Precision matters here, because the comfortable language has been obscuring an uncomfortable reality for too long.
What occurred across Western Europe after 1991 was not underinvestment. It was demolition. The Bundeswehr contracted from nearly half a million personnel at reunification to roughly 180,000 today, with readiness levels that German parliamentary oversight bodies spent years documenting in reports that should have triggered alarm but instead triggered footnotes. The British Army fell below 100,000 regulars — a figure not seen since before the Somme — while simultaneously running down armoured vehicle inventories, artillery stockpiles, and the logistical infrastructure required to sustain anything beyond a short-duration operation. France preserved credible headline figures but quietly gutted the enabling layers beneath them: strategic airlift, air defence depth, munitions reserves. The Netherlands sold off its entire main battle tank fleet in 2011, then spent subsequent years in the awkward position of leasing armour back from Germany once the optics became impossible to ignore.
None of this happened by accident. These were deliberate decisions, taken by consecutive governments across the continent, with full awareness of the trade-offs involved. The fiscal space liberated by defence cuts went into pension commitments, healthcare expansion, and the general architecture of societies that had concluded history had been successfully concluded.
Then came 2014.
Russia’s annexation of Crimea and the subsequent manufacturing of a proxy conflict in eastern Ukraine constituted an unambiguous termination of the post-Cold War order. The NATO response — the Wales Summit pledge committing members to reach two percent of GDP within a decade — communicated something to Moscow, though not what was intended. A decade. Not emergency mobilisation. A decade of gradual adjustment, which the majority of signatories then declined to honour even on those relaxed terms. Germany crossed the threshold only in 2024, under sustained American pressure, two full years after a continental land war had resumed. Belgium, Italy, and Spain remained comfortably below throughout. The signal was clear enough.
The post-February 2022 record is, in some ways, more revealing precisely because it followed an event that eliminated every remaining excuse. The intelligence warnings had been extensive and public. The invasion was not a surprise. And yet the dominant European approach to rearmament has resembled a medium-term procurement review rather than a response to an active continental emergency. The artillery ammunition problem — the central material lesson of Ukraine’s first year — took European industry well over two years to begin meaningfully addressing. American production of 155mm shells scaled from approximately 14,000 monthly to over 100,000 within eighteen months of the invasion. European collective output remained a fraction of that figure throughout the same period.
The explanation lies not in bureaucratic inertia alone but in something more structural: a cognitive dependency that mirrored the financial one.
The working assumption embedded in European defence planning — never explicitly stated because stating it would have required confronting it — was that Article 5 was a delivery mechanism for American power. Not a framework for collective European defence, but access to US extended deterrence, US reinforcement capability, US command infrastructure, US precision strike, US intelligence architecture. European governments were not building a coalition military. They were purchasing political access to an American guarantee at the minimum price Washington would tolerate. NATO, across most of its Western European membership, functioned as a subsidised insurance arrangement — one in which the premiums had been quietly, persistently, and knowingly allowed to lapse.
This is why American disengagement is not a correctable stress on the European security architecture. It is an exposure of the fact that substantial portions of that architecture are American, and cannot be substituted quickly regardless of political will or financial commitment.
The specific capability gaps are not abstract. Theatre-level integrated air and missile defence: the command architecture, the most capable interceptor layers, and the battle management systems are American. Remove them and massed Russian strike packages — the pattern applied to Ukrainian infrastructure repeatedly since late 2022 — would encounter a dramatically degraded response over NATO territory. Intelligence, surveillance, and reconnaissance at operational scale: European contributions are genuine but structurally insufficient to replace American satellite coverage, signals intelligence, and airborne early warning capacity. Nuclear deterrence: France holds an independent deterrent; the United Kingdom’s is entangled with American Trident infrastructure; the remainder of the continent exists entirely within American extended nuclear guarantees. Talk of a European nuclear umbrella describes a political aspiration, not an existing capability. Deep logistics and strategic lift: moving and sustaining large formations across European operational distances requires the kind of pre-positioned, exercised infrastructure that exists in American hands.
Reconstituting these capabilities is measured in years under optimal conditions, and in decades absent genuine emergency footing.
Which is what makes the current Iran conjuncture strategically clarifying.
The proposition that Europe might distance itself from Washington over Iran policy — whatever the substantive merits — presupposes an ability to absorb the defence consequences of doing so. That ability does not currently exist. The nations most directly exposed to Russian pressure are not the ones making autonomy arguments. Finland — which maintained conscription through the entire post-Cold War period and shares over 1,300 kilometres of border with Russia — has not been vocal about strategic independence from Washington. Poland, constructing what will likely be the largest land army in NATO Europe, has not. The Baltic states, which have built territorial defence doctrines explicitly modelled on the lessons of the Donbas and later of the full-scale invasion, have not. These countries understand the gap between declaratory posture and actual capability. Their governments have been building real defence, within their means, for years. But they cannot substitute for American theatre-level assets. They know this.
The political class responsible for this situation operated under a label. It was called smart defence, or efficient burden-sharing, or occasionally, when a degree of candour intruded, strategic patience. Its actual function was a straightforward fiscal transfer: the costs of continental security were externalised onto the United States, and the resulting budgetary room was converted into domestic political programmes. European voters received expanded public services. American taxpayers received the liability. Trump did not manufacture this grievance. He inherited a structural imbalance, recognised its leverage value, and began applying it. The leverage is real because European governments created it, methodically, over three decades.
What genuine remedy requires is not a procurement uplift. It is a mobilisation mindset.
Five percent of GDP on defence is the functional lower bound for rebuilding European military capacity to a level that could perform credible territorial deterrence without American structural support within a relevant timeframe. For reference: West Germany at the Cold War’s height spent around four percent. What confronts Europe now — a peer adversary actively prosecuting a war on the continent’s eastern border, with explicit territorial ambitions and a defence industrial base operating at wartime output — is not the Cold War analogue. It is more acute, and the timeline is shorter.
At five percent, the actual content of the commitment includes: armoured formations with genuine ammunition depth rather than weeks of combat sustainability; layered air defence capable of handling massed drone and missile attack; artillery and the shell production to sustain it beyond the opening phase of a conflict; conscription at meaningful scale, which means accepting that young citizens will serve in uniform for substantive periods; and defence industrial capacity rebuilt for resilience rather than optimised for peacetime efficiency. All of this takes time even with adequate funding — the industrial base requires physical investment, the institutional knowledge has to be recovered, the training pipelines have to be opened. Which is precisely why the correct moment was 2014, and the next available moment is now.
Two honest positions exist for European governments. The first: accept that NATO’s eastern members will operate under a deterrence posture that is partly a bluff, manage the political consequences of that calculation after the fact, and hope the bluff is not tested. This is unstable, but it is at least coherent. The second: remain in a client relationship with Washington, accept whatever terms that relationship currently carries, and trade foreign policy autonomy on questions like Iran for continued access to the security guarantee. Also coherent, and at least honest.
What is not coherent is the current position: conducting foreign policy as though strategic autonomy exists, while maintaining defence expenditure levels that guarantee it cannot, while expecting continued unconditional American security commitments regardless of how those commitments are treated politically. The arithmetic does not work. The capability inventory does not support it. The industrial base does not back it.
European governments have been operating inside a durable fiction — that American commitment to the continent was structurally guaranteed, politically indestructible, and infinitely patient irrespective of European behaviour. That fiction has now been retired. The question that remains is whether Europe’s political leaderships can communicate to their publics what solvency actually costs — in tax rates, in industrial redirection, in the return of conscription, in the acceptance that the posthistorical holiday is over.
The analytical case for treating this as an emergency is closed. Whether the political capacity to act on that case exists is a different question, and a less encouraging one.