The alliance has eight times more money. Moscow has more warheads and a greater willingness to take risks. The most terrifying war could begin with someone else's miscalculation, without any order to launch an offensive at all.
On the night of September 10, 2025, something happened over Poland that Europe had not witnessed since the Cold War. Around two dozen drones crossed the Polish border, and Dutch F-35s, together with Polish fighter jets, opened fire on Russian targets over the territory of a NATO ally for the first time in the alliance's history. Warsaw requested consultations under Article 4 of the North Atlantic Treaty. Nine days later, three Russian MiG-31s spent about twelve minutes in Estonian airspace near the island of Vaindloo, prompting Tallinn to request consultations as well.
No one was killed. No one declared war.
Yet during those ten days, Europe saw the anatomy of a major confrontation unfold at full scale for the first time: an incident, a response, an interpretation of that response, then the next move. The central question now being discussed in military headquarters is simple: who wins if, one day, that chain is not broken?
A war between Russia and NATO is possible. But the possibility of war, the decision to start it, and the ability to achieve victory remain three different things. The most dangerous mistake in this debate is the belief that the outcome can be determined by simply counting tanks, aircraft, or billions of dollars.
In a prolonged conventional war against a united NATO with full U.S. participation, Russia would find itself at a significant disadvantage. A limited regional conflict could allow it to achieve an initial success and then attempt to convert that success into a political victory. In a large-scale exchange of strategic nuclear strikes, the very concept of victory loses its meaning. Between these scenarios lies an enormous zone of uncertainty, and it is precisely there that decisions capable of triggering war are made.
Eight to One: The Number That Misleads Everyone
First, money, because that is where almost everyone begins.
Data published by September 2026 allow the balance of resources to be estimated, but they require caution. Complete annual spending statistics mostly refer to 2025, while many figures for 2026 remain plans and preliminary estimates.
According to the Stockholm International Peace Research Institute, SIPRI, all 32 NATO countries together spent about $1.581 trillion on defense in 2025. Russia's spending was estimated at $190 billion. The difference is roughly 8.3 to one.
The United States alone spent $954 billion, about five times as much as Russia. NATO's European members, using SIPRI's grouping, spent $559 billion, nearly three times the Russian figure. Germany spent $114 billion, the United Kingdom $89 billion, France $68 billion, and Poland $46.8 billion. The combined spending of Germany, the United Kingdom, and France reached $271 billion, exceeding Russia's by about 43 percent. Add demographics and economic output, and the contrast becomes even sharper: nearly one billion people live in NATO countries, compared with roughly 146 million in Russia, while the alliance's combined GDP exceeds Russia's by more than twentyfold.
The numbers are impressive. But they cannot be translated into the claim that NATO is "eight times stronger than Russia."
Military spending is converted at market exchange rates, while the cost of labor, manufacturing, repairs, and construction differs enormously among countries. The same amount of money buys vastly different quantities of work and output inside Russia and inside the United States. Two authoritative estimates illustrate the problem well. SIPRI calculates at market exchange rates. The London-based International Institute for Strategic Studies, IISS, recalculated Russia's 2024 military budget using purchasing power parity in its annual military balance and arrived at about $462 billion, more than all European states were spending combined. Both estimates are legitimate; they simply measure different things. The first reflects weight in the global economy. The second comes closer to measuring how many tanks, shells, and soldiers' salaries the ruble can actually buy domestically. For battlefield arithmetic, in my view, the truth lies closer to the second figure than to the first.
The U.S. budget also finances a global military presence. It pays for fleets in the Pacific, bases in the Middle East, and commitments to Japan, South Korea, and Taiwan. Imagining that the entire American military establishment can simply be transferred to one European theater while the rest of the world is ignored is a fantasy that no one at the Pentagon shares.
The structure of spending matters as well. A billion dollars allocated to a system that will enter service ten years from now is not equivalent to a billion spent on ammunition arriving today. Pensions, base maintenance, nuclear forces, and purchases of field radios all fall under the broad category of defense expenditure, but they affect the outcome of a particular battle in radically different ways.
Money defines the limits of what is possible. Force readiness determines what can be done immediately.
A War of Economic Attrition: How Much Longer Can Moscow Pay?
Despite all these qualifications, the economic gap remains one of NATO's strongest advantages in a long war. A major conflict requires the constant replacement of losses, repair of equipment, training of personnel, and production of ammunition. Initial stockpiles are consumed within weeks, while industry has to keep functioning for years.
Russia is already carrying a heavy military burden. According to SIPRI, it devoted about 7.5 percent of GDP to military purposes in 2025. In a March study, the institute estimated Russian military spending that year at roughly 16 trillion rubles. The initial 2026 plan called for about 14.9 trillion rubles, or 6.3 percent of GDP, although the authors warned that budget revisions were likely, making it inappropriate to treat planned figures as final.
For comparison, during the final years of the Soviet Union, defense consumed, by various estimates, between 15 and 25 percent of Soviet GNP. I remember well how that race ended. An army that only a few years earlier had been described as the second most powerful in the world was disintegrating before our eyes, while officers in the late 1980s and early 1990s were struggling to feed their families. Today's burden is far below the Soviet level, and the comparison is not direct. It simply serves as a reminder: every military economy has a price, even if that price is not paid immediately.
A heavy burden does not automatically mean collapse. A state can redirect resources toward the military for a long time. But additional military spending intensifies competition for skilled workers, machine tools, components, credit, and budget funding. Continuing a war is one task. Waging it without mounting economic costs is something entirely different.
NATO has a much larger financial reserve, but mobilizing it depends on decisions by dozens of governments. At the Hague summit in June 2025, the allies adopted a benchmark for 2035: 3.5 percent of GDP for core defense spending and another 1.5 percent for defense- and security-related expenditures. No one intends to spend the entire 5 percent on weapons. According to NATO's July update, five allies expected to reach the core 3.5 percent benchmark as early as 2026. The countries leading the way are telling: Poland, the Baltic states, and Scandinavia. The closer a capital is to the Russian border, the faster its defense budget is rising.
Years can pass between approving a budget and fielding a combat-capable formation. A factory needs equipment and engineers. An aircraft needs a trained pilot and technical crews. A brigade needs commanders, transportation, communications, maintenance units, and stockpiles.
That is why NATO's potential superiority does not eliminate its vulnerability at the outset.
Sixty Hours to Tallinn: Anatomy of the First Strike
For Russia to achieve a limited success, it does not need to defeat the entire alliance simultaneously. In theory, it would be enough to concentrate forces along one axis, achieve a result before reinforcements arrive, and confront the opponent with a choice: accept what has happened or widen the war.
This describes a possible logic of action, not evidence that a decision to attack has been made.
Ten years ago, the American RAND Corporation conducted a series of war games on the defense of the Baltic states. The conclusion, published in 2016, shocked Brussels: under the force posture then in place, Russian units could reach the outskirts of Tallinn and Riga within 36 to 60 hours. The authors concluded that credible defense would require roughly seven brigades, including three heavy brigades. Much has changed since then. Multinational forward-presence battlegroups have been deployed in the Baltic states and Poland. Germany is establishing a permanent armored brigade in Lithuania, deploying a major formation abroad for the first time since World War II.
In February 2025, Danish military intelligence published an assessment outlining hypothetical timelines for the restoration of Russian capabilities after the end or freezing of the war in Ukraine. It estimated about six months for Russia to regain the ability to wage a local war against a neighboring state, about two years for a regional war against several countries in the Baltic region, and about five years for a major European war without U.S. participation. These timelines came with major conditions: Russia would have to free up resources, while NATO would have to refrain from expanding its own capabilities at a comparable pace. The document assessed the development of military potential; it did not establish a date for a future attack.
The ability to start a war does not prove an intention to do so. But an erroneous belief in an adversary's weakness can bring such a decision closer.
In the first phase, everything would be decided by the forces already present in the crisis zone. The total number of units across all NATO countries is not the same as the number that can immediately enter combat. A brigade already deployed and a brigade that still has to be manned, receive equipment, and cross several borders on overloaded rail networks have very different military value during the opening days of a war. The most vulnerable point has long been known to every military headquarters: the Suwalki Corridor, a roughly 65-kilometer stretch of the Polish-Lithuanian border between Russia's Kaliningrad region and Belarus. Cutting it could isolate the Baltic states from land-based supply routes. That is why the Zapad 2025 exercises held in Belarus last September were monitored hour by hour in Warsaw and Vilnius.
Geography has also changed in NATO's favor. Finland joined the alliance on April 4, 2023, and Sweden on March 7, 2024. Finland's roughly 1,340-kilometer border with Russia became a NATO frontier, approximately doubling the alliance's total land border with Russia. The Baltic Sea became almost an internal NATO sea, while St. Petersburg and the Baltic Fleet base at Kronstadt came within reach from both shores. Finland also brought with it one of Europe's largest artillery forces and a universal reserve training system that it never dismantled after the Cold War. At the same time, enlargement creates new obligations for the alliance: it must now defend a far longer frontier.
The alliance also drew conclusions from the September incidents. On September 12, 2025, NATO launched Operation Eastern Sentry, strengthening air patrols on the eastern flank. Earlier, at the 2022 Madrid Summit, it approved a new force model under which more than 300,000 troops were to be kept at high readiness and moved to a threatened sector within thirty days. Readiness on paper and actual deployment differ as dramatically as a budget differs from a brigade. Exercises in recent years have repeatedly run into the same obstacles: bridges unable to support the weight of modern tanks, shortages of rail platforms, and permit procedures for crossing internal European Union borders. Military planners call this problem "military mobility," and Brussels is allocating billions of euros to address it.
Russia has both an advantage and an Achilles' heel in this theater. The Kaliningrad region, densely equipped with air-defense systems and Iskander missile complexes, can complicate NATO air and naval operations over the southern Baltic. Yet the enclave is surrounded on all landward sides by NATO territory, and sustaining it in a prolonged war would become an exceptionally difficult task. The alliance's principal advantage in a conventional war still lies in the air: thousands of modern combat aircraft, including hundreds of fifth-generation fighters. The experience of Ukraine, however, shows that dense, layered air defenses can deny even a powerful air force uncontested control of the skies for a very long time.
The notion of an easy Russian offensive followed by the inevitable collapse of Baltic defenses has no reliable basis. The opposite promise of an immediate rout of Russian forces is no better supported.
Any initial advance would still have to be consolidated. Captured territory requires supply and defense. The opponent retains the ability to bring up reserves, strike rear areas, and alter the very nature of the war.
An Army Learning in Blood
Russia's combat experience cannot be assessed in a single dimension. The war in Ukraine is exhausting the armed forces, but it is also forcing them to learn. According to Western estimates, total Russian casualties, killed and wounded combined, were approaching one million by mid-2025. At the same time, the military mastered the mass use of drones, electronic warfare, real-time reconnaissance, and new methods of artillery employment. On the night of September 7, 2025, according to the Ukrainian Air Force, Russia launched more than 800 attack drones and decoys against Ukraine in a single strike. No NATO military has yet had to repel an attack of that scale. European armies also lack comparable experience in high-intensity warfare.
In a 2025 study, IISS suggested that through production, refurbishment, and organizational changes, Russian ground forces may have approached their prewar quantity of equipment. Eliminating equipment shortages in units that already existed was projected for late 2026 to early 2027. Equipping the planned expanded force could take until 2030 to 2035 if current rates are sustained. A 2024 presidential decree increased the authorized strength of Russia's active military personnel to 1.5 million. All of these are conditional projections and planning figures, not confirmation that restoration has been completed.
Heavy losses therefore do not make Russia a harmless adversary. Restoring the number of vehicles is also not the same as restoring military quality. An experienced commander, a well-integrated crew, and a cohesive battalion do not roll off the production line with a new tank.
Is America Pulling Back? The Central Unknown in Europe's Defense
For NATO, U.S. participation is decisive. The American contribution includes combat forces, intelligence, satellites, data processing, aerial refueling, and the logistics required for large-scale operations. A study published by IISS in the summer of 2026 noted that without U.S. intelligence support, battlefield surveillance would become less continuous, early warning would deteriorate, and the use of long-range strike systems would become significantly more difficult.
This creates a fundamental distinction. A war against the entire NATO alliance with full U.S. participation and a war against European allies with only limited support from Washington are two different conflicts with two different sets of probabilities.
In the second case, Europe would not become defenseless. It would, however, have to close critical capability gaps in the middle of a crisis, which is the worst possible time for improvisation. My assessment is that limited U.S. participation would primarily increase the likelihood of an initial Russian success. It would not guarantee Moscow victory in a prolonged war.
Article 5: A Promise Without Autopilot
Political cohesion is equally important. Article 5 of the North Atlantic Treaty requires assistance to an ally that has been attacked, including the possible use of armed force. At the same time, the treaty leaves each country the right to determine the actions "it deems necessary." The text contains no single automatic military response to every incident. In the alliance's 76-year history, Article 5 has been invoked exactly once, on September 12, 2001, the day after the attacks on New York and Washington.
That does not make the commitment fictitious. But it creates room for a dangerous miscalculation: an adversary may conclude that the allies will argue for too long or refuse to take serious risks altogether.
A war could begin precisely with such a miscalculation.
A Staircase Without Railings: How Wars Begin That No One Wanted
A major war does not require an initial order for a large-scale offensive. A collision between aircraft, an armed incident at sea, a strike causing mass casualties on allied territory, or the gradual expansion of retaliatory actions could be enough.
Each successive step appears limited to the side taking it. The other side sees preparation for a larger attack. Failure to respond is interpreted as weakness, while a response becomes justification for the next strike.
The precedents are well known. In November 1983, NATO's Able Archer exercise, which rehearsed nuclear strike procedures, was taken seriously in Moscow as possible cover for a real attack, and some Soviet forces were placed on heightened alert. The world passed close to catastrophe without even realizing it. There is an opposite example as well. On November 24, 2015, a Turkish F-16 shot down a Russian Su-24 near the Syrian border, killing the pilot. Turkey was a NATO member, Russia a nuclear power, and bilateral relations collapsed. Yet war did not follow, and within six months the two sides were already restoring contacts. Escalation, in other words, is not a law of nature: diplomacy and deterrence can stop it.
Public rhetoric over the past year has made this staircase steeper. On December 2, 2025, Russian President Vladimir Putin said: "We are not planning to go to war with Europe, but if Europe suddenly wants to fight us and starts, then we are ready right now." Nine days later, on December 11, NATO Secretary General Mark Rutte, speaking in Berlin, delivered a line repeated across European capitals: "We are Russia's next target." Each side insists that it is merely responding to the other's threat. It is the classic security dilemma in its purest form.
The Shells That Are Missing: The Trap of Impressive Numbers
In a prolonged confrontation, the advantage will go to the side that remains combat-capable after the initial stockpiles are exhausted. This is where simplified comparisons become especially dangerous.
The number of shell casings manufactured is not the same as the number of complete rounds ready for use. Announced production capacity is not the same as actual deliveries. Reports on armored vehicle deliveries often combine newly manufactured and refurbished platforms. Thousands of vehicles in storage reveal nothing about how many are operational.
The European Union set a goal of reaching annual production of two million artillery shells by the end of 2025, while Western estimates of Russian output ranged from three to four million per year. Millions more shells supplied by North Korea were added to the equation. These figures migrate from report to report, but they almost never specify exactly what is being counted: shell bodies, complete rounds, 152-millimeter or 155-millimeter ammunition, newly manufactured or refurbished stocks. Reliable analysis requires comparing like with like. Attractive aggregate figures without such clarification merely create the appearance of precision.
4,400 vs. 4,215: The Arithmetic of the End of the World
There is a threshold beyond which industrial and economic superiority ceases to provide an acceptable path to victory. That threshold is nuclear weapons.
According to SIPRI's January 2026 estimate, Russia possessed roughly 4,400 warheads in its military stockpile, of which 1,718 were considered deployed. Its total inventory, including warheads awaiting dismantlement, was estimated at 5,545. The United States had a military stockpile of about 3,700 warheads, including 1,770 deployed, and a total inventory of 5,042. France's military stockpile was estimated at 290 warheads, including 280 deployed, while the United Kingdom had 225, including 120 deployed. All of these figures are approximate. A deployed warhead is not necessarily maintained in a state of immediate launch readiness.
Arithmetically, the combined military stockpiles of the United States, France, and the United Kingdom amount to 4,215 warheads versus Russia's 4,400. But three national arsenals cannot simply be treated as a single centralized force: each country independently decides whether to use its nuclear weapons. Another detail is often overlooked: a substantial part of Russia's numerical advantage consists of nonstrategic, or so-called tactical, nuclear weapons, precisely the category assigned the role of an instrument of "limited" escalation in military debates. Since 2023, some of these weapons have been stationed in Belarus. In November 2024, Moscow updated its nuclear doctrine, stating that aggression by a nonnuclear state carried out with the participation or support of a nuclear-armed state would be regarded as a joint attack by both.
For deterrence, the decisive factor is the ability to inflict unacceptable retaliatory damage; raw numbers are secondary. A modest numerical advantage in warheads does not provide the ability to disarm an adversary and escape retaliation.
The expiration of New START on February 5, 2026, added another layer of uncertainty. For the first time in half a century, the world's two largest nuclear powers were left without legally binding limits on their strategic arsenals and without mutual inspections. Less transparency means more room for worst-case assumptions.
Another precedent is relevant here. In the spring of 1999, India and Pakistan, both already nuclear-armed, fought in the mountainous Kargil region. Pakistan calculated that its nuclear umbrella would allow it to consolidate a limited territorial gain. India responded with conventional forces but did not cross the Line of Control, and two months later Pakistani units withdrew. A limited war under the shadow of nuclear weapons proved possible. The side that started it did not secure victory.
NATO may possess the resources to repel an attack without having a safe means of forcing Russia into unconditional surrender. Russia may retain the ability to inflict catastrophic damage on NATO without possessing a reliable path to victory over the entire alliance.
It is therefore essential to distinguish between defensive success and an attempt to completely defeat a nuclear power. For NATO, victory would mean repelling an attack, preserving allied territory, and bringing the war to an end. A limited Russian political victory could theoretically consist of coercing the opponent into concessions or undermining trust within the alliance. A local success in the opening days guarantees neither outcome.
Five Billion: The Price of the Word "Victory"
In a large-scale nuclear exchange, the question itself changes. A study published in Nature Food in 2022 estimated that one scenario involving a Russia-U.S. nuclear war could result in more than five billion deaths from a global food crisis. The result depends on specific assumptions about fires, soot injected into the stratosphere, and declining agricultural yields, so it cannot be treated as a precise prediction of the consequences of every nuclear war. The scale, however, makes the central point clear: the catastrophe would not be confined to the countries directly at war.
From Baku, this arithmetic is especially visible. The South Caucasus lies between the two sides of this hypothetical conflict, the Middle Corridor passes through the region, and Azerbaijani gas, about 13 billion cubic meters of which was exported to Europe in 2024, has become part of European energy security. A major war in Europe would strike everyone connected to it by pipelines, railroads, and trade, regardless of which flags flew over the trenches.
Forty years ago, the leaders of the two superpowers had already found the right formula. In November 1985 in Geneva, Ronald Reagan and Mikhail Gorbachev signed a joint statement declaring: "A nuclear war cannot be won and must never be fought." Thirty-seven years later, in January 2022, the five permanent members of the United Nations Security Council repeated the same sentence word for word. A month and a half later, the largest war in Europe since 1945 began.
My Verdict
In a short regional conflict, either side could achieve an initial success, depending on concrete readiness in the specific area where the fighting begins. In a prolonged conventional war against a united NATO with full U.S. participation, the alliance's aggregate resources give it substantially better prospects. If American participation were limited, the position of the European allies would become significantly more difficult, but a Russian victory would still not be predetermined. A large-scale strategic nuclear exchange would leave no sustainable winner in political or economic terms.
My forecast is testable. There will be no direct war between Russia and NATO before the end of 2028. But before the end of 2027, at least one more ally will request consultations under Article 4 because of Russian drones, aircraft, or activities in the Baltic, while the number of air and maritime incidents over the Baltic region will continue to grow.
The most dangerous situation will arise if Moscow concludes that NATO will not respond, while Western capitals conclude that Russia will not dare to act. Both sides would then discover that their assumptions were wrong only after the fighting had begun, when correcting those mistakes would already be too late.
Each side has enough strength to disrupt the other's plans. Neither has any guarantee that a war it starts will end on terms it finds acceptable.
Three MiG-31s spent about twelve minutes over the Estonian island of Vaindloo. The island is tiny, with an old lighthouse standing on it. In the fall of 2025, those twelve minutes were enough to trigger consultations in Brussels. Next time, they may be enough only for a decision.