NATO leaders met in Vilnius this week only around 30 kilometres from the Belarusian border, with one question dominating much of the summit: what exactly does the alliance mean when it says Ukraine’s future belongs inside NATO? Since the 2008 Bucharest summit, NATO has maintained that Ukraine will eventually become a member. Fifteen years later, Ukraine is fighting the largest European war since 1945 partly because Russia attempted to prevent its geopolitical movement towards the West, yet the alliance is still unwilling to provide Kyiv with either immediate membership or a clear timetable for receiving it.
The Vilnius summit moved Ukraine somewhat closer. NATO removed the requirement that Ukraine complete a formal Membership Action Plan, created a new NATO-Ukraine Council and reiterated that Ukraine would eventually join when allies agree and conditions are met. At the same time, the G7 announced a framework through which individual countries will negotiate longer-term security commitments involving military equipment, training, intelligence, cyber capabilities and development of Ukraine’s defence-industrial capacity. Those arrangements are intended to make Ukraine substantially harder to attack while leaving formal NATO membership for later.
This compromise makes strategic sense. It is also exactly why Ukraine remains dissatisfied. NATO membership carries something bilateral assistance cannot reproduce fully: Article 5, under which an armed attack against one member is treated as an attack against all. The deterrent value comes not simply from weapons but from an adversary believing an attack could trigger war with the United States and the rest of the alliance. Ukraine currently receives enormous Western military assistance, but Western governments retain discretion over how much they provide, when they provide it and what Ukrainian forces may do with it.
That difference between assistance and guarantee is fundamental. Assistance says that if you are attacked, I intend to help you. A guarantee says that attacking you creates an obligation for me. The first preserves flexibility for the guarantor. The second creates greater deterrence precisely because it reduces that flexibility.
Ukraine understandably wants the second. Russia has already demonstrated that it is willing to use force against Ukrainian territory, first through the seizure of Crimea and intervention in the Donbas in 2014 and then through the full-scale invasion beginning in February 2022. A post-war settlement that simply returns Ukraine to the same ambiguous security position could leave Kyiv facing the possibility that Russia rebuilds its military and attacks again later.
This is why Zelenskyy’s government argues that eventual NATO membership is not merely symbolic Western integration. It is an attempt to change Russia’s expected payoff from future aggression. If Moscow believes another attack on Ukraine automatically risks conflict with the entire alliance, the expected cost rises dramatically.
That is deterrence.
The problem is that the mechanism making NATO membership valuable to Ukraine is exactly what makes admitting Ukraine during the current war dangerous to existing NATO members. If Ukraine joined tomorrow while Russian forces continued occupying and attacking Ukrainian territory, Article 5 would immediately raise the question of whether the United States, Britain, France, Germany and other allies were obligated to enter direct military confrontation with Russia.
NATO therefore faces a contradiction it cannot solve through stronger wording. It wants to demonstrate that Russia cannot obtain a permanent veto over Ukrainian membership through aggression, because allowing invasion to determine alliance boundaries would reward war. Yet admitting Ukraine while the war continues could transform a Russia-Ukraine war into a NATO-Russia war involving several nuclear powers.
The alliance is therefore attempting to separate Ukraine’s long-term security architecture from the current military conflict. Support Ukraine sufficiently to survive and improve its bargaining position now, while postponing the full collective-defence guarantee until conditions make that guarantee less likely to trigger immediate war.
Strategically, that is understandable. From Kyiv, it can look like being asked to fight first for the security required to receive the security guarantee it wanted partly to prevent the fighting.
This creates a credibility problem. NATO says Ukraine will eventually join, but “eventually” is not a date and “when conditions are met” does not define the conditions precisely. Ambiguity gives allies room to respond to future circumstances. It also gives Russia reasons to believe that continued pressure might keep Ukraine permanently outside the alliance.
A promise intended to reassure Ukraine can therefore weaken deterrence if Moscow doubts whether the promise will ever be implemented.
This is why credible commitments are difficult in international relations. A government today cannot perfectly bind a government ten years from now. Biden can promise long-term American support, but another president or Congress may have different preferences. Germany can announce multi-year assistance, but future coalitions control future budgets. International agreements can increase the political cost of reversal, but they cannot eliminate political change.
Ukraine knows this unusually well because of the 1994 Budapest Memorandum. Ukraine inherited a large nuclear arsenal after the collapse of the Soviet Union and agreed to relinquish those weapons while Russia, the United States and Britain provided security assurances concerning Ukrainian sovereignty and territorial integrity. Those assurances did not amount to an Article 5-style defence commitment, and Russia later violated Ukraine’s borders. Ukrainian officials consequently have strong reasons to distinguish carefully between assurances, commitments and guarantees. Later Ukrainian negotiations over the G7 framework would explicitly emphasise that bilateral security arrangements should not substitute for NATO membership.
This does not mean Ukraine made an obvious mistake by surrendering nuclear weapons in the 1990s. The weapons were embedded in complicated Soviet command and maintenance systems, international pressure for denuclearisation was enormous, and the economic and diplomatic circumstances of the newly independent Ukrainian state were radically different from today. Retrospective claims that Kyiv could simply have retained a fully independent nuclear deterrent often ignore those constraints.
But political memory does not require the historical counterfactual to be simple. The lesson many Ukrainians draw is straightforward: a written assurance is valuable only to the extent that other states are willing to bear costs when the assurance is tested.
That is what makes NATO membership qualitatively different. Article 5 does not mechanically launch every member into identical military action, but the entire credibility of NATO depends on adversaries believing members will defend one another. If Russia attacked Poland or Lithuania and the alliance refused to respond meaningfully, the security value of NATO would collapse across Europe.
Institutional reputation therefore reinforces the commitment. NATO would not merely be defending one member. It would be defending the credibility of the alliance itself.
Ukraine currently sits outside that institutional mechanism.
The G7’s proposed long-term security commitments attempt to compensate by increasing Ukraine’s own capacity for deterrence by denial. Instead of relying entirely on a promise that Western armies will intervene, Ukraine would possess enough air defence, artillery, intelligence, training, industrial capacity and modern military technology that Russia expects another invasion to fail or become prohibitively expensive.
This distinction between deterrence by punishment and deterrence by denial matters. Article 5 partly deters through punishment: attack one ally and you risk retaliation from many. Building a heavily armed Ukrainian military deters through denial: attack Ukraine and you encounter a state capable of preventing you from achieving your objectives.
The two approaches can reinforce one another, but they create different risks. A direct NATO guarantee maximises deterrence if credible but raises escalation risks. A heavily armed non-member Ukraine reduces vulnerability while preserving Western discretion, but may never make Moscow as cautious as an Article 5 guarantee would.
The emerging strategy therefore appears to be making Ukraine extremely difficult to defeat while postponing the decision about whether it should be formally impossible to attack without confronting NATO.
This will be expensive. Ukraine’s military will require enormous resources even after active fighting eventually ends. Weapons need maintenance, ammunition stocks need replenishment, soldiers need training and domestic defence production needs reconstruction. A country already facing enormous infrastructure damage could therefore need to devote a substantial share of national resources to defence for years.
Security has an opportunity cost.
Every billion spent on air defence cannot simultaneously rebuild housing, schools or hospitals unless foreign financing expands total resources. Ukraine’s post-war development model may therefore depend heavily on whether Western states continue financing security while private and public investment supports reconstruction.
This creates another reason long-term commitments matter. Businesses deciding whether to build factories in Ukraine will care not only about tax rates and labour costs but about whether another invasion appears likely. Security guarantees therefore have an economic value beyond military deterrence because they change expected returns on investment.
A factory whose expected life is twenty years is difficult to finance if investors assign a high probability to another war within five.
Security can therefore operate as economic infrastructure.
The stronger Ukraine’s future deterrence becomes, the lower the geopolitical risk premium attached to investment should become. This means NATO integration, bilateral guarantees and military capacity can influence borrowing costs, capital flows and reconstruction long before another conflict occurs.
The debate is consequently not simply about soldiers and borders. It concerns whether Ukraine can become economically normal enough that millions of refugees feel confident returning and businesses believe long-term investment is rational.
There is also a larger European question. NATO’s enlargement after the Cold War extended collective defence across Central and Eastern Europe. Finland has now joined after Russia’s invasion transformed its security calculations, while Sweden is moving towards membership. Russia’s attempt to prevent NATO expansion has therefore produced additional expansion along its own borders.
This demonstrates how security policy can generate the opposite of its intended outcome. Russia invaded Ukraine partly within a broader argument that Western military institutions were moving too close to Russian territory. The invasion convinced previously non-aligned Finland and Sweden that remaining outside NATO had become more dangerous than joining it.
Threat perception is relational.
One country’s attempt to increase security can reduce another’s security, producing actions that ultimately leave the first country facing an even less favourable strategic environment.
That is the security dilemma again, although describing it does not imply moral equivalence between invasion and alliance membership. States retain agency over the methods they choose. Russia chose a full-scale invasion; Finland chose accession to a defensive alliance.
Understanding the strategic interaction does not erase responsibility for those choices.
Ukraine’s membership question is harder because admitting it would change NATO’s geographic and military obligations much more substantially. Ukraine has a long border with Russia, unresolved occupied territory and an active war. Existing allies therefore want to avoid making promises they are uncertain they would fulfil.
That caution is rational because an incredible guarantee can be more dangerous than no guarantee. If NATO formally promises to defend Ukraine but Moscow doubts that Germany, France or the United States would actually fight, Russia might test the commitment. NATO would then face the choice between a devastating escalation and publicly abandoning its own guarantee.
Deterrence depends on promises being believable.
This is why stronger language is not automatically stronger security.
A carefully limited commitment that governments will actually fulfil may deter more effectively than an enormous promise everyone privately doubts.
Ukraine understandably wants the largest possible commitment because it bears the direct risk of Russian attack. NATO members want commitments calibrated to risks their own populations are actually willing to accept.
That difference in preferences cannot be eliminated through diplomatic rhetoric.
It has to be negotiated.
The Vilnius compromise therefore may be frustrating precisely because it reflects the genuine distribution of risk. Ukraine wants certainty. NATO wants flexibility. Russia benefits from ambiguity. Western governments are attempting to create enough long-term certainty to strengthen deterrence without eliminating so much flexibility that they become direct parties to the current war.
There is no elegant solution.
The best arrangement may be one that nobody considers sufficient because every stronger arrangement increases another risk.
This is often what international security looks like when nuclear powers are involved. Policy is not choosing between safety and danger. It is choosing which configuration of danger appears least likely to produce catastrophe.
The summit nevertheless matters because it shifts the time horizon. Western policy during the first months of the invasion concentrated heavily on keeping Ukraine alive. The discussion now increasingly concerns what Ukrainian security should look like years after the current phase of fighting ends.
That is strategically important because Russia may believe time favours it. If Moscow expects Western political support to weaken eventually, it has incentives to prolong the conflict and wait. Long-term commitments attempt to change that expectation by communicating that Ukraine’s military capacity will not disappear simply because headlines move elsewhere.
This is another form of deterrence: deterrence through persistence.
The message is that Russia cannot simply outwait democratic attention spans.
Whether Western governments can make that message credible is uncertain. Democracies change leaders, budgets and priorities. Their flexibility is one of their strengths and one of the difficulties of long-term security policy.
Institutions can help by embedding cooperation deeply enough that changing course becomes politically and administratively costly. Training programmes, defence-industrial partnerships, intelligence networks and multiyear procurement create relationships that survive individual leaders more easily than annual political declarations.
This is why the architecture announced around Vilnius may matter even without immediate NATO membership. It begins institutionalising Ukraine’s military integration with Western states while leaving the final treaty commitment unresolved.
That is not what Zelenskyy wanted.
It may still be strategically significant.
The difficult question is whether it produces a stable bridge to NATO membership or becomes a comfortable substitute that allows allies to postpone membership indefinitely.
Ukraine clearly fears the second possibility.
Its scepticism is rational because temporary arrangements have a tendency to become permanent when changing them creates new risks.
The West therefore needs enough specificity that “Ukraine’s future is in NATO” eventually means something more than a sentence repeated at summits.
At the same time, Ukraine needs a security architecture capable of functioning before membership becomes politically possible.
That is the contradiction Vilnius is trying to manage.
Ukraine needs to be secure enough to join NATO.
Part of the reason it wants to join NATO is to become secure.
For now, Western governments are attempting to fill the gap with weapons, training and long-term commitments.
Whether that gap becomes a bridge or a permanent waiting room may determine the security architecture of Europe for decades.
