A ceasefire is nowhere in sight. Putin sees no interest in ending his war. Even so, the EU should start preparing for the many post-war challenges Ukraine will face.
The European Union’s accession promise loses credibility if it continues to keep the most significant benefits of accession until the end of the process. Its two political options for bringing membership forward – Membership First and Accession Treaty First – must both pass the same four tests.
Four scenarios for Ukraine’s future after 2027: from continued war to a lasting peace deal, and what role the EU must play in each
Ukraine has systematically removed humans from roles that don’t require human judgment, letting machines absorb the danger soldiers would otherwise face. What UGVs can and can’t do, and what lessons European institutions can draw from Ukraine’s experience.
Russia’s full-scale invasion of Ukraine began in a nuclear atmosphere. This was not new: since 2014, Russian nuclear capacity had shaped Western thinking about how far support for Ukraine should go. But on 24 February 2022, Vladimir Putin made the implication explicit, warning outside powers of unprecedented “consequences” if they dared to interfere. The Kremlin has repeated the message on a regular basis through further nuclear threats, exercises, doctrinal revisions, and the announced deployment of non-strategic nuclear weapons to Belarus.
The purpose of this signalling has been to deter direct intervention and to alter the Western political calculus of assistance to Ukraine. Moscow creates uncertainty; it turns every Western decision into a question of escalation and makes hesitation appear prudent.
Russia’s temporary occupation of Chornobyl and its continuing control of the Zaporizhzhia nuclear power plant showed how a conventional war can generate nuclear risk through disrupted power supplies, cooling systems, staffing, and communications around civilian nuclear infrastructure. This is not the same as nuclear bombing. Yet the possibility of catastrophe still shapes political choices.
European governments have tried to support Ukraine while avoiding a direct NATO-Russia war. They did not disagree about that goal. They disagreed about whether it required accepting Moscow’s self-declared limits on Ukraine’s self-defence. During the first two years of the full-scale invasion, this disagreement surfaced with each major capability decision: tanks, combat aircraft, long-range missiles, and permission to strike military targets inside Russia. Most restrictions were ultimately loosened, but only after months of argument and worsening conditions on the battlefield. This gave Russia valuable time to prepare.
The division was not between those who supported Ukraine and those who did not (that is a different argument). It was a division of speed, scale, and political imagination. Some states became leaders in military assistance and in removing self-imposed restrictions. Others remained cautious, slower to rearm, and reluctant to permit Ukraine to use supplied weapons beyond its borders. Too often, those with the greatest capacity lacked the urgency, while those with the greatest urgency lacked the capacity.
Why, under the same broad pattern of Russian nuclear signalling, are some European states more willing than others to sustain military aid, strengthen their own defences, and relax restrictions on Ukraine’s use of weapons?
The clearest pattern is geographical. States most exposed to conventional Russian pressure generally gave more, rearmed faster, and moved earlier to ease restrictions. Nuclear status and democratic quality correspond less consistently with these choices. Proximity changes how governments weigh the risk of restraint in aid to Ukraine. Those facing Russia at close range have generally judged that risk higher than those further away.
This policy brief compares 31 European donor states across three observable choices:
These choices are compared with three possible explanations: direct strategic exposure via a land border with Russia or Arctic Council membership (among European donors, the five Nordic states); independent nuclear status; and democratic quality, measured through the V-Dem Liberal Democracy Index.
This indirect design reflects a fundamental problem in deterrence research. The absence of a particular escalation cannot tell us for certain whether a particular deterrence succeeded or failed. No escalation may mean that support for Ukraine worked. But it may mean that Western ‘caution’ worked. It is also possible that the policy of self-restraint has actually increased the likelihood of future escalation, with that threshold simply not having been crossed yet. Or maybe the feared escalation was never likely in the first place. The purpose here is therefore not to calculate a single nuclear coefficient in European policy. It is to compare the political environments in which governments decided whether to arm Ukraine, rearm themselves, or relax their own caveats. This will help identify the factors that have divided Europe over security policy in the past, with the aim of mitigating such divisions in the future.
Stanley Kubrick’s Dr. Strangelove offers a useful image of the older nuclear world. The Soviet Union keeps its doomsday machine secret until an American bomber is already approaching its target. A deterrent announced too late cannot deter. But the film also belongs to a geography of distance. The bomber has to cross the Arctic, leaving time for telephone calls, arguments and increasingly desperate efforts to stop it. Nowadays the interval between the application of pressure and the need to act is much shorter.
Today’s real-life Moscow has not made the mistake of the Soviet protagonists of Kubrick’s film. Its nuclear warnings have been public, repetitive, and hard to miss. The question is: what do European governments fear more – a wider war triggered by action, or one made more likely by restraint?
The map below traces that divide. Five patterns stand out.
All six NATO members that share a land border with Russia – Estonia, Latvia, Lithuania, Norway, Finland, and Poland (with Kaliningrad) – rank among Ukraine’s leading military donors when their allocations are measured against the size of their economies. Latvia, Lithuania, and Poland also border Belarus, which Russia used as a staging ground for its invasion in 2022.
These countries differ in economic size, military capacity, and domestic politics. But they share a strategic fact: battlefield dynamics in the Russia-Ukraine war and the precedent of allied restraint transform their own security environment more quickly and more directly than that of European states further away.
Sustaining military assistance at this level for more than four years reflects not a temporary political preference but a durable national consensus across governments, parliaments, opposition parties, the media, and public opinion. For them, military aid to Ukraine has national security merit as well as the value of solidarity. The logic is close to “deterrence by delivery of arms”: states can avoid direct intervention while helping Ukraine deny Russia military success.
Denmark and Sweden also rank among Ukraine’s nine leading military donors relative to GDP. Greenland gives Denmark an Arctic security frontier; Sweden belongs to the same northern security environment through its geography and closely integrated Nordic (European Arctic) defence planning. Together with the six NATO states bordering Russia, they account for eight of the nine leading donors. For these countries, Russia is encountered either at the frontier or across the Arctic.
The Netherlands tests the difference between distance and exposure. It ranks eighth among Ukraine’s leading military donors relative to GDP, even above Poland, and is the only one with neither a land border nor an Arctic frontier. However, the destruction of MH17 in 2014 killed 196 Dutch nationals – the largest single-incident foreign civilian toll of Russia’s war against Ukraine. Russia’s war was therefore not a distant event in Dutch political memory. This does not necessarily establish causation, but it may help explain why physical distance did not produce the same political distance. Crucially, this is not an abstract historical memory of moral solidarity, but rather a reduction in political distance in response to the present threat. For example, the Czech Republic, Hungary, and Slovakia have historical reasons to fear Moscow, given the events of 1956 in Budapest and 1968 in Prague. Yet the absence of a shared border with Russia and of direct casualties from its current aggression means that this historical memory is not activated in the same way, nor does it place these countries among the leading supporters of Ukraine.
Iceland belongs to the Arctic Council, yet it has no standing armed forces or military equipment to transfer. The question is whether Russian military power figures in a state’s own security calculations, and whether that state has the means to act on them.
Military aid alone is an imperfect test of resistance to nuclear coercion. For states directly exposed to Russia, keeping Russian forces engaged and degraded in Ukraine serves an immediate national interest: it buys time for their own rearmament. The May 2024 debate over strikes inside Russia offered a harder test. Were governments prepared to accept a greater risk of escalation in order to make their support militarily effective?
Russia’s offensive on Kharkiv showed the practical cost of the restrictions. Ukraine was being attacked by forces operating from just across the border, while many of the weapons supplied by its partners could not be used against the bases, launchers, and troop concentrations inside Russia sustaining that attack. The question became: should fear of escalation continue to grant Russian forces a protected rear?
Conventional precision strikes offered a way out of the false choice between paralysis and nuclear escalation. They could degrade long-range firepower, airfields, and military production without targeting population centres or strategic nuclear forces. The dispute in May was not simply about how far Ukraine could strike. It was about whether weapons already transferred could be used to hit the launchers, bases, and supply lines inside Russia that were driving the assault on Ukraine. On 31 May 2024, the United States eased its restrictions, initially only for targets connected to the Kharkiv offensive. As NATO’s largest military power and Europe’s principal provider of extended nuclear deterrence, Washington set an important benchmark. National permissions varied in scope, and some were narrower than others. Public support for lifting restrictions even before the United States did so is an important indicator of a willingness to risk worsening relations with Russia without the backing of the main ally, as part of a broader security strategy. The same countries lead on both measures. All nine leading military donors relative to GDP supported some relaxation by 31 May: the six NATO states bordering Russia, Denmark and Sweden as Arctic states, and the Netherlands. France, the United Kingdom, and Czechia did so as well. In other words, they were willing to waive certain restrictions until they received a political signal from Washington. Where Russian conventional force was perceived as an immediate threat, the risks of restraint appeared to outweigh the risks of escalation.
Defence spending points in the same direction. Between 2021 and 2025, the six NATO states bordering Russia raised military expenditure by an average of 1.52 per cent of GDP, compared with 0.41 per cent across the rest of the sample. Every European NATO member spending more than three per cent of GDP on defence in 2025 was exposed to Russia through a land border or the Arctic. For these governments, support for Ukraine was not an alternative to national rearmament. It formed part of the same defence strategy: weaken Russia’s capacity for aggression now, and reduce its ability to dictate the terms of the next crisis.
Ukraine is, of course, the extreme case. In Gallup International’s End of Year 2023 survey across 41 countries, Ukrainian respondents were among those most likely to see no risk of nuclear weapons being used, despite living under both Russia’s attack and its nuclear threats. This is not a simple geographical effect; it points instead to the experience of an ongoing conventional war. For a society already living with missile strikes, occupation, and battlefield losses, nuclear escalation can seem less immediate than the violence already under way. The survey cannot separate this from a second possibility: that Ukrainians, unwilling to be seen yielding to nuclear blackmail, answered with deliberate defiance. Either way, the response reflects a population whose reading of the nuclear threat has been reshaped by direct experience of war – the opposite of the paralysis Russia’s nuclear threats are meant to induce.
The Nordic cases also caution against a simple democracy-explains-it story. They sit near the top of the V-Dem Liberal Democracy Index, and several have been among Ukraine’s strongest supporters. Yet they do not isolate a democratic effect. These states are also embedded in the Baltic-Arctic security environment; Finland and Norway share land borders with Russia (albeit ones of vastly different lengths). Across the wider sample, a high V-Dem score does not consistently track greater military aid, faster rearmament, or earlier support for lifting weapons restrictions.
Democracy may, however, matter at a later stage. Once a government commits, strong democratic institutions can make it easier to stay the course: they provide continuity, sustain public trust, and give rival parties a basis for keeping a common line. Where isolationist and populist forces have gained ground, support for Ukraine has more readily become a domestic political fault line. But democratic quality does not directly predict how much economic cost or escalation risk a government will accept.
The conclusion is therefore conditional. Proximity supplies the strategic motive; democratic institutions may help sustain the response. Neither factor, by itself, determines how far a government will go.
For much of NATO’s history, the division of responsibility seemed clear. The United States supplied extended nuclear deterrence, while the next-largest Western allies held most of the alliance’s high-end military capability. Eastern members were treated less as contributors to allied strategy than as consumers of it – their warnings about Russia often heard as regional anxiety, pushing everyone towards escalation rather than shared threat assessment. The familiar alliance schism followed: states on the frontier feared abandonment, while those further away worried about being drawn into a conflict they did not control.
States bordering Russia now devote the largest shares of their economies both to arming Ukraine and to rearming themselves. Europe’s larger powers retain far greater resources in absolute terms, but commit a smaller relative share of them. Capacity and effort are no longer concentrated in the same places.
France and the UK illustrate the difference between political leadership and material commitment. Both supported loosening restrictions on strikes inside Russia. Yet their increases in defence spending, and their military assistance relative to GDP, remained below those of the leading border and Arctic states. Between 2022 and 2026, British bilateral military aid amounted to 0.61 per cent of 2021 GDP, and French aid to 0.25 per cent. Denmark committed 2.95 per cent and Estonia 2.82 per cent – roughly five times the British share and more than ten times the French one. Where a government finds the money, and where it does not, itself reflects how directly it feels the threat.
The same alignment appeared in May 2026, when Mark Rutte urged allies to allocate 0.25% of their GDP to support Ukraine, but it did not secure unanimity. According to reporting by The Telegraph, the UK, France, Italy, Spain and Canada opposed the target, while the states already meeting it – the Netherlands, Poland, and the Nordic and Baltic countries – pushed it forward. The alignment is the argument of this brief in miniature: the burden is heaviest for those nearest Russia, and the reluctance greatest among those furthest away. Ukraine cannot depend on a new political fight each time it needs ammunition or air-defence missiles.
This does not make Europe’s larger powers security consumers. Nuclear forces, industrial depth, and high-end military capabilities remain concentrated among them. But the greatest relative expenditure, and much of the political exposure, now lie on the frontline. Europe’s emerging problem is a separation between where capability resides and where urgency is strongest.
In any case, we can observe different dynamics in the assessment of risks. Geographically vulnerable states fear being abandoned because of the western European- and US-perceived risk of future nuclear escalation – in other words, they might fear finding themselves tomorrow where Ukraine is today. As a result, it is strategically important for them to reduce the fear of escalation as a factor shaping decisions on support for allies.
This does not mean that the frontline and Arctic-exposed states are acting recklessly. On the contrary, during the war, the Baltic states and Poland have responded cautiously to military incidents involving Russia, such as drone and missile incursions. They support Ukraine not because they discount escalation, but because they judge that allowing fear of escalation to shape every decision would invite further coercion. In this sense, their aim is to alter Moscow’s calculation by showing that its nuclear signalling will not deliver a weaker Ukraine or a divided alliance.
Russia’s latest nuclear campaign comes at a moment when its conventional strategy is yielding less than the Kremlin needs. Its grinding offensive has not produced a decisive breakthrough. Ukraine, meanwhile, has carried the war into Russia’s rear: against refineries, logistics networks, and air bases. The message behind Russia’s new nuclear rhetoric is blunter than before: stop Ukraine from bringing the war home to us.
Russia can intensify its bombardment of Ukraine, and it is doing so. But bombardment alone is a poor route to the political and territorial result it seeks. Its other conventional options are unattractive. Among other costs, a new offensive from Belarus would turn a useful rear area into a vulnerable theatre, exposing Russian forces and supply routes to Ukrainian long-range strikes. A new mass mobilisation could add men, but would not by itself solve Russia’s central military problem: not how many troops it can place near the front, but how they can be effectively armed and deployed. Russia has expanded its force in and around Ukraine to more than 700,000 personnel. Yet the larger that force has become, the harder it has been to concentrate it and employ it at scale on a drone-saturated battlefield.
That leaves more dangerous ways of changing the conflict. Russia could mount a more calibrated hybrid attack against a NATO member – sabotage, a cyberattack, or a deniable violent incident, designed to force an argument over whether an armed attack has occurred and what collective response it requires. Or it could gamble on limited nuclear use in Ukraine: a strike small enough, in Moscow’s calculation, to horrify the West without bringing NATO directly into the war. Recent Kremlin-linked reporting has floated such a possibility while also stressing that there is no sign of actual preparations. That ambiguity is part of the pressure.
Neither the hybrid attack nor limited nuclear use guarantees Russia a military victory. Both are designed to split the alliance at the moment it needs to act. A hybrid attack tests whether members can agree on what they have just witnessed; for frontline states, hesitation would revive the oldest fear of alliance politics: abandonment. A limited nuclear strike inside Ukraine would leave little doubt about the facts, but it would test whether the West is prepared to act to stop Russia, and could again divide NATO countries over whether direct military intervention is necessary or should be avoided.
This is the fault line revealed by the comparison in this policy brief. The states closest to Russia have already treated Ukrainian resistance as forward defence; they would see a weak response as evidence that conventional aggression still pays. Russia does not need to defeat NATO in battle. It needs its members to disagree about the price of an answer.
Europe cannot stop Russia from issuing nuclear threats. It can stop them from working. Europe may never know whether Moscow’s next warning is a bluff, genuine preparation, or an attempt to influence a particular decision. It does not need to know. If nuclear bluff were allowed to work as effectively as nuclear weapons themselves, Russia’s influence would be virtually unlimited.
Ukraine has already shown how this works. It faces the highest risk of becoming the arena for the use of Russian nuclear weapons, yet it maintains the will to resist in the face of threats. Europe must finally make that resilience collective. Geography cannot be removed from Europe’s security map, but the financial and political burden it creates can be better shared.
The policies outlined in the recommendations will reduce Europe’s vulnerability to nuclear rhetoric and undermine Moscow’s objectives: a weaker Ukrainian military, an alliance that pauses at every new threat, and European societies frightened into retreat. That means sharing the burden before a crisis, keeping Ukraine’s resistance supplied, and making a new Russian threat a reason for collective action rather than another European pause.
In the event of a refusal to supply a new type of weapon requested by Ukraine, or a refusal to lift restrictions on the use of weapons already being supplied, supplying governments should circulate twin risk assessments to the NATO-Ukraine Council:
Ukraine should attach its own operational assessment and two allies (at least one from the northern or eastern flank) should comment on the supplying government’s assessment.
A new Russian warning would become a claim to be tested against evidence and operational consequences, rather than a cue for another automatic pause. Russia’s leverage is greatest when a government’s fear remains private and unexamined; placing that fear before allies turns it into a European question.
European NATO allies should commit, for five years, to annual military support for Ukraine worth at least 0.25 per cent of GDP, additional to their own defence budgets, sustained regardless of the war’s phase. The form can vary – transfers from national stocks, joint procurement, Prioritised Ukraine Requirements List (PURL) packages, contracts with Ukrainian producers, or long-term production orders against agreed Ukrainian requirements. The floor should not depend on whether Russia is currently attacking.
The states furthest from Russia have so far resisted such a floor (see Section 5), but Ukraine cannot depend on a new political fight each time it needs ammunition or air-defence missiles. A common floor would also ease the disproportionate burden now carried by the frontline and Nordic states, which already exceed it.
Predictability is itself a form of deterrence. Nuclear warnings work by reopening a settled decision; a fixed commitment leaves none to reopen. A public five-year commitment would make it harder for Russia to use a new nuclear warning, or pressure on a single capital, to reopen the basic question of whether Ukraine will be armed. Moscow would face a Ukraine whose means of resistance were already assured.
Russia’s nuclear signalling works when it is treated as a dispute between Moscow and NATO alone. The point is not to win over the whole world on the war, but to secure a narrower agreement that most states can accept: that using or threatening nuclear weapons is a line the world will not tolerate. The European Council should therefore keep a conditional economic and export-control package ready, while the European External Action Service (EEAS) uses bilateral channels, the G20, the UN, the International Atomic Energy Agency and the Organisation for the Prohibition of Chemical Weapons to work with China, India, Turkey and Gulf partners. The aim is not to recruit them into a Western sanctions coalition, but to secure a clear common position: no political cover for nuclear escalation, no recognition of territory or ceasefire terms won by using or threatening nuclear, chemical or biological weapons, and greater scrutiny of Russian trade, finance, logistics and nuclear cooperation. The task is to fix that position before a crisis, so that Russia cannot trade a breach of the taboo for a weaker Ukraine or the normalisation of the nuclear factor in international relations.
Russian nuclear signalling creates an asymmetric political effect: Moscow issues warnings while European governments reopen debates over restraint. Europe should answer by bringing Ukraine into its own deterrence cooperation, so that Russian warnings meet a visible European response rather than a pause, without turning every warning into a nuclear crisis.
France’s March 2026 forward-deterrence doctrine, together with closer nuclear coordination with the United Kingdom and Germany, provides a starting point. European governments should use this framework to involve Ukraine in unclassified exercises, conventional-support projects and standing crisis consultations. Ukraine can provide operational lessons from its experience of Russian escalation. This would place Ukraine within Europe’s strategic depth and reduce Moscow’s ability to turn nuclear signalling into delays or restrictions on support.
The European Union’s accession promise loses credibility if it continues to keep the most significant benefits of accession until the end of the process. Its two political options for bringing membership forward – Membership First and Accession Treaty First – must both pass the same four tests.
You will now receive EPIK’s latest analysis, publications, event invitations, and other updates directly in your inbox.
You will now receive EPIK’s latest analysis, publications, event invitations, and other updates directly in your inbox.
Go Home Page
The form has been successfully submitted. We value your time, so please expect a call or email from our manager shortly.
Go Home Page
You will now receive EPIK’s latest analysis, publications, event invitations, and other updates directly in your inbox.
You will now receive EPIK’s latest analysis, publications, event invitations, and other updates directly in your inbox.
You will now receive EPIK’s latest analysis, publications, and event updates directly in your email.
Go Home Page
You will now receive EPIK’s latest analysis, publications, event invitations, and other updates directly in your inbox.
You will now receive EPIK’s latest analysis, publications, event invitations, and other updates directly in your inbox.
Go Home Page
The form has been successfully submitted. We value your time, so please expect a call or email from our manager shortly.
Go Home Page
You will now receive EPIK’s latest analysis, publications, event invitations, and other updates directly in your inbox.
You will now receive EPIK’s latest analysis, publications, event invitations, and other updates directly in your inbox.
You will now receive EPIK’s latest analysis, publications, and event updates directly in your email.
Go Home Page
