For almost three months, I have ended posts about Ukraine with some version of the same qualification: Russia has assembled the capability to invade, Western governments believe the risk is serious, but an invasion has not happened and should not be treated as inevitable.
On Thursday morning, that uncertainty ended.
Vladimir Putin announced what he called a “special military operation”, and Russian forces attacked Ukraine from several directions. Missiles struck targets across the country. Russian troops entered from Russia, occupied Crimea and Belarus. Fighting reached the approaches to Kyiv. Ukrainians began sheltering underground or leaving their cities, while men of fighting age were required to remain in the country as the government mobilised for defence.
President Volodymyr Zelensky remained in Kyiv as Russian forces approached the capital. The United States had reportedly offered assistance in leaving; Zelensky instead stayed and has appeared repeatedly in videos with senior officials to demonstrate that the government remains in the city. Whatever happens militarily, that decision already matters politically because one of the central questions in a war is whether citizens believe the state they are being asked to defend still exists.
Ukraine, unexpectedly to many observers, is still fighting.
It is far too early to know how the war ends. Russia possesses overwhelming advantages in many categories of conventional military power, and several days of resistance cannot tell us the eventual outcome. But the rapid collapse that many feared has not happened. Russian forces have encountered serious resistance, and Kyiv remains under Ukrainian control tonight.
For months I have been writing about deterrence, credible commitments, sanctions, intelligence disclosures, NATO, European strategic autonomy and the security dilemma. This week forces a fairly uncomfortable conclusion.
Deterrence failed.
The United States publicly revealed intelligence about Russian preparations. Britain and other countries supplied weapons to Ukraine. Western governments threatened severe sanctions. Emmanuel Macron, Olaf Scholz and other leaders negotiated directly with Putin. NATO warned that an attack would fundamentally change relations with Russia.
Putin attacked anyway.
But saying deterrence failed is not the same thing as saying every attempt at deterrence was pointless or obviously mistaken. That would be reasoning backwards from the outcome. A seatbelt has not necessarily been irrational because somebody dies in a crash while wearing one. A policy should be judged partly on whether it changed probabilities given the information available when the decision was made, not solely on whether the worst outcome eventually occurred.
The more useful question is why the expected costs were insufficient.
Western strategy appears to have rested on an assumption that Russia would compare the benefits of invading Ukraine with the enormous economic, military and diplomatic costs an invasion would create and conclude that war was not worth it. That is a rational framework. What matters is how each variable was valued.
Perhaps Western governments underestimated how important control over Ukraine’s geopolitical direction had become to Putin. Perhaps Putin underestimated the sanctions Russia would face, the Ukrainian resistance his forces would encounter or the unity the invasion would produce among Western governments. Perhaps both sides correctly understood the other’s position but valued the stakes so differently that no deterrent short of direct military intervention could close the gap.
We do not yet know.
But one assumption now appears particularly weak: that economic interdependence would necessarily impose enough restraint to prevent major war in Europe.
Russia sold enormous quantities of energy to European countries. European businesses invested in Russia. Russian companies depended on Western financial markets and technology. Germany constructed an entire energy relationship around Russian gas. A war severe enough to rupture those connections appeared economically irrational.
It happened anyway.
This does not prove the liberal argument that economic interdependence reduces conflict is completely wrong. Economic connections raise the cost of war, and raising the cost can reduce its probability.
The mistake is treating reducing probability as eliminating possibility.
Economic interests are one interest among several. Governments may accept enormous economic losses when leaders believe territorial, ideological or security interests are more important.
Europe appears to have discovered the limits of economic leverage this week.
Germany’s reaction may be the clearest evidence of how dramatically the strategic environment has changed. Chancellor Olaf Scholz addressed the Bundestag today and described the invasion as a Zeitenwende, a turning point. His government plans a €100 billion special fund for the Bundeswehr and has committed Germany to spending more than 2% of GDP on defence. Berlin has also reversed its longstanding reluctance to send weapons into active conflict zones and agreed to provide weapons to Ukraine.
Only two weeks ago, I wrote that Europe possessed enormous economic power without equivalent independent military capacity.
Germany has apparently reached the same conclusion in considerably more dramatic circumstances.
This is a historic change because post-war German foreign policy was built partly around restraint. Germany’s history created deep suspicion of militarism. Economic integration, diplomacy and European institutions became central instruments of German power.
That model achieved extraordinary things.
It helped embed Germany inside a peaceful European order.
But every strategic doctrine rests on assumptions about the environment in which it operates. If the environment changes, policies that once represented wisdom can become vulnerabilities.
The challenge is determining whether the environment genuinely changed or whether governments are overreacting to one shock.
Russia’s invasion is large enough that I think the first interpretation is difficult to avoid.
A major European state has used military force against a neighbouring sovereign country after demanding changes to the European security order. That directly challenges the assumption that large-scale territorial war in Europe had become essentially obsolete.
Defence spending therefore looks different today than it did a week ago.
This does not mean every additional euro spent on the military automatically increases security. Military budgets can be wasteful like any other budgets. Weapons purchased without coherent strategy merely create expensive inventories.
But underinvestment can also become expensive when deterrence fails.
Germany is now discovering that the fiscal savings created by decades of relatively low defence spending came partly from operating inside a security environment protected heavily by American capabilities.
The €100 billion announcement is therefore not simply military policy. It is recognition that economic power cannot always substitute for coercive capability.
The sanctions announced this week are also extraordinary. Western governments have moved against major Russian banks and individuals, imposed export controls and, this weekend, announced measures intended to restrict the Russian central bank’s access to international reserves. Several governments have also moved towards excluding selected Russian banks from the SWIFT financial messaging system.
These measures demonstrate another kind of Western power.
The United States and European Union occupy central positions in global finance. Russia can possess nuclear weapons and a large military while remaining vulnerable to institutions operating through dollars, euros, banks, payment systems and advanced technology.
That is economic statecraft at a scale rarely attempted against an economy as large and internationally connected as Russia’s.
But sanctions should still be evaluated through the distinction I have repeatedly made this year between punishment and persuasion.
Punishing Russia for invading Ukraine is one objective.
Changing Russian behaviour is another.
Sanctions can devastate economic relationships without necessarily persuading Putin to reverse the invasion. If the political objective is ending the war, policymakers need to understand what economic pressure can realistically accomplish and how quickly.
There is also a humanitarian consequence. Russian citizens did not individually decide to invade Ukraine. Broad economic sanctions can reduce living standards for people who possess little influence over Putin’s decision.
That does not automatically make sanctions wrong. Russia’s government initiated a war against another state, and allowing aggression to impose no meaningful costs would create dangerous incentives.
It means the distribution of those costs should remain part of the analysis.
The same applies to Europe’s decision not to intervene directly militarily.
Ukraine is being attacked, yet NATO is not entering the war.
Morally, that can feel inadequate. If Western governments believe Russia is committing an act of aggression, why not defend Ukraine directly?
Because Russia possesses nuclear weapons.
That fact places an enormous boundary around Western policy.
Ukraine is not a NATO member. There is no Article 5 commitment requiring American, British, French or German forces to fight on its behalf. Sending NATO forces into direct combat with Russian forces could transform a terrible regional war into a confrontation between nuclear powers.
This is the brutal distinction between supporting a country and guaranteeing it.
Western governments are supplying weapons, intelligence, financial assistance and sanctions. They are not willing to accept the risks of direct war with Russia.
Ukraine therefore has to fight principally with Ukrainian soldiers.
The Budapest Memorandum I wrote about last Sunday looks even more consequential now. Ukraine relinquished inherited Soviet nuclear weapons in the 1990s while receiving assurances concerning its sovereignty and borders. Russia has now launched a major military attack against the state whose territorial integrity it had committed to respect.
Again, that does not prove Ukraine could easily have maintained an independent nuclear arsenal or that keeping those weapons would necessarily have produced a safer outcome.
But the signal to other governments is troubling.
The international system asks states to believe that sovereignty, treaties and institutions can provide security without nuclear weapons. Ukraine’s experience may strengthen the opposite conclusion: when survival becomes existential, independent deterrence is more reliable than international assurances.
If that lesson spreads, the long-term damage could extend far beyond Europe.
There is also something important happening inside Ukraine that military comparisons failed to capture.
Before the invasion, analysts could compare numbers of tanks, aircraft, missiles and soldiers. Russia appeared overwhelmingly stronger.
Those comparisons remain important.
They did not measure political willingness.
Ukraine’s resistance demonstrates that military capacity is partly endogenous to political legitimacy. A soldier defending a state they believe represents their community may fight differently from one defending an institution they expect to collapse.
Zelensky’s behaviour matters in this context. His decision to remain in Kyiv is not simply symbolic theatre. Leadership can solve a coordination problem.
If citizens believe the president is fleeing, officials may flee. Soldiers may conclude the government is finished. Local leaders may negotiate surrender. The expectation of collapse can create collapse, exactly as happened in Afghanistan last August.
If the president remains visibly present, the opposite signal is transmitted.
The government expects to survive.
Others may therefore continue acting as though it will.
This does not determine military outcomes. Leadership cannot stop missiles.
But expectations affect morale, and morale affects military effectiveness.
Afghanistan taught us how rapidly a state can collapse once enough people believe collapse is inevitable.
Ukraine, so far, is demonstrating the reverse mechanism.
The comparison should not be pushed too far. Ukraine and Afghanistan have completely different histories, institutions, militaries and political identities.
But the contrast illustrates why state capacity cannot be purchased entirely through equipment.
Ukraine’s political cohesion may currently be generating military value that spreadsheets underestimated.
Russia may have miscalculated this.
Again, we do not know enough yet to say confidently what Putin expected. But any Russian plan based on rapid political collapse would face serious problems if Ukrainian institutions continue functioning and citizens continue resisting.
This is where war exposes the difference between conquering territory and controlling a society.
Russia can destroy Ukrainian military infrastructure.
It can potentially capture cities.
That does not automatically create a government Ukrainians accept.
If the objective extends beyond military concessions towards fundamentally changing Ukraine’s political orientation, the task becomes much harder.
A population that considers an occupying force illegitimate can impose enormous long-term costs even after conventional military resistance declines.
Military victory and political victory remain different things.
That was true for America in Afghanistan.
It would be true for Russia in Ukraine too.
There is another remarkable consequence of the invasion: it may be producing some of the outcomes Putin said he wanted to prevent.
NATO looks more unified than it has in years.
Germany is dramatically increasing defence spending.
European countries are sending weapons to Ukraine.
Finland and Sweden are participating closely in Western discussions.
Eastern European NATO members are receiving additional forces.
This is a classic strategic problem. A state can achieve a tactical objective while worsening the broader environment it was trying to change.
If Russia’s goal was to reduce Western military influence around its borders, an invasion that causes European governments to rearm and NATO to strengthen its eastern presence may be strategically counterproductive even if Russian forces achieve battlefield successes.
That conclusion is still preliminary. The war is four days old.
But it illustrates why policy has to be evaluated through second-order effects.
Actions change other actors’ behaviour.
Russia does not choose policy in a static environment. Its invasion changes Europe, which changes the strategic environment Russia faces afterwards.
This is the same feedback mechanism that made the security dilemma dangerous before the war.
The difference is that the feedback is no longer theoretical.
Perhaps the largest analytical mistake would now be treating everything that happened before Thursday as obviously leading here.
It did not feel inevitable while it was happening.
Diplomats were still negotiating.
Macron believed there was space for compromise.
Scholz travelled to Moscow.
Russia repeatedly denied invasion plans.
Western governments debated exactly how likely an attack was.
Some analysts believed the military mobilisation was primarily coercive diplomacy.
Those possibilities were reasonable given incomplete information.
Now we know which possibility occurred.
That should update our understanding without rewriting our memory of uncertainty.
Hindsight creates an illusion that history moves towards the outcome that eventually happened.
People living through history do not experience it that way.
They experience probabilities.
This matters because the next decisions are also being made without knowledge of the future.
Will Kyiv hold?
Will negotiations begin?
How severe will sanctions become?
Will Russia escalate?
Will Western unity survive economic costs?
Will the conflict remain inside Ukraine?
Nobody knows.
Governments are again choosing among risks rather than certainties.
The priority now should be preserving Ukrainian sovereignty while preventing the war from expanding into a direct NATO-Russia confrontation. Those objectives can conflict. The more assistance NATO gives Ukraine, the more Russia may perceive Western involvement. Too little assistance makes Ukrainian defeat more likely. Direct intervention creates risks that are qualitatively different because nuclear weapons sit behind the confrontation.
There is no risk-free policy.
That sentence has appeared repeatedly in my writing this year because it seems to describe politics better than most slogans do.
The question is which risks governments are willing to accept, who bears them and what objective justifies them.
For Ukrainians tonight, these are not theoretical questions about deterrence or alliance design. People are sleeping in metro stations. Families are crossing borders. Soldiers and civilians are being killed.
That human reality should constrain how abstract the analysis becomes.
International relations often turns countries into units on a map: Russia wants security, NATO wants credibility, Europe wants stability, Ukraine wants sovereignty.
Countries do not suffer.
People do.
The purpose of analysing strategy is not to make war sound intellectually interesting. It is to understand why political systems failed to prevent people from experiencing what Ukrainians are experiencing now.
For months, the question was whether deterrence could prevent Russia from deciding that invasion was worth the cost.
It could not.
The question now becomes whether the international response can ensure that using military force to change another country’s political future ultimately costs more than it gains.
That matters for Ukraine.
It also matters for every government watching what happens next.
Because international rules do not enforce themselves. They survive when violating them produces consequences strong enough that other states decide violation is not worthwhile.
Russia has tested that system.
Ukraine is paying the immediate price.
The rest of the world now has to decide what the test means.
