Extended Deterrence

Extended deterrence describes the arrangement under which a nuclear-armed state, most prominently the United States within NATO, commits to treat an attack on an allied but non-nuclear state as equivalent to an attack on itself, and to respond accordingly, including with nuclear retaliation if necessary. The purpose is to deter an adversary from attacking the allied state at all, by making clear that doing so would trigger the guarantor's full deterrent capability rather than merely the ally's own, typically more limited, defences. Extended deterrence carries a structural credibility problem that no amount of treaty language or reassurance fully resolves: its effectiveness depends entirely on an adversary's, and an ally's, belief that the guarantor would genuinely accept catastrophic risk to itself, including nuclear retaliation against its own territory, on another state's behalf. This was recognised from the earliest years of the Cold War, sometimes framed as the question of whether a guarantor would 'trade one of its own cities for an ally's city,' and the answer has never been definitively demonstrable, only assertable through policy, deployment and institutional commitment. Arrangements such as NATO's nuclear sharing, under which US nuclear weapons are stationed in allied states (Belgium, Germany, Italy, the Netherlands and Turkiye) and delivered, under US custody and authorisation, by allied aircraft, exist specifically to make the American commitment appear structurally difficult to withdraw from, since the promise of extended deterrence was never considered sufficient on its own. Extended deterrence's credibility has come under renewed strain in the 2020s amid the broader collapse of arms-control architecture (including New START's 2026 expiry) and questions about the durability of the American commitment to Europe, prompting the first serious open European debate in decades about a security architecture not dependent on a US nuclear guarantee. The concept is distinct from, but complements, conventional deterrence, and the credibility problem inherent to extended nuclear deterrence has been cited as a driver of Europe's post-2022 push to build more independent conventional defence capability.

• 1 Extended deterrence: a nuclear-armed state commits to treat an attack on an ally as an attack on itself, deterring the adversary via the guarantor's own deterrent.
• 2 Structural credibility problem: rests on belief that the guarantor would accept catastrophic risk to itself on an ally's behalf, which cannot be conclusively proven.
• 3 NATO's nuclear sharing (US weapons in Belgium, Germany, Italy, Netherlands, Turkiye, delivered by allied aircraft under US custody) exists to reinforce this credibility.
• 4 The United Kingdom rejoined NATO's airborne nuclear mission in 2025, procuring F-35A aircraft for the role.
• 5 Credibility strained in the 2020s by arms-control collapse (New START's 2026 expiry) and doubts about sustained US commitment to Europe.
• 6 Has prompted the first serious open European debate in decades on a security architecture independent of the US nuclear guarantee.
• 7 Distinct from, but complements, conventional deterrence; Ukraine's war reinforced the importance of credible conventional capability alongside nuclear guarantees.

European planners' growing willingness to publicly discuss a security architecture not underwritten by Washington was cited as evidence that confidence in America's extended nuclear deterrence guarantee has already been discounted, regardless of any formal reassurance offered afterward.