Extended deterrence is the promise by a nuclear-armed state to use its military power, potentially including nuclear weapons, in defense of an ally rather than only itself. Within NATO, this promise is bound up with Article 5 of the North Atlantic Treaty, which holds that an armed attack against one member is to be treated as an attack against all. Since 1949, the credibility of this guarantee has rested heavily on the United States: its nuclear arsenal, its forward-deployed conventional forces, and the perceived willingness of an American president to risk a wider war, potentially a nuclear exchange, to defend a country thousands of miles from Washington. Whether that guarantee remains credible in 2026 is sharply contested, particularly as the Trump administration's defense strategy emphasizes that European allies should “take primary responsibility” for their own conventional defense while the United States provides “more limited” support.¹⁾ The dispute is not merely about troop numbers or weapons systems; it concerns whether deterrence is fundamentally a matter of demonstrated capability, declared intent, or some combination that resists clean measurement until it is tested in a crisis. Because deterrence that fails is only visible in retrospect, and deterrence that succeeds leaves no observable event, partisans on every side of this debate are arguing from inference rather than direct evidence, which is part of why the disagreement persists.

Advocates of this position argue that Article 5 has never been formally invoked outside the September 2001 response to the attacks on the United States, and that the absence of a direct Russian attack on any NATO member since 1949, including during periods of acute crisis such as the Cold War's most dangerous years and the post-2022 war in Ukraine, is itself evidence that deterrence has worked exactly as intended.²⁾ They contend that critics conflate rhetorical ambiguity from individual American officials with an actual change in treaty commitments or military posture, and that no presidential administration has withdrawn from NATO or renounced Article 5 despite decades of public hand-wringing about American resolve. On this view, the relevant test is not whether doubts are voiced in op-eds or parliamentary debates but whether the United States maintains the forward-deployed forces, nuclear-sharing arrangements through the Nuclear Planning Group, and joint exercises that constitute the “hardware” and “software” of assurance.³⁾ Proponents further argue that deterrence requires only a relatively low threshold of credibility to function against the adversary, while the higher threshold is for reassuring allies; as British defense minister Denis Healey is often paraphrased as having observed, deterring Moscow may take only a small fraction of the credibility needed to fully reassure Europeans, meaning that allied anxiety does not by itself indicate that deterrence toward Russia has weakened.⁴⁾ They also point to recent reinforcement rather than retreat: the United Kingdom's decision to join NATO's dual-capable aircraft mission and purchase American F-35A aircraft, and the 2025 Northwood Declaration in which Britain and France affirmed a European dimension to their nuclear deterrents, are read as evidence that the alliance's overall deterrent posture is adapting and strengthening rather than collapsing.⁵⁾ Finally, some who hold this position argue that public statements emphasizing burden-sharing, such as the 2026 National Defense Strategy's call for allies to take on more conventional responsibility, reflect a long-standing and legitimate demand that European states spend adequately on their own defense, not an abandonment of the underlying nuclear guarantee, which remains formally intact.

Advocates of this position argue that extended deterrence depends on perceived willingness as much as physical capability, and that repeated public statements from American officials questioning the value or permanence of US commitments to Europe constitute exactly the kind of erosion that matters most, regardless of whether any treaty document has been formally altered.⁶⁾ They note that the 2026 National Defense Strategy, unlike the Nuclear Posture Reviews of 2010, 2018, and 2022, did not explicitly affirm extended deterrence and instead described US support to allies as “critical but more limited,” a textual shift they read as a material signal rather than rhetorical noise.⁷⁾ On this view, deterrence is a psychological and political relationship between the guarantor and the adversary's leadership, and an adversary that comes to doubt American resolve, whether due to troop drawdowns, ambiguous statements, or visible policy shifts toward prioritizing the Indo-Pacific, may act on that doubt long before any crisis tests Article 5 directly. Proponents also argue that the credibility of the American “tripwire” forces in Europe depended not only on their presence but on a broader political context that has shifted, leaving Russia potentially able to calculate that a limited incursion would force Europe to choose between capitulation and an implausible leap to strategic nuclear exchange, with no credible intermediate rungs on the escalation ladder.⁸⁾ They point to material developments as corroborating evidence: Russia's relocation of nuclear warheads to Belarus and joint nuclear exercises with Belarusian forces in May 2026 are cited as a sign that Moscow itself perceives an opening to test or pressure the alliance's resolve.⁹⁾ Finally, advocates of this position note that European leaders, including German Chancellor Friedrich Merz and Polish Prime Minister Donald Tusk, have publicly raised the previously unspoken question of whether the United States can be relied upon at all, treating the open airing of this doubt among allied governments, not only adversaries, as itself a measure of how far credibility has already fallen.¹⁰⁾

A related position, distinct from a simple assessment of whether the old guarantee holds or fails, argues that the debate over credibility is being superseded by a deliberate strategic redesign. Proponents of this view, associated with the Elbridge Colby Defense Strategy Viewpoint, argue that the relevant question is not “is the old extended deterrence credible” but “what kind of deterrence should replace it.” This approach emphasizes “denial defense,” meaning the buildup of forces capable of physically stopping an attack rather than relying primarily on the threat of retaliation or punishment, on the grounds that denial is more credible in contexts where the deterring power's political stakes are limited relative to the defender's.¹¹⁾ On this view, encouraging European states to build their own conventional and nuclear escalation capacity, exemplified by France's “forward deterrence” proposal and the Northwood Declaration, is not a retreat from extended deterrence but a more sustainable and credible architecture for it, since a defense resting on multiple independent centers of decision-making may be harder for an adversary to discount than one resting on a single, potentially wavering, guarantor. Critics of this reframing, found among both of the positions above, respond that whatever its theoretical merits, the shift toward burden-sharing and denial defense has so far been announced faster than the underlying European capabilities, particularly air defense, theater nuclear forces, and command-and-control architecture independent of the United States, have been built, leaving a credibility gap during the transition itself.¹²⁾

Participants across these positions broadly agree that extended deterrence has both a capability dimension, the physical forces and weapons available, and a perception dimension, the adversary's and allies' beliefs about resolve, and that the two do not always move together. There is also broad agreement that the war in Russia's invasion of Ukraine, and Russian nuclear signaling during that conflict, has materially increased the salience of the credibility question regardless of one's view of how it should be answered. Most participants further agree that European conventional and nuclear capability has historically been, and remains as of 2026, insufficient to substitute fully for the American contribution to NATO's deterrent posture, even as they disagree about how urgently or completely that needs to change.

• Main Topic: Extended Deterrence
• History: Extended Deterrence History, NATO - History
• Viewpoints: elbridge-colby-defense-strategy-viewpoint, european-nuclear-autonomy-viewpoint, nato-atlanticist-viewpoint
• Consensus: extended-deterrence-capability-perception-consensus
• Related Debate: friendly-proliferation-nuclear-allies-debate

~~REFNOTES~~