European Military Sovereignty

Defence systems running on American components are operable at American discretion.

1. The Spine and the Stakes ===========================

The United States bore most of the financial and strategic burden of European defence and was repaid through trade access, intelligence cooperation, dollar dominance, and the political cohesion of a unified Western bloc.

Europe accepted American military primacy and used the protected space to build a distinctive political and social model: rule of law and human-rights commitments, plural democracy with regulatory limits on power, social rights and welfare, multilateral institutional architecture, and the rule of law as constitutional foundation across all four.

That model is now at stake. In 2025 the alliance partner that protected the space for it began to name it as ideological deviance.

On 14 February 2025 US Vice President JD Vance delivered a speech at the Munich Security Conference. He told European leaders that the biggest security threat to Europe was not Russia or China but the threat from within. He attacked European governments for what he called censorship of free speech. He criticised European migration policy as anti-democratic.

He praised political parties that European democracies had identified as threats to constitutional order. He conditioned American support on European compliance with American positions. European leaders called the speech a wrecking ball, an ideological war, a declaration of culture war against the United States’ European allies.[1]

In January 2026 the Trump administration assembled the Board of Peace, a group of countries invited to participate in Middle East negotiations. France, Germany, the United Kingdom, Italy, Norway, Sweden, Poland, and Spain declined. Saudi Arabia, Turkey, Qatar, Hungary, and Belarus accepted.

Canada was invited and the invitation was publicly rescinded after Prime Minister Carney criticised American economic coercion at Davos. The Board’s working geometry was visible: a coalition of states whose values aligned with the current administration’s positions, with traditional democratic European allies excluded.[2]

In the same months the Trump administration repeated public threats against Denmark over Greenland, did not rule out military force, and made tariff threats explicit. The threats were retracted within weeks but they exist in the historical record. Greenland is European territory under Danish sovereignty. The willingness to put it on the negotiating table is now established.

We cannot remain Europe without being able to defend Europe. We cannot yet.

The Board of Peace was the cheapest test. Eight European democracies declined and the cost of declining was bounded: exclusion from a negotiation. The next tests will not be as cheap.

If American policy conditions F-35 software updates on European compliance with Middle East positions, or NATO intelligence sharing on agreement with American positions on Ukraine, or Patriot missile resupply on alignment in some other policy domain, the cost of holding European values rises sharply.

With current capability Europe does not have the option of holding the values when the cost rises sufficiently. Capability is what turns values into commitments. Without it, values are preferences, held when convenient and surrendered when expensive. The decision window for building capability is 2025 to 2030. What follows is what building it would require.

2. What Makes Europe Europe ===========================

Europe is geographically a continent and politically an institutional project. The institutional project is what is at stake when American policy names European values as ideological deviance, and what would be lost if Europe surrendered the values to retain American protection.

The first is the rule of law and human rights. The European Convention on Human Rights, in force since 1953, binds 46 states and gives individuals direct standing against their governments. The International Criminal Court, founded by European-led negotiations from 1998, holds individual leaders accountable for the gravest international crimes.

The General Data Protection Regulation, in force since 2018, treats personal data as a fundamental right requiring affirmative consent rather than a commodity available to commercial extraction. Due process for migrants, including non-refoulement obligations under the Refugee Convention, governs European treatment of asylum seekers.

The rule of law itself is enforceable through the Court of Justice of the European Union, which has struck down member-state actions that violate constitutional principles.

The second is plural democracy with regulatory limits on speech. The Digital Services Act, in force since 2023, requires very large online platforms to assess and mitigate systemic risks to civic discourse, electoral integrity, and democratic deliberation. The Audiovisual Media Services Directive constrains broadcast content along democratic norms.

National constitutional courts (the German Bundesverfassungsgericht most prominently) have ruled repeatedly that political parties seeking to undermine constitutional democracy can be limited or banned. The European model treats absolute speech freedom as compatible with democracy only when discourse-shaping infrastructure is itself accountable.

The Vance Munich speech named these limits as censorship.

The third is the social rights and welfare model. Universal healthcare across the European Union and adjacent states. Strong labour protections, including collective bargaining rights, paid parental leave, and statutory minimum wage frameworks. Public housing, public education through tertiary level, public pensions.

The social market economy as constitutional principle in the German Basic Law and analogous national provisions across the continent. The model treats market outcomes as legitimate only when bounded by the social commitments that make democracy substantive.

The fourth is multilateral institutional commitment. Membership in the United Nations, in the World Trade Organization, in the World Health Organization, in the International Atomic Energy Agency, and in dozens of other multilateral bodies. Climate policy through the Paris Agreement, with European leadership in the Conference of the Parties process.

International humanitarian law as treaty obligation. Diplomatic recognition of Palestine by most European states. Adherence to ICC obligations. The European model assumes that international order works through institutions, not through bilateral coercion.

These four commitments are not policy preferences. They are constitutional architecture, embedded in treaty law, statute, and institutional practice. Europe is what they are. Take them away and the territory remains. The institutional project does not.

The Vance Munich speech named all four as deviance. American positions on free speech reject European platform regulation. American positions on migration reject European due-process commitments. American withdrawal from the Paris Agreement, the WHO, and the ICC rejects European multilateralism. American social policy rejects the European welfare model as economically inefficient.

The current US administration positions itself in opposition to each of the four root commitments. American support for Europe is now conditional on European compliance with these positions.

Compliance would not preserve Europe. It would dissolve it.

3. The Doctrine Imported ========================

The dependencies under which European militaries operate are not passive. They are doctrine compiled into software, hardware, and procurement architecture. Buying the dependency is buying the doctrine. The doctrine that is being imported has named effects.

On 28 February 2026, the first day of the 2026 Iran war, a US guided weapon destroyed the Shajareh Tayyebeh Elementary School in Minab, Hormozgan province, southern Iran. One hundred and fifty-six people were killed, including one hundred and twenty children. The targeting database had marked the building as part of an Islamic Revolutionary Guard Corps compound.

Satellite imagery from 2013 supported that designation. Imagery from 2016 showed the school had been walled off and given separate civilian entrances with three gated public access points. The targeting data was at least ten years out of date. The strike proceeded anyway.[3]

The data was processed through Palantir’s Maven Smart System, the analytics platform used by the US Joint Force during the 2026 Iran operations. Maven fused nine separate military intelligence systems into a single interface. It compressed the kill chain from hours to minutes. Across the operation the US military struck five thousand five hundred to six thousand targets in Iran.

The first thousand strikes were executed in twenty-four hours. The system enabled twice the daily strike rate of prior operations while relying on significantly fewer human analysts.[4]

The system’s marketed advantage is target discrimination. Independent assessment puts Maven’s classification accuracy at sixty per cent against eighty-four per cent for experienced human analysts. The system is twenty-four percentage points less accurate than the analysts it replaces. Its actual advantage is throughput.

The trade-off is the design choice: faster targeting at lower precision, more strikes per day, higher operational tempo, predictable civilian casualties as the cost of the rate.[5]

In Gaza, the Lavender system gives Israeli officers an average of twenty seconds to authorise strikes on residential buildings before bombing. Lavender produces lists of tens of thousands of targets identified by AI association with Hamas. Officers report that the system’s output is treated as fact rather than recommendation. Civilian casualties have been documented in the thousands.

Common errors include strikes on family homes after a target’s phone was passed to a son or brother.[6]

These are the same institutional pattern in different uniforms. Compressed verification windows. Acceptable civilian-casualty thresholds calibrated to operational tempo. Tempo treated as the goal and accuracy as the trade-off. Stale data accepted as current. The institutional logic is consistent: throughput first, precision second, civilian casualties as the price of the rate.

That doctrine is now compiled into Palantir Gotham and Maven, into the Lockheed Martin mission data files generated at Eglin Air Force Base for F-35 operations, into NATO standardisation agreements that align European targeting protocols with US doctrine, into the analytics platforms procured by European intelligence services and police forces.

European militaries that procure these systems do not just buy software. They buy the institutional logic that says ten-year-old data is sufficient justification to strike a building full of children, and that twenty-four percentage points less accuracy is an acceptable price for higher throughput. Europe holds itself to standards that say it is not.

The dependencies Europe is procuring encode standards that say it is.

The vendor’s position on European values is now public. In April 2026 Palantir’s CEO Alex Karp and head of corporate affairs Nicholas Zamiska published a 22-point manifesto on X. Among its claims: the postwar neutering of Germany and Japan must be undone. Some cultures have produced vital advances; others remain dysfunctional and regressive.

Hard power in this century will be built on software. The manifesto calls for compulsory national service and embraces religion in public life. The post is officially Palantir corporate communication and reached over twenty-one million views within days.

The CEO of the company that runs F-35 mission systems, sells Maven to the US military for kill-chain compression, embeds in the French Direction générale de la Sécurité intérieure and German state police forces, and operates as a CIA-seeded venture, has publicly named the post-war European settlement as ideological deviance to be reversed.

Mark Coeckelbergh of the University of Vienna described the document as technofascism. Yanis Varoufakis warned of advanced AI combined with militarised state power. Buying Palantir is buying from a vendor whose corporate ideology is the explicit reversal of what makes Europe Europe.[7]

This is buying a vendor that does not believe in the values the product is supposed to defend.

4. The Buttons ==============

The dependencies are not passive. Each is an active leverage point. The American government has the technical and legal authority to press them. The history shows the buttons get pressed.

The F-35 Lightning II is the case. Fourteen European countries have committed to approximately 668 aircraft. The programme will define European air power for thirty to forty years. Mission data files telling each aircraft how to identify threats, counter electronic warfare, and engage targets are generated at a classified facility at Eglin Air Force Base in Florida.

Software updates, maintenance diagnostics, and operational capabilities are controlled by Lockheed Martin and the US Department of Defense. European partner nations do not have access to the source code.

The Operational Data Integrated Network, replacing the earlier Autonomic Logistics Information System, transmits flight data, engine diagnostics, radar signatures, electronic warfare intercepts, and mission records back to servers managed by Lockheed Martin in the United States. Every sortie reports back. The country that provides the capability also observes its use.[8]

In February 2026 the Dutch Defence Secretary publicly warned that European nations might need to jailbreak F-35 software if the US withheld upgrades amid deteriorating transatlantic relations. The word was chosen deliberately. It is the language of a consumer trapped inside a proprietary ecosystem, applied to the most advanced combat aircraft in European service.[9]

Turkey demonstrated what happens when the lock is tested. In 2017 Turkey signed an agreement to purchase the Russian S-400 air defence system. The United States warned that operating a Russian radar alongside the F-35 was incompatible. When Turkey took delivery of the S-400 in July 2019 the US removed Turkey from the F-35 programme within five days.

Turkey had ordered 100 aircraft and paid approximately 1.4 billion dollars. Turkish companies manufactured over 900 parts for every F-35 in production. The US Department of Defense estimated retooling at 500 to 600 million dollars. Turkey made a sovereign procurement decision and was ejected from the most consequential Western military programme of the twenty-first century within a working week.[10]

The other buttons follow the same logic. International Traffic in Arms Regulations gives the US government veto authority over any satellite or weapon system containing American components. Most current European military systems carry US components and are therefore subject to unilateral US veto.

Patriot missile resupply was throttled to Ukraine in 2024 to 2025 as American policy reassessed support. NATO intelligence sharing was conditioned on alignment with American positions on Ukraine. GPS M-Code access can be restricted to US-only use or degraded for non-aligned users.

Starlink battlefield communications were denied to Ukrainian operations in Crimea in September 2022 by Elon Musk’s personal decision. Drone boats lost connectivity, drifted, and washed ashore. Communications infrastructure that European militaries increasingly depend on is now demonstrated to be subject to a private American citizen’s risk assessment.

In September 2025 the NATO Communications and Information Agency selected Oracle Cloud Infrastructure for mission-critical workloads across the alliance’s command, control, communications, computers, intelligence, surveillance, and reconnaissance systems. Oracle is a Delaware corporation. The CLOUD Act applies to Oracle regardless of where the data is held.

The selection means that NATO C4ISR runs under US legal jurisdiction. NATO is a multi-jurisdictional alliance and US-jurisdiction infrastructure has a defensible role in NATO-collective workloads.

The structural problem is that European-only operations, including national defence functions and European-pillar-of-NATO actions that do not include US forces, currently route through the same infrastructure. An alliance of thirty-two nations, defended by the cloud of one.[11]

The British nuclear deterrent demonstrates the dependency at the highest layer. The United Kingdom holds title to 58 Trident II D5 missile bodies, kept in a communal pool at the Strategic Weapons Facility at Naval Submarine Base Kings Bay in Georgia. Maintenance is performed at Kings Bay periodically. Many components, including warhead aeroshells and tritium reservoirs, come from US suppliers.

Operationally the Prime Minister can authorise launch independently, including against the United States. Logistically the system depends on continuing American cooperation. A 2014 British American Security Council report concluded that without US cooperation UK nuclear capability would have a life expectancy measured in months.

The 2025 Northwood Declaration expanded UK-French nuclear coordination short of system migration. The UK government formally ruled out switching to French M51 missiles in early 2025. The political environment has shifted significantly since.[12]

Europe could have built much of this. Operation Dunhammer (Paper 6) documented that the US National Security Agency conducted multi-year intelligence collection against Eurofighter, Saab, and the Danish defence company Terma during Denmark’s competition to replace its F-16 fleet. Lockheed Martin’s F-35 won.

The fact that the United States needed to spy to win the competition is the answer to whether European industry could have built it. If Eurofighter and Saab had been incapable, no spying would have been required.[13]

5. The Capability Gaps and the Build Base =========================================

The capability gaps Europe faces are concrete and quantified. NATO’s own assessments confirm dependence on US enablers across the major strategic categories. The United States operates 246 military satellites; European NATO members manage 49. The United States fields approximately 150 aerial intelligence-surveillance-reconnaissance aircraft; European NATO has dozens.

The US has over 900 ISR uncrewed aerial vehicles to Europe’s roughly 200. Airborne command-and-control platforms run 120 to 35. Air-to-air refuelling tankers stand at 450 to 150. The estimated time for Europe to reach self-sufficiency on space-based ISR is five to ten years.

Battlefield command-and-control, long-range strike, and suppression of enemy air defences are five-year programmes if pursued seriously.[14]

The gaps are bounded. They are not technologically impossible. They are procurement-and-investment problems with calculable answers.

The build base exists. National armies remain Europe’s largest military assets, with combined active personnel of around 1.4 million across European NATO members, around 16 million reserves where mobilisable, and substantial industrial capacity across all major combat domains.

National defence budgets reached approximately 280 billion euros in 2024 across European NATO and are projected to rise above 480 billion euros by 2026 under ReArm Europe acceleration.[15]

France maintains the only fully autonomous strategic deterrent in Western Europe. The force de frappe comprises four Le Triomphant-class ballistic missile submarines, M51 missiles built by ArianeGroup, and warheads designed and produced in France. The full supply chain is sovereign. Approximately 290 deployed warheads.

In March 2025 President Macron opened a strategic dialogue with European partners on the role of French deterrence in collective European defence. This was the first time a French president has formally offered nuclear umbrella discussions to European allies.

The offer is the most consequential structural development in European nuclear policy since France’s withdrawal from NATO’s integrated command in 1966.[16]

The United Kingdom maintains an operationally autonomous deterrent with the dependency limitations described above. The 2025 Northwood Declaration commits the UK and France to deeper nuclear coordination including cross-secondment of officers, joint patrols, and aligned doctrine short of system migration. The Lancaster House framework (2010, expanded 2025) provides the broader bilateral basis.

The 2010 Teutates Treaty already shares a sub-critical nuclear testing facility in France.

The conventional defence-tech sector has produced European companies at venture scale. Helsing reached a 12 billion euro valuation in 2025 as Europe’s leading frontier defence-AI company, with Bundeswehr and other European militaries as anchor customers. Mistral reached a 6.2 billion dollar valuation as Europe’s frontier foundation-model company.

ICEYE operates 62 synthetic-aperture-radar satellites with a 2.4 billion euro valuation, providing imagery used by Ukraine to identify thousands of Russian vehicles. Quantexa builds enterprise analytics with growing defence and intelligence applications. Unseenlabs provides French-developed satellite electronic warfare.

The kernel of capability for sovereign defence analytics, satellite ISR, and frontier AI exists.[17]

Major joint programmes are in motion. The Future Combat Air System (France, Germany, Spain) develops a sixth-generation fighter intended to enter service in the 2040s. The Global Combat Air Programme (United Kingdom, Italy, Japan) pursues a parallel sixth-generation programme. The Main Ground Combat System (France, Germany) develops a next-generation tank to replace Leopard 2 and Leclerc.

The A400M Atlas provides multinational strategic airlift. The A330 MRTT provides multinational air-to-air refuelling. Programmes have political friction but they exist and they work when the political will is sustained.

Non-EU European integration deepens steadily. The United Kingdom was invited in November 2022 to participate in the Permanent Structured Cooperation Military Mobility project. Switzerland was invited in 2025 to join Military Mobility, marking its first significant European defence integration. Norway, Serbia, and Ukraine hold cooperation agreements with the European Defence Agency.

The 2025 European Defence Fund work programme allows Ukrainian entities to participate for the first time. The architecture of European defence is increasingly continental rather than EU-bounded.[18]

The financial commitments are unprecedented. ReArm Europe, endorsed by EU leaders in March 2025, mobilises 800 billion euros in defence spending. SAFE provides 150 billion euros in concessional loans for joint procurement. Germany committed 35 billion euros to military space spending by 2030 in September 2025, alongside constitutional debt-rule loosening to enable defence expansion.

France committed an additional 4.2 billion euros to military space for 2026 to 2030 in November 2025, with overall defence spending rising to 64 billion euros annually by 2030. Italy’s IRIDE earth-observation constellation has 1.1 billion euros committed through Italy’s National Recovery and Resilience Plan. National-track capacity is real and rising fast.[19]

The build base exists. The capability gaps are bounded. The political moment has produced unprecedented financial commitments. What is missing is the structural framework within which all of this can be assembled into a coherent European military sovereignty.

6. Defending Without Empire ===========================

Capacity is not the constraint. Choice is. Europe has the engineering, financial, and industrial capacity to build almost any military capability the United States builds. The question is which capabilities are necessary for defending Europe and which capabilities, in being built, would make Europe something other than Europe. Force-projection capability at imperial scale is a values choice.

Building it would import the institutional habits of empire that Europe has spent eighty years moving past.

The American military-industrial complex grew through two and a half centuries of near-continuous conflict. The United States has been at war for two hundred and twenty-five of the two hundred and fifty years of its existence. Its institutional habits reflect that history. A defence industrial base whose employment serves as regional economic policy.

A national-security state with significant classified domains outside ordinary oversight. A doctrine of forward-deployed force, expeditionary readiness, multi-theatre projection. An intelligence community embedded in domestic surveillance through the FISA, FBI, and Department of Homeland Security architectures. Targeted-killing programmes, mass surveillance, drone-strike doctrine.

Defence employment as a regional economic policy. Defence exports as a fiscal strategy. Defence research and development as the engine of broader American innovation. The American MIC has no cap. It expands as long as it can find the political constituencies to justify the spending. The spending sustains the constituencies. The constituencies sustain the spending.

Each generation of the loop makes the next harder to break.

Building a European defence capability that imports these institutional habits would import the problems that come with them. Some of the habits are not values-neutral choices.

Mass surveillance of civilian populations, automated kill-chain compression below meaningful-human-control thresholds, civilian-data ingestion at scale, defence-export-as-fiscal-strategy that requires selling to questionable regimes: these are not different American preferences. They are values violations under European constitutional law and under international humanitarian law.

The European model needs a defence industrial base. The question is what kind. Five constraints distinguish a defence base built in the European mode from one that replicates the American MIC.

Civilian primacy and democratic authorisation. The decision to delegate any portion of the kill chain to algorithms is constitutional. Whether to have automated targeting at all, under what conditions, with what safeguards, belongs to legislatures, courts, standing expert panels, and the public, not to defence ministries operating in classified domains.

Parliamentary oversight is necessary but not sufficient.

The European model requires multi-layer oversight:

constitutional or treaty-level prohibitions on absolute limits including meaningful human control over lethal force;

parliamentary approval requirements for AI systems above thresholds;

standing independent expert panels with technical and legal capacity;

judicial review for compliance with international humanitarian law;

civilian operational inspectors-general;

civil-society institutions including data protection authorities;

and international standards including extension of the EU AI Act risk-classification regime to defence AI, which is currently exempted under national-security provisions. Each layer catches what the others miss.

The exemption itself is a values choice that the political moment requires reversing.

Doctrine and posture. Defensive territorial focus, not expeditionary force-of-choice. International humanitarian law embedded at all command levels with judicial review. Regional rather than global deployment. Refusal of pre-emptive policing, of mass surveillance integration with civilian populations, and of automated kill-chain compression below meaningful-human-control thresholds.

Political economy. Always capable, not always at war. The American MIC is perpetually mobilised because perpetual conflict justifies it.

The European model should be perpetually capable:

sustained baseline investment in long-lead capabilities so that the build base exists before the threat materialises;

scaled programmatic decisions for shorter-cycle commitments tied to threat assessment;

industrial-layer reforms that prevent capture;

high cap on defence-spending share of GDP with explicit declared-crisis surge provisions for crossing the cap;

sunset reviews on specific capability commitments tested against current threat.

The build-time problem is real. Sovereign satellite ISR is a ten-year programme.

Fighter platforms are fifteen-to-twenty-year cycles. These cannot be conditional on annual threat assessments. The capability has to exist before the threat. Always ready, but not always at war.

Civil-military separation. No mass-surveillance integration with civilian populations. Separation of military intelligence from civilian policing. Defence-AI procurement decisions removed from defence-ministry exclusive authority and embedded in the multi-layer oversight architecture above. Privacy and data-protection authorities given role in defence-AI oversight, not just civilian.

Capability built in the European mode. Defence research and development as supplier from the broader civilian innovation base where civilian alternatives are competitive, while the driver of European innovation lies elsewhere.

Foundation AI models, autonomous systems, cyber capabilities, communications technology, materials, and sensors procured from Horizon Europe, the Sovereign Tech Fund, civilian sovereign-fund vehicles, and national R&D agencies, with defence-specific R&D continuing only where civilian alternatives do not exist (fighter platforms, naval combat systems, submarine systems, specialised munitions).

The American MIC drives American innovation through defence procurement.

Europe should innovate through other channels and let defence procure from that ecosystem.

These five constraints are themselves a values commitment. A military that is civilian-overseen, defensive, not perpetually mobilised, separated from civilian surveillance, and procuring from a broader civilian innovation base is a different kind of military. It reinforces the values it is built to defend.

A military built without the constraints, however capable, becomes the institutional habits of the country that built it. Europe building defence without these constraints would import the habits of empire and become what we are defending against. The constraints are not decorative. They determine what Europe becomes by 2050.

7. What to Build ================

What follows is a structural framework. What follows does not recommend specific weapons systems, specific deployments, or specific operational doctrines. Those decisions belong to military professionals.

What citizens and policy people can address is the framework within which those decisions are made: how defence is procured, how it is financed, how it is governed, what dependencies are accepted, what doctrines are imported through them.

The argument is that the framework Europe currently uses is producing the wrong outcomes regardless of which specific decisions military professionals make within it. Changing the framework changes what military professionals can do. The decisions that operate within the new framework remain theirs.

Five reform areas constitute the framework.

Industrial reform

The European Defence Fund’s 8 billion euros of grants for collaborative research and prototype development through 2027 should convert to an equity-investment model. Returns flow to the EU budget pro rata to contribution. Defence companies must be commercially viable rather than subsidy-dependent. The fund grows over time rather than being annually re-allocated.

The mechanism is the same as the In-Q-Tel model that produced Palantir, where an approximately two million dollar seed-stage investment in 2003 to 2005 anchored a company now valued at approximately 250 billion dollars. The model is well-understood and produces extraordinary returns when the structural fit is right.

SAFE’s 150 billion euros in concessional loans for joint procurement should preserve the loan model, which is appropriate for procurement smoothing. Loan-based financing of joint procurement is structurally different from grant-based financing of capability development. Each has its place.

Juste retour has no place at the defence-industrial layer.

The OCCAR global-balance rule, the consortium-formation requirements that drive the European Defence Fund, and similar mechanisms that allocate work geographically rather than by capability fragment European industrial competitiveness in the same way they have fragmented European space industrial competitiveness (Paper 17).

Capital should flow to the most efficient producer. Returns should flow back proportional to capital contribution.

National defence procurement remains national. The reform applies to the industrial layer that supplies national militaries, not to the militaries themselves. Position 1 sovereignty for the military function is preserved. Position 2 industrial integration is the goal at the industrial layer.

Demand floor

Procurement rules across regulated defence operations create the market floor that European defence companies need to be commercially viable.

Defence ministry procurement above value thresholds should require sovereign source as a procurement criterion: critical military communications must be European-jurisdiction; critical military analytics must be European-owned; targeting-relevant data infrastructure must be European-jurisdiction.

Galileo Public Regulated Service should be the default for European military positioning, navigation, and timing, with dual-mode receivers that maintain GPS interoperability. The Safran Skylight receiver, in production with sixteen thousand test cases completed under the European GEODE programme, combines GPS M-Code and Galileo PRS in a single dual-mode unit.

The mandate does not replace GPS; it adds Galileo as cryptographically authenticated redundancy. When GPS is jammed by Russian equipment costing thousands of euros across European airspace, Galileo continues. When GPS is spoofed, Galileo’s Open Service Navigation Message Authentication catches the spoof and the receiver rejects the false signal.

When both signals jam, inertial navigation continues with high-confidence degradation flag. The mandate is procurement-rule, with no violation of the EU-US GPS/Galileo Agreement, which restricts only civilian-signal mandates and not military procurement decisions.

NATO procurement is a multi-jurisdictional matter where the alliance includes the United States. NATO collective workloads can defensibly use US-jurisdiction infrastructure. The structural ask is to carve out European-pillar workloads onto European-jurisdiction infrastructure: French national defence functions, German national operations, European-pillar-of-NATO actions that do not include US forces. The same alliance can run multi-jurisdictional cloud at the collective layer and European-jurisdiction cloud at the European-pillar layer. The current architecture does neither.

Capability gaps

The capability gaps named in Section 5 require concrete programmatic responses. The categories are well-defined and the kernel of capability exists across most of them.

Sovereign defence analytics is the most consequential. The European version of the In-Q-Tel-and-anchor-customer model applies. A European venture vehicle, structured as the EDF in equity form or as a parallel instrument, takes early-stage equity in Helsing, Quantexa, Mistral, and other European defence-analytics companies.

European intelligence services and militaries provide the anchor-customer commitment with initial procurement contracts in the tens of millions, scaling. France goes first, given political alignment with the Macron strategic-autonomy doctrine. Germany follows under the 35 billion euro military space mandate. The United Kingdom follows under the Northwood Declaration cooperation framework.

Three to five years to first deployments at scale, five to ten to alternative capability at scale.

The key design constraint is that the European venture is a different Palantir. The platform should be designed from the start to do the necessary functions and refuse the harmful ones.

Multi-source intelligence integration, force-protection analytics, counter-terrorism investigation tools, cyber-defence analytics, fraud and corruption detection, open-source intelligence aggregation: these functions are operationally legitimate and commercially supportable.

Mass surveillance of civilian populations, automated targeting with sub-human-analyst accuracy, throughput-over-precision doctrine, kill-chain compression that compresses human review windows, pre-emptive policing of civilian populations, domestic political surveillance, cross-data-domain civilian profiling: these are not different American preferences. They are values violations.

The European platform should refuse them as a structural commitment, not an aspiration. The technical architecture should embed audit trails, human-review-window minimums, accuracy floors, and doctrine-of-use restrictions. Procurement contracts should include clauses preventing the institutional habits Palantir has imported elsewhere.

The customer base for the European platform is European intelligence services, European militaries, European border agencies, Europol, and national counter-terrorism agencies. These services are themselves the subject of the next paper in this series at the intelligence layer.

The operational-military dependency connects directly to the analytical-intelligence dependency (Paper 19). The same European venture vehicle, the same anchor-customer mechanism, builds capability across both. The institutional design that distinguishes European-Palantir from American-Palantir operates at both layers.

The other capability gaps are bounded category problems. Sovereign satellite intelligence-surveillance-reconnaissance is a ten-year programme cross-linked to the implementation set out elsewhere in this series at the space-sovereignty layer. Sovereign command-and-control infrastructure is a three-year programme as the European-pillar carve-out from NATO mission-critical workloads.

Sovereign battlefield communications require the IRIS² constellation accelerated where possible, with procurement-mandate for European source on critical military comms in the interim.

Strategic enabler closure (air-to-air refuelling through A330 MRTT scaling, strategic lift through A400M, suppression of enemy air defences) is a multi-billion-euro multi-year programme that fits inside ReArm Europe’s allocated capacity.

Integrated air and missile defence requires consolidation of IRIS-T, Aster, and Meteor under a European framework rather than fragmented national procurement. ITAR-free supply chain is the IRIS² model applied across military supply chains generally (Paper 17).

National sovereignty preserved

Position 1 sovereignty is preserved for national militaries. National command, doctrine, force structure, and deployment all remain national. The reforms apply to the industrial layer, the demand floor, the procurement framework, and the institutional design, not to the military structure itself.

Coordination occurs through existing frameworks. The Macron strategic dialogue offer extended in March 2025 provides the political mechanism for nuclear-cooperation expansion among interested European partners. The Northwood Declaration provides the UK-French framework. The Lancaster House Treaties (2010, with 2025 expansion) provide the broader bilateral basis.

The Permanent Structured Cooperation framework, the European Defence Agency, and the European Defence Fund provide multilateral architecture. Existing joint programmes (FCAS, the Global Combat Air Programme, MGCS, A400M, A330 MRTT) continue with reformed financing. They compete for anchor-customer commitments rather than receive subsidy.

Whichever programme wins major procurement commitments scales. Whichever does not, does not. The model that produced SpaceX produced Ariane 6’s competitors. Ariane 6 did not get that model.

The non-EU integration that has accelerated since 2022 should expand. UK Military Mobility participation (invited November 2022) deepens through ongoing UK-EU defence-cooperation discussions. Switzerland’s 2025 invitation to join Military Mobility expands. Norway, Serbia, and Ukraine retain European Defence Agency cooperation.

Ukraine participates in the 2025 European Defence Fund work programme. The architecture of European defence is continental, including democratic European states whether or not they are EU members.

The United Kingdom nuclear logistics question deserves explicit treatment. UK Trident is operationally autonomous and logistically dependent on the United States. The 2014 British American Security Council assessment that UK nuclear capability would have a life expectancy measured in months without US cooperation has not been refuted by subsequent events.

The UK government formally ruled out switching from Trident to French M51 missiles in early 2025. The political environment has shifted significantly since. The Vance Munich speech, the Iran war, the Palantir kill-chain operations, and the broader sovereignty crisis have altered the calculation that produced the original ruling.

The migration is technically achievable on a Dreadnought-successor timeline. The crisis condition that would force reconsideration exists in latent form.

Europe should plan for both contingencies: a continued US-dependent Trident with domestic UK maintenance redundancy buildout to extend the months-of-life-expectancy to years, and a contingency UK migration to French logistics or joint UK-French next-generation platform on a 2050s decision timeline. The 2025 Northwood Declaration provides the political framework for either path.

Institutional design

The institutional design is the European pillar of NATO operating as the European pillar should: capable, alternative-having, value-aligned, sustainable. The alliance gets stronger when its European pillar is real. The constraints are the difference between defence capability that defends what Europe holds dear and defence capability that imports what Europe rejects.

8. The Next Test ================

Greenland is the test case where the threat horizon stretches longest. Direct invasion today is unlikely; it would trigger NATO Article 5 and shatter the alliance.

Coercion today is plausible: economic pressure on Denmark, political pressure on Greenland’s autonomous government, information operations on Greenlandic public opinion, expansion of US military footprint at Pituffik Space Base, mineral and Arctic-resource leverage.

Information coercion through US-controlled platforms, including the kind of sustained influence campaign that could shape Greenlandic public opinion toward independence and association with the United States, is genuinely plausible over the next decade.

The medium and long horizons matter more. By 2050 Arctic ice melt will have transformed Greenland’s strategic value. The Northwest Passage and Northeast Passage become navigable for substantial portions of the year. Direct Europe-Asia shipping shaves weeks off current routes. Arctic mineral and energy resources become economically extractable at scale.

Strategic naval positioning shifts north. Greenland’s value to any major power rises by an order of magnitude in the next twenty-five years.

The procurement decisions Europe makes now operate on the same horizon. F-35s ordered today are in service through 2060. The Dreadnought-class submarines being designed today are in service through 2070. The strategic environment in which those platforms operate includes a transformed Arctic with much higher stakes for Greenland. Procurement is policy made for forty years.

The Arctic-melt scenario means the threat horizon for Greenland coercion (or worse) extends well beyond current alliance assumptions.

The decisions of 2025 to 2030 determine which scenarios Europe can defend against in 2035 to 2050. The Vance Munich speech, the Board of Peace, the Greenland threats, the Iran war, the Palantir kill-chain operations, and the Karp manifesto are not separate events. They are the same political development at different points in 2025 to 2026. The development is continuing.

What is at stake is whether Europe remains Europe. The values that distinguish the European model, the institutions that embed those values in law and treaty, the constitutional architecture that makes them commitments rather than preferences: these depend on the capability to defend them when defending them is costly. The Board of Peace cost nothing to decline.

The next test will not be that cheap. The capability decisions made now determine whether Europe can hold its values when the cost rises sufficiently.

Bruno Le Maire has named the European space procurement mistakes publicly. Mario Draghi has recommended the procurement reforms in detail. Friedrich Merz has named the dependency in the bluntest terms in Munich. President Macron has opened strategic dialogue on nuclear cooperation with European partners. Eight European democracies declined the Board of Peace.

Germany has committed 35 billion euros to military space. France has committed an additional 4.2 billion euros to military space. Spain funds PLD Space directly. Italy funds IRIDE through national channels. The European Launcher Challenge runs outside juste retour. Twenty-two member states have separately decided that the existing procurement framework does not produce competitive capability.

What Europe has not done is coordinate the decisions into a coherent European military sovereignty. The diagnosis is settled. The proof of concept is operational. The political alignment exists at the institutional top. The financial commitments are unprecedented. The kernel of capability exists across the major dimensions. The non-EU integration is deepening. The values architecture is in place.

The implementation has not happened.

We cannot remain Europe without being able to defend Europe. We cannot yet. The question is whether we will. The window is now.

[1] JD Vance, Vice President of the United States, address to the Munich Security Conference, 14 February 2025. Full transcript published on the White House website. European leaders including German Foreign Minister Annalena Baerbock and Munich Security Conference Chair Christoph Heusgen issued sharply critical responses.

[2] The Board of Peace, Middle East negotiation framework launched January 2026. Coverage in Reuters, Politico Europe and Le Monde. France, Germany, the United Kingdom, Italy, Norway, Sweden, Poland and Spain declined to participate; Saudi Arabia, Turkey, Qatar, Hungary and Belarus accepted.

[3] Strike on Shajareh Tayyebeh Elementary School, Minab, Hormozgan Province, Iran, 28 February 2026. Casualty figures reported by Iranian Red Crescent and confirmed by independent monitoring organisations including Airwaves and Action on Armed Violence.

[4] Palantir Maven Smart System, US Department of Defense Maven analytics programme. Operational use confirmed in DoD briefings and reporting in The Washington Post, Reuters and 404 Media, 2024-2026.

[5] Comparative classification accuracy assessments of automated targeting versus experienced human analysts in operational contexts. See US Government Accountability Office reports on AI-enabled targeting, 2024-2025; independent analyses by Stop Killer Robots and the International Committee for Robot Arms Control.

[6] +972 Magazine and Local Call investigation, ‘‘Lavender’’: The AI machine directing Israel’s bombing spree in Gaza, April 2024. Reporting by Yuval Abraham documenting Lavender’s operational use and the twenty-second authorisation window observed by Israeli intelligence sources.

[7] Palantir Technologies post on X (formerly Twitter), 18 April 2026. 22-point manifesto adapted from The Technological Republic: Hard Power, Soft Belief, and the Future of the West by Alex Karp and Nicholas Zamiska, 2025. Reached over 21 million views within days.

[8] F-35 Lightning II European partner nation aggregate orders, 2025. Approximately 668 aircraft committed across the United Kingdom, Italy, Netherlands, Norway, Denmark, Belgium, Poland, Finland, Germany, Romania, Czechia, Greece, Switzerland and Israel.

[9] Dutch Defence Minister statements, February 2026, reported in NRC Handelsblad, De Volkskrant and Reuters. Comments referenced the dependency of F-35 operational software updates on US authorisation.

[10] Russia-Turkey S-400 procurement contract signed 12 September 2017. US suspension of Turkey from the F-35 programme announced 17 July 2019, six days after S-400 deliveries to Turkey began. Turkey had ordered approximately 100 F-35A aircraft and paid approximately 1.4 billion US dollars in milestone payments.

[11] 1958 US-UK Mutual Defence Agreement and 1963 Polaris Sales Agreement, as amended for Trident II D5. UK Trident missile bodies are drawn from a communal pool with US Navy at the Strategic Weapons Facility Atlantic, Naval Submarine Base Kings Bay, Georgia. UK title approximately 58 missile bodies.

[12] British American Security Information Council (BASIC), Trident Commission, Final Report, July 2014. Concluded that the operational viability of the United Kingdom’s Trident system absent US cooperation would be measured in months rather than years.

[13] Emmanuel Macron, President of the French Republic, address to the nation on European strategic autonomy, 5 March 2025. Opened formal dialogue with European partners on the role of French nuclear deterrence in collective European defence.

[14] Helsing GmbH Series D funding round, June 2025. Approximately 600 million euros raised at a post-money valuation of approximately 12 billion euros. Lead investor Prima Materia.

[15] Mistral AI SAS Series C funding round, September 2025. Approximately 1.7 billion euros raised at a post-money valuation of approximately 6.2 billion US dollars.

[16] Federal Ministry of Defence (Germany), military space spending commitment of approximately 35 billion euros through 2030, announced September 2025. Routed through national procurement to German firms including OHB, RFA, Isar Aerospace and HyImpulse.

[17] NATO and Center for Strategic and International Studies space inventories, 2024-2025. Aggregate satellite counts: United States approximately 246; European NATO members approximately 49 across reconnaissance, signals intelligence, communications and navigation.

[18] ICEYE Oy, synthetic-aperture-radar constellation operations and Series D funding, 2024-2026. Approximately 62 SAR satellites in orbit; valuation approximately 2.4 billion euros.

[19] Council of the European Union, Permanent Structured Cooperation (PESCO) Council Decision, 2025. Switzerland invited as a third state to participate in the PESCO Military Mobility project, the first significant Swiss participation in EU defence cooperation.