The Myth of Inevitable Dependence

Every successful European refusal followed the same pattern.

Inevitability has been sold to Europe before

In 1979, when the European Monetary System was first proposed in the form that became the euro, Robert Mundell, Milton Friedman and the consensus of American academic economics held that monetary union among heterogeneous European states could not work.

The structural diagnosis was that without a fiscal union, without labour mobility on American scale, without a federal political authority to manage asymmetric shocks, a single currency would either fragment in its first crisis or impose unbearable costs on its members. By 1999, eleven currencies had been replaced. By 2002, the notes and coins were in circulation.

The euro is now twenty-six years old, has weathered the 2008 financial crisis, the 2010-2012 sovereign debt crisis, the 2015 refugee inflow, the 2020 pandemic, the 2022 energy shock, and the 2026 Iran war and gas crisis. Its dissolution has been predicted in every decade. It has not dissolved.

In 1989, every major NATO briefing held that the division of Europe was a fact of the post-war settlement that would not be undone in our lifetimes. The 1986 Reykjavík summit had improved relations without changing the structural reality. The Soviet Union’s nuclear arsenal was a permanent fixture. East Germany’s Stasi state was not going anywhere. By November 1989, the Berlin Wall had fallen.

By October 1990, German reunification was complete. By December 1991, the Soviet Union had dissolved.

In 1947, when George Marshall announced the European Recovery Program at Harvard, the consensus among American economists and European political leaders was that recovery from the war would take generations. The continent’s industrial base was destroyed. Its currencies were collapsing. Hunger was widespread. By 1955, European industrial output had passed pre-war levels.

By 1960, the recovery was complete enough that the founding members of the European Economic Community were running trade surpluses with the United States.

In 1957, when the Treaty of Rome was signed, the consensus among American commentators and European sceptics was that integration would not survive its first decade. National interests would reassert themselves. The Common Market would fracture on agricultural prices. The Coal and Steel Community had been a wartime expedient that would not generalise.

The Treaty of Rome is now sixty-eight years old. The community it created is the largest single economy on earth.

Each of these was the structural consensus of its time, repeated across academic economics, policy commentary, government briefings and elite discussion until it became, for those listening, simply how the world worked. Inevitability is a story Europe has been sold many times.

Each time the story has been refused, the building that followed produced the institutions Europeans now consider permanent.

The story is being sold again now. The current iteration is that European digital dependency on the United States cannot be undone. That hyperscalers cannot be displaced. That European industrial policy in tech will fail because European industrial policy in tech has failed. That the architecture is built and the markets are won and the question of digital sovereignty is settled.

The story is older than the dependency.

They sell us inevitability

They sell us inevitability. We believe them. We can refuse again.

The “they” is concrete. In 2025, the technology sector spent €151 million lobbying European Union institutions in Brussels, up 33 per cent in two years and a record. Meta tops the list at €10 million. Microsoft, Apple and Amazon spend €7 million each. Google and Qualcomm sit at €4.5 million. Ten firms account for €49 million combined, roughly a third of total digital-sector lobbying.

An additional €9 million per year flows to consultancies, public-relations firms and think tanks that translate the lobbying into respectable analysis.

The think tanks are funded by the same firms. The Center for European Policy Analysis (CEPA), whose supporters include Amazon, Google, Microsoft, Meta and Apple, published a 2025 article calling EuroStack a “€5-trillion mirage” and “a recipe for stagnation.” The author was William Echikson, a senior fellow whose career includes six and a half years at Google.

The framing recasts sovereignty as nostalgia and dependency as openness: “Europe can spend trillions trying to duplicate the past, or it can use openness to own the future.” The Atlantic Council lists Meta, Amazon and Apple among its million-dollar contributors.

CSIS provides the “splinternet” frame, treating European digital autonomy as a balkanisation of a global commons that happens, by the framing’s logic, to be American-anchored.

The Information Technology and Innovation Foundation calls EU sovereignty “protectionism, pure and simple.” The Chamber of Progress, a US tech industry coalition, manufactures the round €5 trillion figure that other actors recycle.

The trade bodies coordinate the campaigns. AmCham EU, BSA, CCIA Europe and ITI issued joint statements through 2022, 2023 and 2024 demanding the removal of sovereignty requirements from the EU Cybersecurity Certification Scheme for Cloud Services.

ITI specifically commended the Commission for “improvements in the most recent EUCS draft that removes sovereignty requirements.” The March 2024 draft duly stripped the sovereignty tier.

The European requirement that sovereign cloud certification require EU headquarters and immunity from foreign-jurisdiction orders was, on the record, removed at the request of trade bodies funded by US hyperscalers.

The selling extends to the executive branch. Microsoft’s Brad Smith pledged in April 2025 to expand European data centre capacity by 40 per cent and “go to court” against any US government order to cease European operations. The pledge was timed to the German federal sovereign cloud procurement, which Microsoft duly won via Delos Cloud, the Azure-anchored partnership with SAP.

Meta’s Mark Zuckerberg announced in January 2025 a pivot to “work with President Trump to push back on governments around the world that are going after American companies,” framing European regulation as anti-American hostility. The Trump administration tied tariffs explicitly to European digital rules in February, August and December 2025.

The Europeans participate.

The European Round Table for Industry, the chairs of sixty leading European multinationals, came out against the EU’s tech sovereignty package in 2025, arguing that “in the absence of international partnerships, European companies would be forced to rely on less optimal services.”

The German federal government awarded its sovereign cloud contract to Delos Cloud (Microsoft Azure plus SAP) and signed Oracle framework contracts worth €4.8 billion to 2030.

The French Cloud de Confiance scheme licences Bleu (Capgemini and Orange reselling Microsoft Azure) and S3ns (Thales reselling Google Cloud) to wear the sovereign label over US technology.

In November 2025 the European Commission launched the Digital Omnibus package, proposing to weaken AI Act and GDPR provisions under a “competitiveness” frame, pacing the dilution that lobbyists had requested.

The selling has recurring moves. Cost shock: round numbers, usually trillions, presented without methodology or counterfactual. The €5 trillion figure travels faster than the analysis behind it. The protectionism frame: a categorical disqualifier deployed once and never argued. Once an action is labelled protectionist, no further analysis is required.

The splinternet warning: any European autonomy is balkanisation of a global commons, the global commons in question being one Europe does not own. Past-failure-as-prophecy: GAIA-X collapsed because hyperscalers captured it, but the failure is reported as proof that the project category is doomed.

Sovereignty-washing: rather than oppose sovereignty, hyperscalers and their European partners offer a sovereign-labelled version of the existing dependency that absorbs the political demand without changing the architecture.

The moves do not need to be made by every actor. They need only to be in the air, repeated often enough that the reader who has not heard them argued nonetheless takes them for granted. That is how myth-making works. The story does not need to convince everyone individually. It needs only to be present in the room when decisions are made.

The myth is the mechanism

The architecture is maintained by belief in its inevitability. Believing the myth produces the apathy that prevents action. The apathy preserves the architecture. The preserved architecture validates the myth. The selling is the dependence.

This is the operational mechanism. When a Chief Information Officer at a European hospital chooses Microsoft 365 over a European alternative, the choice is rational at the level of the individual decision: Microsoft is more functional, more familiar, more universally compatible. The rationality of the individual choice is what aggregates into the dependency.

Aggregated dependency is read by the next procurement officer as the established standard. The established standard becomes the basis on which the next CIO makes the same choice. Each individual rationality compounds into a collective trap that no individual can escape.

Refute the myth and the structure becomes contestable. Not because the architecture changes immediately, but because the rationality calculation shifts.

A CIO who knows that European alternatives are operational, that the Buy European procurement floor is being mandated, that the Sovereign Tax Wrapper makes European-listed equity tax-advantaged for European pension contributions, that the openDesk migration at the ICC produced a working result, is choosing under different premises.

The cumulative behaviour of CIOs choosing under different premises produces a different aggregate. The aggregate is the architecture.

The myth is the mechanism. The historical refusals are evidence the mechanism can be broken.

The refusals

Each historical European sovereignty build began with the inevitability claim of its moment. Each refused the claim. Each produced the institutions Europeans now consider permanent.

Airbus, 1970.

The 1965 Plowden Report, commissioned by the British government and read across Western European capitals, concluded that Europe could not compete in commercial aerospace at scale. American manufacturers had the order books, the launch capacity, the political backing and a domestic market large enough to spread fixed costs that European national champions could not match.

Boeing and Lockheed dominated commercial aviation. Douglas Aircraft, Convair and the rest of the second-tier American industry were already failing or merging. The structural conclusion was that European commercial aviation, if it was to exist at all, would exist as a junior partner to American leadership.

In 1970, the governments of France, Germany and the United Kingdom signed the Airbus consortium agreement. Spain joined in 1971. The structure was deliberately European-multinational, with risk-sharing partnerships across countries that no single nation could have funded alone. The launch aid was repayable.

The marketing strategy, under Roger Béteille, targeted American carriers directly. Eastern Air Lines bought twenty-three A300s in 1978 and the project survived its first decade. The A320 family, launched in 1988, took the single-aisle market. By 2003, Airbus was delivering more aircraft annually than Boeing. By 2024, it had been the larger producer for two decades.

The British government, having read the Plowden Report, had withdrawn from the original consortium in 1969 and rejoined as a minority partner only when the project was already operating. The reason given at the time was that Airbus would not be a commercial success.

Galileo, 2003.

GPS was, until 2003, the world’s navigation system. It was American military infrastructure operated by the US Air Force, with civilian access provided by political grace rather than treaty obligation.

The 1999 Kargil War demonstrated the implications: the United States degraded civilian GPS over the Indian Ocean for ninety minutes during the conflict, depriving Indian and Pakistani forces of accurate positioning at a critical moment. Brussels noticed.

The structural argument against Galileo was that GPS already existed, was good enough, and could not be displaced. The cost of building a parallel constellation was estimated at €5 billion at programme initiation; it has been higher in practice.

The benefit was the abstract one of independence from American military infrastructure during a period when American military infrastructure was understood to be allied. The American government opposed the project, raising concerns about signal interference with planned GPS military signals and threatening countermeasures. The French and British defence establishments were initially sceptical.

The European Space Agency took on the work despite the doubt. The first Galileo satellite launched in December 2005. The constellation reached full operational capability in 2020. Galileo now provides more accurate civilian positioning than GPS itself, with the Public Regulated Service offering encrypted signals for European critical infrastructure.

The 2026 Iran war was the first major conflict during which European militaries did not depend on American positioning for navigation in contested theatres.

CERN, 1954.

In 1954, particle physics required pooled European budgets at a scale no single nation could carry. The American argument, made repeatedly across the early 1950s, was that European physicists should join American institutions or accept second-rate facilities. The discipline was American-led, with Brookhaven, Berkeley and the early Argonne setting the centres of gravity.

European reconstruction was barely complete. Strategic priorities were elsewhere.

The CERN Convention was signed by twelve member states in July 1953 and entered into force in September 1954. The structure pooled scientific capability across countries that had been fighting each other less than a decade earlier. The political achievement was at least as large as the scientific one.

CERN now operates as a multilateral organisation with twenty-four member states, employs roughly 25,000 staff and visiting scientists, and operates the Large Hadron Collider, the most powerful particle accelerator in existence. The Higgs boson was confirmed at CERN in 2012.

The American particle-physics programme has, since the cancellation of the Superconducting Super Collider in 1993, played a junior role.

GSM, late 1980s.

In the late 1980s, the dominant mobile telephony standard in the United States was AMPS, the Advanced Mobile Phone System, an analogue cellular technology developed by Bell Labs and operated by AT&T’s spinoff Bell Operating Companies.

The American consensus was that mobile telephony would follow American standards, as had been the case for landline telephony, satellite communications and most of computing. European national champions were running incompatible analogue systems (TACS in the United Kingdom, NMT in the Nordic countries, Radiocom 2000 in France) that would not survive the transition to digital.

Conférence Européenne des Administrations des Postes et des Télécommunications, the European post-and-telecoms regulator, took a different decision. The Groupe Spécial Mobile working group, established in 1982, designed a digital cellular standard from first principles.

The standard was made mandatory across European national telecoms regulators in 1987 via a memorandum of understanding signed by thirteen countries. The first commercial GSM network operated in Finland in 1991. By 2000, GSM had displaced AMPS internationally.

By 2010, GSM and its descendants (GPRS, EDGE, UMTS) were the dominant mobile telephony standard worldwide, with the United States adopting them under different brand names. American firms eventually built handsets and chipsets to a European-designed standard.

The euro, 1999.

In 1999, when the euro launched as a unit of account, the consensus among American academic economists was that monetary union among heterogeneous European states could not work.

Robert Mundell’s optimum currency area theory, originally formulated to argue against fixed exchange rates within the United States, was applied to argue against the euro on grounds that European labour markets were insufficiently mobile, fiscal transfers were politically impossible and asymmetric shocks would tear the system apart.

Milton Friedman wrote in 1997 that the euro would survive its first crisis only by becoming the prelude to political union, and that political union was impossible.

The euro launched as a unit of account on 1 January 1999. Notes and coins entered circulation on 1 January 2002. The system has weathered the 2008 financial crisis, the 2010-2012 sovereign debt crisis (which required institutional reform under stress), the 2015 refugee crisis, the 2020 pandemic, the 2022 energy shock and the 2026 Iran war and gas crisis.

Twenty-six years after launch, the euro is the second-most-traded currency in the world by foreign exchange volume, the second-largest reserve currency, and the unit of account for an economy of 350 million people.

Ariane, 1979.

In the 1970s, American launch policy was that European satellites that would compete with American commercial interests would not be launched on American rockets.

The 1973 decision to deny launch capability to the Franco-German Symphonie communications satellite, on the grounds that it would compete with Comsat, was the moment European space-faring states understood they did not have launch sovereignty. The structural argument was that European launchers could not be commercially viable.

American Delta and Titan launchers had economies of scale and a captive Department of Defense customer. European demand was too small. Independent launch capability was a sovereignty luxury Europe could not afford.

The European Space Agency funded Ariane through a French-led consortium. Ariane 1 first flew successfully in December 1979. Arianespace, the commercial operator, was incorporated in 1980. By 1984, Ariane was capturing more than half of the commercial communications-satellite launch market. Ariane 5 dominated commercial launch from 1996 until SpaceX entered the market in the 2010s.

Ariane 6 entered service in 2024. Europe has had sovereign launch capability for forty-six years.

EURATOM, 1957.

In 1957, the United States controlled the global market for fissile material, enrichment technology and reactor design. The Atoms for Peace programme, announced by Eisenhower at the United Nations General Assembly in December 1953, was constructed to channel civilian nuclear development under American supervision.

The structural argument against an independent European nuclear treaty was that Europe could not develop civilian nuclear power outside the American framework, and that doing so would jeopardise the broader Atlantic alliance.

The EURATOM Treaty, signed alongside the Treaty of Rome in March 1957, established a separate legal framework for civilian nuclear cooperation among the six founding EEC members. The structure took civilian nuclear development outside ordinary competition law and provided for joint research, supply and safeguards. Sixty-eight years later, the treaty still operates.

Europe runs the world’s largest civilian nuclear programme. France’s nuclear reactor fleet provides 70 per cent of French electricity, the highest share in the world.

The next refusal

The historical builds were each completed before the moment of refusal had passed. The current build is being attempted now.

In the spring of 2025, France, Germany, Greece, Italy, Spain, Estonia and Finland announced the European Cable Initiative, a joint procurement and capability programme for submarine cable infrastructure under European ownership and operation.

The structural reason was that 99 per cent of intercontinental data passes through submarine cables,

that the cables landing in Europe were predominantly American and Asian-owned at the carrier and consortium level,

that cable repair and surveillance capability had been concentrated in Five Eyes intelligence-sharing arrangements,

and that the 2022-2024 Baltic Sea cable damage incidents demonstrated European vulnerability to actors who controlled cable infrastructure but not European jurisdictions.

The programme combines several mechanisms. Member-state procurement preference for European cable carriers. Joint funding through the Connecting Europe Facility. Capacity-building for cable-laying and cable-repair vessels (Alcatel Submarine Networks operates seven; Orange Marine operates nine; the European cable repair fleet is now the largest in the Atlantic).

European cable security through naval cooperation under PESCO. The chokepoint is named. The alternative is being built. The mechanisms are in place. The political will is forming.

The refusal is unfinished. The pattern is alive.

What can fail

The pattern fails when the refusal is not actually made.

GAIA-X was announced in 2019 as a European federated cloud project. Its founding documents committed to European data sovereignty, European-controlled infrastructure and European competition law. By 2021, Microsoft, Amazon, Google and Palantir were members of the GAIA-X working groups. By 2022, the technical specifications had been altered to accommodate hyperscaler participation.

By 2024, Frank Karlitschek, founder of Nextcloud and a participant in the early discussions, said publicly that “I think it was actively pushed in [its current] direction by the hyperscalers.” European federated cloud is operationally possible. The project failed because the actors it was meant to displace were admitted as architects.

The inevitability case treats the GAIA-X failure as evidence that the project category is doomed. The actual mechanism is the opposite. The project failed because the refusal was not made. The hyperscalers were invited into the room where the alternative to hyperscaler dependence was being designed. The result was predictable.

GAIA-X is proof that European federated cloud requires the actual refusal of hyperscaler participation in its governance. The actors a sovereignty project is meant to displace cannot be the architects of the project.

ASML is the converse case. The Eindhoven-headquartered firm is the world’s only producer of extreme ultraviolet (EUV) lithography systems, the machines required to produce sub-7-nanometre semiconductors. The company is European-built, European-headquartered and European-listed. By the structural test, ASML is a sovereignty success.

The structural test is incomplete. ASML’s products fall under Wassenaar export-control regimes that give the United States effective veto over which countries ASML can sell to. In 2019, the Dutch government was instructed by Washington to deny export licences for ASML EUV machines to China. The Dutch government complied. In 2024, the same constraint was extended to less-advanced ASML products.

The Dutch industry minister of the day stated that the Netherlands was making the decisions; the actual decision-making was occurring in Washington and being transmitted via export-control coordination. ASML is European in form. It is American-controlled in the operational sense that matters most.

ASML is the cautionary case. The build can succeed and the sovereignty can still fail if the protection mechanisms are not in place. The lesson is that the refusal of inevitability has to include the refusal of capture-by-acquisition and capture-by-export-control as well as the refusal of capture-by-procurement.

The pattern is alive

The historical refusals continue. The pattern is operating now in dozens of smaller acts that together constitute the daily practice of European sovereignty work.

In 2025, the Danish government announced its exit from Microsoft 365 across the Ministry of Digital Affairs. The migration target was openDesk, the open-source productivity suite from Germany’s Zentrum für Digitale Souveränität (ZenDiS).

On 31 October 2025, the International Criminal Court announced the same migration, having spent sovereign budget to escape the chokepoint that Microsoft had operationalised against the court’s prosecutor in February 2025. France launched Tchap and Olvid as sovereign messaging and video conferencing platforms for ministerial use.

Mistral, the French large-language-model company, became operational and began winning enterprise contracts in 2024 and 2025.

The Italian journalist Francesco Cancellato, the Mediterranea Saving Humans co-founders Luca Casarini and Giuseppe Caccia, and Father Mattia Ferrari brought their cases of Paragon Graphite spyware targeting to Italian courts and Citizen Lab, producing the disclosure that brought Italian government complicity to public attention.

NOYB and the Schrems legal apparatus have produced two CJEU rulings (Schrems I in 2015 and Schrems II in 2020) and continued enforcement actions through 2024-2026 that have made US data transfer arrangements operationally precarious. The Norwegian Data Protection Authority advised in February 2025 that European businesses should prepare exit strategies from the EU-US Data Privacy Framework.

ZenDiS exists, employs engineers and ships software. AMI Labs, having moved to the United States in the early 2020s, returned to Europe in 2024. The Nordic asset-management constituency, primarily through the Nordic Blockchain Association and aligned pension funds, has begun building investable instruments for European sovereignty assets in advance of the Sovereign Taxonomy.

Two hundred organisations have signed the EuroStack declaration. Volt Europa carries the political channel. CISPE represents the European cloud sector. The Mistral and AMI Labs and ZenDiS and Bleu and openDesk constellation is collectively a European technology stack in early formation.

These are smaller acts than the historical builds. They are also more numerous. They are evidence that the refusal is in motion.

We have refused before. We are refusing now in smaller ways. We can refuse again at the scale the digital architecture requires.

The Blueprint applies the pattern

The accompanying Blueprint applies the historical pattern of refusal and build to the digital domain. Its rules are not novel speculation. They are the historical mechanisms operationalised across the digital stack: pooled multinational structure, long time horizons, explicit opt-out from the dominant system, clear threat perception, insulated funding.

The Position Framework Rule (Blueprint Rule 1) is the Galileo move applied to data. Name the dependency at every layer of the seven-layer stack and require Position 1 or 2 sovereignty for strategic infrastructure. The Buy European Rule (Rule 2) is the Airbus move. Government mandate creates the demand floor; private capital invests to meet it. The State-Aid Carve-Out (Rule 3) is the EURATOM move.

A treaty-level exemption from ordinary competition law that lets strategic infrastructure scale outside the rules that govern ordinary commerce. The Invest Don’t Subsidise Rule (Rule 4) is the Galileo and Airbus financing mechanism. Equity not grants. Returns recycle. The architecture pays for itself.

The Binding FDI Screening with Mandatory Golden Share Rule (Rule 5) is the protection mechanism that ASML did not have. The Statutory Hostile-Actor Prohibition (Rule 6) is the threat-perception mechanism that the EUCS sovereignty tier would have institutionalised had it not been removed in March 2024. The Algorithm Standards Sovereignty Rule (Rule 7) is the GSM move applied to standards-setting.

Europe writes the standards Europeans use. The Free State-Sponsored Citizen Tools Rule (Rule 8) is the public-broadcasting model applied to digital citizen infrastructure.

The Blueprint operationalises. Paper 26 finances. Paper 27 makes the urgency case. The pattern that built the institutions Europeans now consider permanent is the pattern that builds the digital infrastructure Europeans cannot yet take for granted.

The window is open

The myth is sold daily by people who profit when we believe it. The architecture deepens with each procurement decision made under the myth. The talent leaves with each year European sovereignty work is treated as embarrassing. The political imagination contracts with each commentariat round of “but is this realistic?”

Inevitability is a story that has been refused before. The euro was refused. The Berlin Wall was refused. The Marshall Plan trajectory was refused. The Treaty of Rome was refused. Airbus, Galileo, CERN, GSM, the euro, Ariane, EURATOM. Each began as something the consensus held could not be done. Each is now infrastructure Europeans operate without remembering it was once dismissed.

The myth is sold. We can stop believing it. The pattern of refusal and build is European, repeated, alive.

They sell us inevitability. We believe them. We can refuse again.