NATO Is Already Over. Nobody Has Said So.

The treaty persists. The credibility that made the treaty operative does not. The replacement is being built — and the Europeans began the work in 2017.

The North Atlantic Treaty is short. Its operative passage, the famous Article 5, is fifty-two words long. The words commit signatories to consider an armed attack on one member as an attack on all and to take “such action as it deems necessary, including the use of armed force” in response. The text has been static since 1949. The treaty has been invoked exactly once — after September 11, 2001 — and the invocation, however symbolic, was a real institutional act with real institutional consequences.

The text is not the alliance.

What made NATO operative was the credibility structure around the text, not the text itself. The credibility had specific components. American military preponderance — not merely capability but the will to use it — was the first. American institutional continuity across administrations — the assumption that what one administration signed, the next administration would honor — was the second. American elite consensus on the alliance’s strategic value — broadly bipartisan, broadly held by the relevant career foreign-policy professionals — was the third. American political stability — the absence of regime-level uncertainty about the country’s basic strategic posture — was the fourth.

These four conditions were what converted fifty-two words into a security architecture that organized the western half of the international system. The treaty was the legal instrument. The credibility was the operating system. When credibility decays, the legal instrument becomes ceremonial.

NATO is not the treaty. It is the credibility. When the credibility goes, the treaty is what remains, and what remains is decoration.

This is the analytical move worth making early in the argument, because the conventional commentary on NATO continues to treat the legal instrument as the substance. The summits happen. The communiqués are issued. The defense ministers meet. The ritual proceeds. The conventional coverage treats the ritual as the alliance. The ritual is not the alliance. The ritual is what the alliance does. The alliance is what makes the ritual operative.

The credibility structure that made the ritual operative has been decaying since 2014, accelerating since 2017, pausing under the Biden administration without resetting, and resuming at scale since the beginning of 2025. By mid-2026, the question is not whether the credibility has decayed. The question is whether the rituals have any operational content left in them, and the answer the European institutions are arriving at — by their actions, not their statements — is that the operational content is now substantially elsewhere.

The Decade-Long Crack

The decay did not begin in 2025. It began in 2014, with the Russian annexation of Crimea and the American-European response. The response was institutionally proportionate to the legal instrument and institutionally inadequate to the security challenge. Sanctions were imposed. Diplomatic protests were issued. The alliance held the line of no NATO troops engaged. The Russian operation succeeded. The territorial fait accompli was absorbed by the international system. The lesson the alliance drew from 2014 was that the cost of resisting territorial revision was higher than it was willing to pay. The lesson Moscow took from 2014 was the same lesson, read from the other side of the ledger.

The Europeans began their internal reassessment quietly. The reassessment accelerated in 2017.

Emmanuel Macron took office in May 2017 and delivered the Sorbonne speech on European sovereignty in September of that year. The speech is read in retrospect as one of the foundational documents of European strategic autonomy. The vocabulary it introduced — sovereignty applied to European institutions, strategic autonomy applied to European security, European army discussed as an operational rather than rhetorical possibility — was unprecedented in postwar European political discourse. The audience for the speech was domestic French and European. The actual reader the speech was constructed for, by Macron’s own subsequent acknowledgment, was the American executive who had taken office four months earlier.

The 2017–2020 period was the discovery phase for European architecture. PESCO — Permanent Structured Cooperation — was activated in December 2017, twenty-five member states binding themselves to defense-integration commitments that had been dormant since the Maastricht Treaty’s 1992 provisions. The European Defence Fund was established in 2017 and operationalized in 2021. The EU Global Strategy of 2016 was reread in 2017 as the foundation for a security architecture that did not assume continued American leadership. The work was procedural, technocratic, and unhurried. It was also continuous.

In November 2019, Macron told The Economist that NATO was experiencing “brain death.” The statement was reported in the American foreign-policy press as a Macron provocation, an entry in the running French complaint about American leadership. The reading was wrong. The statement was a diagnosis. The diagnosis was that the alliance’s strategic mind — the shared analysis of threat, the coordinated planning, the integrated political-military judgment — was no longer functioning at the level the security environment required. Macron was naming what the European institutions had spent two years preparing for.

The Europeans were not naive about American reliability. They have been preparing for American unreliability since 2017.

This is the heterodox claim the American foreign-policy commentary has not fully metabolized. The conventional reading of European behavior between 2017 and 2024 has been that the Europeans were dilatory — slow to spend on defense, slow to act on their own strategic autonomy rhetoric, dependent on American leadership and unable to wean themselves from it. The conventional reading reads European hesitation as European weakness. The hesitation was something else. It was sequencing.

The Biden administration’s 2021–2025 term was the test case for the European strategic-autonomy project. The administration’s posture — reassuring on alliance commitments, active on Ukraine support, traditional in its foreign-policy personnel and institutional practice — gave the Europeans the option to demobilize their strategic-autonomy work. Most American commentary assumed they had. They had not. The strategic-autonomy track slowed during the Biden period; it did not stop. The architecture that had been built in 2017–2020 was maintained, refined, and quietly extended. The European institutions had read the 2017 experience as a forward indicator, not a one-time event. They were preparing for the contingency of a return.

The return arrived in January 2025.

2025

The pattern of 2025 and the first half of 2026 has been continuous and structural. The American executive has treated NATO as a transactional arrangement rather than an alliance commitment. The transactional language has been explicit: members who have not met defense-spending targets are described as delinquents; the commitment to mutual defense has been conditioned on payment; the alliance has been reframed as a debt-collection operation in which American security guarantees are services rendered against unpaid bills.

The pattern of intelligence-sharing disruption has been the more concrete signal. Intelligence sharing — particularly with Ukraine, but also bilaterally with allies whose interests have diverged from the administration’s positions — has been used as a coercive instrument. The disruption is reversible in principle; in practice, the use of the instrument has been the demonstration that the instrument is available, and the demonstration is what changes the calculation of allies who had assumed the intelligence relationship was a constant.

The pattern of conditional language attached to mutual-defense commitments has been, perhaps, the most consequential. The executive’s public statements have repeatedly tied American defense of allies to those allies’ compliance with specific demands — defense spending, trade postures, immigration cooperation, particular procurement decisions. Each conditional statement is an Article 5 credibility hit. The hits are not cumulative in a linear sense. They are cumulative in the sense that they alter the underlying assumption that allies make about American commitment, and once the assumption is altered, the alliance is operating on different terms even if the legal instrument remains unchanged.

The military-deployment posture in Europe has shifted. Specific troop reductions have been announced and partially implemented; specific basing decisions have been reopened; specific exercises have been canceled or curtailed. The cumulative effect is a reduction in the operational presence that, more than any treaty text, had constituted the alliance’s deterrent posture. The deterrent was the troops, the bases, and the exercises. Reducing them is the alliance, in the operational sense.

The Europeans have received these signals at the resolution that the European institutions are designed to receive them. The signals have been processed not as anomalies requiring American reassurance but as confirmations of the 2017 diagnosis. The European response has been to accelerate the architecture they began building eight years earlier, on the assumption that the underlying conditions are now permanent rather than episodic. The administration that took office in January 2029 — whichever administration that is — will inherit not a transient disruption but a structural reconfiguration that has been in motion for nearly a decade.

The American foreign-policy commentary has, for the most part, continued to cover the alliance as if the 2025–2026 events are a perturbation that will be corrected by the next administration. The Europeans have stopped operating under that assumption.

What They Are Building

The architecture built since 2017 and accelerated since 2025 has named components, institutions, and budgets. The visible structure is the following.

The ReArm Europe Plan, announced in March 2025, mobilizes approximately €800 billion over four years to strengthen European defense capabilities. The plan is funded through a combination of national fiscal expansion, EU-level borrowing, and the relaxation of the Stability and Growth Pact’s defense-spending rules. It is, in scale, the largest coordinated European military mobilization since the Second World War. The plan’s existence is not contested; its implementation timeline is on schedule; its capacity-generation effects are visible in defense-industrial output figures across the major European producers.

The EU Defense Industrial Strategy, published in 2024 and accelerated through 2025, sets production-capacity targets for European defense manufacturers and provides EU-level procurement coordination through the European Defense Agency. The strategy’s most consequential effect has been the scaling up of ammunition production — the bottleneck the Ukraine war exposed, which the European industrial base has now been substantially restructured to address. By mid-2026, European 155mm artillery shell production has roughly tripled from the 2022 baseline. The bottleneck has not been fully closed; it is closing.

The Bucharest Nine — Poland, Romania, Bulgaria, the Czech Republic, Hungary, Slovakia, Estonia, Latvia, and Lithuania — has consolidated into a regional security architecture under predominantly Polish leadership. The B9’s coordinated planning, joint procurement, and shared deterrent posture have made the eastern flank the most militarily prepared region of the alliance. Poland’s defense spending, above 4% of GDP and rising, has produced the largest land army in Europe outside Russia. The Polish military-industrial expansion, sustained Polish acquisition of South Korean armor and artillery, and the construction of the Eastern Shield border defense system represent a national mobilization of a scale Europe has not seen in two generations.

The Nordic-Baltic spine — Finland’s 2023 accession, Sweden’s 2024 accession, and the integrated defense planning that has followed — has converted the Baltic Sea into what European planners refer to as a NATO lake, though the relevant analytical question is whether the architecture they are building still requires the NATO designation to function. The Nordic-Baltic Eight has coordinated air-policing, integrated air-defense, and joint maritime operations at a level that does not require American command-and-control infrastructure.

The Franco-German nuclear conversations represent the architecture’s most consequential and least-discussed element. The conversations have been quiet, episodic, and ongoing since 2017. They have intensified in 2025 and 2026. The substantive content is the question of whether the French force de frappe — France’s independent nuclear deterrent — could be extended into a European deterrent under some institutional arrangement that does not require American consent. The answer is procedurally complex and politically delicate. It is also no longer the closed question it was in 2017. Macron has spoken publicly about the question multiple times since 2020; Merz has not foreclosed it; the discussions are real.

The German rearmament, financed initially through the €100 billion special fund established in 2022 and now expanded substantially under the Merz government, represents the largest peacetime German military expansion since 1955. The constitutional debt brake reforms, the procurement reorganization, the personnel-recruitment overhaul, and the doctrinal reorientation are all in motion. Germany in 2026 is not Germany in 2014. The trajectory is toward a German military that can serve as the central conventional force on the European continent, paired with French nuclear coverage and Polish forward defense.

The United Kingdom’s post-2024 reconvergence with European defense structures — Lancaster House sequel arrangements with France, expanded defense coordination with Germany under the Trinity House Agreement, the Strategic Defence Review 2025, the AUKUS recalibration — has positioned the British defense establishment as a hinge between the European architecture and the residual transatlantic relationships. The British relationship with the European architecture is not membership; it is functional integration on the specific dimensions where British capability is necessary.

The cumulative effect, by mid-2026, is a security architecture that does not require American command-and-control infrastructure to operate, does not require American nuclear coverage to deter, does not require American intelligence sharing to function, and does not require American political continuity to plan over a multi-decade horizon. The architecture is not complete. It is operational.

The Europeans are not catching up. They are building what comes next. The work began in 2017 and is now eight years along.

The American foreign-policy commentary’s lag in recognizing the scale of this construction is itself an artifact of the commentary’s institutional assumptions. The relevant analytical instruments — the think tanks, the journals, the policy-research institutions — have been calibrated for decades to read European defense activity through the lens of transatlantic burden-sharing, the question of whether the Europeans are doing enough relative to American expectations. The lens is the wrong lens. The Europeans are not doing burden-sharing. They are building a sovereign security architecture under conditions in which American expectations have ceased to be a relevant input to their planning.

The Question Nobody Is Asking

The American foreign-policy debate continues to operate inside the question of how to lead NATO. The question is the wrong question. NATO, as it has operated since 1949, requires American leadership to function. Without American leadership, what survives is a different institution operating under the same name. The decade-long crack is the transition between the institutions. The transition is now substantially complete.

The right question is the one no American foreign-policy institution is positioned to ask: what is the United States when it is no longer the alliance’s spine?

For seventy-five years, American foreign policy has assumed alliance leadership as both means and end. The means assumption was operational — that American power projects through alliances, that the alliances multiply American capacity, that the alliances reduce the per-unit cost of American security commitments. The end assumption was constitutive — that what American foreign policy is is the leadership of the alliance system. The American military’s global posture, the dollar’s reserve status, the sanctions architecture, the intelligence-sharing relationships, the diplomatic resource allocation, the foreign-aid architecture, the international-financial-institution leadership, the trade-system design — all of these were built around alliance leadership as both means and end.

When the alliance leadership ceases to be the operative role, the entire infrastructure faces a question it has not faced since 1945. The infrastructure does not adapt to the new condition automatically. It requires redesign. The redesign requires a strategic analysis that names what American power is, and is for, in the absence of alliance leadership. The strategic analysis has not been done. The institutional process by which the strategic analysis would be done has been substantially disabled — career foreign-policy professionals reassigned, the planning apparatus reconfigured, the intelligence-community analytical function redirected, the diplomatic corps thinned by departure and reassignment.

The Europeans are designing their replacement architecture because they have been working on it for nine years and because the work has been institutionally continuous. The Americans are not designing theirs because the contingency was never planned for at the institutional level and because the institutional capacity that would have done the planning has been redirected to other purposes. The infrastructure of American power is operating on assumptions that the environment no longer supports, and there is no organized process underway to update the assumptions.

This is the structural condition that the alliance-coverage commentary has not yet named. The condition is not that the United States is failing to lead the alliance. The condition is that the alliance is being replaced, and the United States does not have a strategic answer for what it is in the world the replacement creates.

The 2029 administration, whichever administration that is, will inherit this question. The question will arrive without a prepared answer because the preparation has not been done. The country that arrives at the question will be a country that has spent four years not preparing for it, and the institutional infrastructure that would have done the preparation will be in a different configuration than it was in 2024. The strategic answer will have to be improvised in real time, against a security environment that has already been substantially restructured by the European architecture being completed in parallel.

This is not a prediction of American foreign-policy failure. It is a description of the analytical task that has been deferred. Deferral has costs. The costs compound.

Latency

For readers whose arrangements have European exposure — citizenship pathways, residency permits, asset holdings, family situations, professional commitments — the structural transition has operational implications that the alliance-coverage commentary has not yet reached. European jurisdictions whose security premium has been historically discounted because of the assumed American underwriting are being repriced as the underwriting becomes uncertain. Some jurisdictions are net beneficiaries of the new architecture; some are net losers. The differentiation matters at a level that the prior alliance assumption did not require Americans to attend to.

The Nordic and Baltic jurisdictions are the architecture’s structural beneficiaries. Their security is now provisioned by the integrated regional architecture in ways that do not depend on American posture. Their pathways for residency and citizenship — Finland, Sweden, Estonia, Latvia, Lithuania — have a security premium that is rising as the architecture matures. The Polish trajectory is in the same direction.

The Mediterranean and Southern European jurisdictions are in a more complicated position. They are inside the European architecture but at its lower-priority periphery. Italy, Spain, Portugal, Greece — their security calculus is more dependent on the European architecture’s southern extension, which is less developed than the Eastern flank. The pathways through these jurisdictions remain available, and remain valuable, but the security premium they carry is more contingent.

The Western European core — France, Germany, the Benelux, Austria — sits at the center of the architecture being built. The pathways through these jurisdictions are the most expensive, the most procedurally demanding, and the most stable. The premium is high and rising.

The British position is the most ambiguous and the most likely to be repriced. The UK is functionally integrated into the European architecture without being institutionally part of it. The relationship is workable but conditional. The premium attached to British residency depends on the durability of the functional integration, which is conditional on political variables that are themselves in motion.

The window in which the security-premium repricing remains uncompleted is the window in which arrangements made at prior prices retain their pre-repricing terms. The window is open. It is closing on the schedule that the architecture’s completion sets, not on the schedule that the American electoral calendar sets. The alliance is being replaced. The arrangements that depended on the prior alliance are being repriced.

The treaty persists. The credibility that made the treaty operative does not.

The replacement is taking shape.

---

The Long Memo is what I write when the architecture moves faster than the commentary moves. If this piece named something you had not yet seen named, the rest of the work runs in the same register.