laws that decide who is really in charge. 1 The theatre How the illusion is built and sold. Four everyday tricks: the European headquarters that hides an American parent, the data residency promise that changes nothing, the partnership that is really a resale, and the sovereign artificial intelligence that runs on foreign chips. 2 The money Who actually owns European technology. American asset managers, private equity buyers and venture capital sit on the share registers of the continent's champions, and a large flow of capital and fees leaves for the United States every year. 3 The machine The physical stack that makes independence hard. The chips that train the models, the cloud platforms that host them, the cables that carry the traffic and the data centres on European soil that answer, in the end, to companies based abroad. 4 The law Why a contract cannot fix it. The CLOUD Act and FISA Section 702 reach across borders, the Microsoft Ireland case showed how, and European privacy law collides with all of it. The legal trap explained without the jargon. 5 The scorecard The evidence, out in the open. Fifty companies rated on five kinds of dependence: ownership, technology, capital, infrastructure and legal. How the scoring works, what the average of 4.5 out of 10 means, and where each firm lands, from the chip makers to the cloud champions. 6 The triangle Europe does not sit alone. It sits between the United States, which owns the stack, and China, which is building its own. This chapter maps the squeeze, and shows why being the customer of both is not the same as being a power in its own right. 7 The price What the dependence actually costs, in three currencies: the money that leaves each year, the strategic leverage handed to others, and the European futures lost when talent, profit and control all drain in the same direction. 8 The trap Why Europe stays stuck even when everyone can see the problem. The comfortable incentives, the sovereignty washing that sells dependence as independence, and the people and institutions that quietly profit from keeping things as they are. 9 The blueprint What real sovereignty would take, and what it would cost. A concrete plan with a price tag, the role of European capital, tax and public purchasing, the five traps to escape, and a clear set of actions for governments, companies, investors and ordinary citizens. 10 The Dossiers The sourced file behind every claim: a profile for all 219 companies and organisations named in the report, each with its ownership, its dependencies, its five-dimension European sovereignty score and a verdict. Searchable, filterable and sortable. E Epilogue Where the argument leaves Europe, and the choice that remains once the evidence is on the table. A Appendices A timeline of the key events, a plain language glossary, a note on how the sourcing works and acknowledgements, with the full bibliography on its own page. Every number in the report traces back to a source you can open. B Bibliography All 1,320 sources behind the report on their own page, each tagged by the authority of its origin, with search, an authority filter, sorting and pagination. About the authors Four perspectives, one investigation Marc Filipan leads the technical research and argument. The wider team contributes market operations, data governance, privacy, and theoretical review; the dossiers make every vantage point and the publisher's interest explicit. MF Marc Filipan CTO & co-founder, Dweve AI/ML researcher & systems architect · 25+ years · multi-exit entrepreneur · open-source contributor Marc Filipan is CTO and co-founder of Dweve. For more than 25 years, he has designed and built AI and machine-learning systems, neural-network architectures, infrastructure, fintech, cybersecurity, and high-throughput platforms. His work ranges from model and systems research to production engineering, including deterministic and low-bit AI, real-time data processing, fraud prevention, private infrastructure, and open-source software. At Dweve, Marc leads the technical architecture behind its binary, neurosymbolic, and auditable AI stack. His focus is on making AI systems more efficient, reproducible, inspectable, and practical for organizations that cannot rely on opaque black boxes. Core expertise Binary neural networks Distributed infrastructure AI systems engineering Cybersecurity Real-time data processing “ Every problem has a solution if you dig deep enough. I don't believe in impossible, only in systems we haven't understood yet. ” PV Peter Vieveen Chief AI strategy officer & co-founder, Dweve Theoretical review · former DAMA NL president · professional trainer · 35+ years Peter Vieveen is Chief AI Strategy Officer and co-founder of Dweve. He contributes theoretical review and context across data management, governance, privacy, education, and responsible technology adoption. His long experience in professional education helps test how technical claims translate into policy, organisations, and practice. Peter has trained hundreds of professionals across European organizations, universities, and public-sector institutions. Since 2017, he has lectured at the University of Applied Sciences Utrecht and CIBIT Academy, and he served as president of DAMA Netherlands from 2012 to 2023. At Dweve, he connects AI strategy with data governance, privacy, ethics, and the practical reality of using AI responsibly in organizations. Core expertise Theoretical review Data management & governance GDPR & privacy AI ethics Professional education “ True mastery comes from understanding both the theoretical foundations and the practical applications, knowledge that stands the test of time. ” BH Bouwe Henkelman CEO and co-founder, operations and growth, Dweve Trading floor veteran · 40+ years · risk management expert · partnership builder Bouwe Henkelman combines decades of high-stakes trading experience, from the Amsterdam exchange floor to electronic markets, with entrepreneurial drive. After building and running his own trading operations and investment holdings for more than 20 years, he recognised the practical potential of AI that can truly think and adapt. His understanding of risk, market dynamics and relationship building shapes Dweve's approach to enterprise partnerships and strategic expansion. At Dweve, Bouwe leads operations, business development and enterprise partnerships, and makes sure the technology translates into real business value for European organisations. He served on the DSI Disciplinary and Disputes Committee from 2000 to 2003 and as secretary of the Euronext Disciplinary and Disputes Committee from 1995 to 2003, with more than 40 years in derivatives trading and market making. Core expertise Strategic operations Enterprise partnerships Risk management Market analysis Business development “ Great technology without great execution is just expensive research. ” HG Harm Geerlings CEO and co-founder, product and innovation, Dweve Data innovation pioneer · CEDPP certified · DAMA NL co-founder · 18+ years Harm Geerlings is a CEDPP-certified data protection professional and a serial founder in the data intelligence space. He has spent more than 15 years building solutions at the intersection of data governance, privacy and intelligence, including helping major Dutch banks and government agencies achieve data-driven transformation, and co-founding DAMA Netherlands, where he serves as a board member. At Dweve, Harm drives product strategy, the innovation roadmap and the mission to create AI that thinks and reasons while keeping data within European boundaries. He is a guest lecturer at Utrecht University and a serial founder of data companies. Core expertise Data governance GDPR and privacy AI ethics and data sovereignty Business intelligence Data protection “ AI should not come at the cost of data control. I believe in systems that reason autonomously while keeping data firmly within European boundaries. ” The publisher Dweve hello@dweve.com Dweve is a European AI infrastructure stack built around deterministic, binary, and neurosymbolic computation rather than conventional floating-point, GPU-first LLMs. It combines low-bit and binary compute, symbolic reasoning, exact provenance, long-context retrieval, agent orchestration, and distributed and private deployment into one integrated stack. Its distinguishing bet is methodological: that binary, deterministic, and neurosymbolic approaches can make AI energy-efficient, auditable, and reproducible enough for regulated European environments in finance, healthcare, infrastructure, defence, and government. Dweve is also the publisher of this report. Where the report weighs Dweve's own approach, that interest is disclosed in the text and argued in the open rather than hidden. The Sovereignty Illusion · An investigative report on European digital dependence. Published June 2026. Every figure is drawn from public filings, parliamentary records, corporate disclosures and verified reporting, and is listed in the appendices. (function () {…