Does a NATO-Russia War Have to go Nuclear? Not Necessarily

Russia’s nuclear arsenal is terrifying. That doesn’t mean Europe gets to use nuclear war as an excuse for failing to build conventional defenses

An unarmed Minuteman III intercontinental ballistic missile launches during an operational test at Vandenberg Air Force Base, Calif. DoD

Hey friends.

There’s one response I hear almost every time I write about the possibility of a future Russian attack on NATO. “Well, if NATO and Russia ever start actually shooting at each other, we’re all dead anyway.”

Not necessarily. And that line of thinking is dangerous in and of itself because it gives Putin a deterrence win and a lot of freedom to maneuver.

Sure, Russia has thousands of nuclear weapons. So does the United States. Britain and France have their own arsenals. NATO and Russia trading fire sounds like the opening five minutes of a movie where somebody eventually gets vaporized while staring wistfully at a photograph of their family. Spielberg eat your heart out.

I understand the instinct.

This is also why the fictional depictions of a US/Russia war, like the 1984 movie Red Dawn or Tom Clancy’s Red Storm Rising were so striking. They each show a non-nuclear conflict which, at the time, subverted expectations.

But I also think this idea, that a war with Russia must go nuclear, is the single most dangerous idea currently floating around European defense debates. History tells us that conventional war between nuclear powers has stayed conventional every time it’s actually happened.

The Cold War was stranger than we remember

Most people remember the Cold War as forty-five years of Washington and Moscow glaring at each other over piles of nuclear weapons, wisely never pulling the trigger.

Broadly true, I suppose. Except for the part where Americans and Soviets actually shot each other.

During the Korean War, Soviet pilots secretly flew MiG-15s against American aircraft over what became known as MiG Alley. Moscow disguised the whole operation, painting Chinese or North Korean markings on the aircraft and ordering pilots to speak Korean phrases over the radio instead of their own language, a cover story that occasionally fell apart the moment somebody got excited mid-dogfight and started yelling in Russian.

American pilots knew something unusual was happening. US Sabre pilots even nicknamed the more skilled Soviet aviators “Honchos.”

Soviet archives later confirmed what had long been an open secret in Air Force circles, that Soviet combat pilots were directly fighting American pilots in the skies over Korea.

So, the two emerging nuclear superpowers on Earth had their fighter pilots trying to kill one another. The world didn’t end. Although, Stalin restricted where Soviet pilots could operate specifically because Moscow wanted to avoid the fighting turning into an acknowledged Soviet-American war, and UN pilots were similarly barred from hitting the Soviet-run bases across the Chinese border.

Both sides were fighting and both sides were also building fences around the fight in real time. The danger was that nobody could guarantee exactly where the escalation stopped once shooting started, which is why both governments worked so hard to build the fences in the first place.

Korea isn’t the only example.

In 1969, Soviet and Chinese troops fought a series of bloody clashes along their shared border, five years after China tested its first atomic bomb and while the Soviet Union held one of the largest nuclear arsenals on the planet.

They fought anyway, along the Ussuri River and later near the Soviet-Kazakh border, serious enough that Washington grew genuinely alarmed the fighting could spiral into something much larger, and both Beijing and Moscow started nuclear civil-defense preparations as tensions climbed.

Negotiations eventually pulled both governments back from the edge. Nuclear weapons didn’t prevent that conventional combat.

Then Kargil. India and Pakistan openly tested nuclear weapons in 1998. One year later, they fought a real conventional war in Kashmir, infantry at altitude, artillery firing, Indian combat aircraft flying strike missions, soldiers dying by the hundreds.

RAND later described Kargil as proof that even declared nuclear arsenals might not meaningfully dampen military competition between hostile neighbors. India deliberately kept much of that fighting geographically constrained instead of letting it spread across the entire frontier.

Three separate cases. Three sets of nuclear-armed adversaries. None of them ended in a strategic exchange.

Now, I’m very aware that three cherry-picked cases don’t prove that nuclear armed countries can slug it out while avoiding full-scale nuclear war, but there’s actually a name for this pattern in strategic theory: the stability-instability paradox, first developed by Glenn Snyder (PDF).

Nuclear weapons can make the largest possible war less likely while leaving real room for smaller wars underneath that ceiling.

Two adversaries who both know that invading the other’s homeland or threatening its government’s survival risks nuclear catastrophe become extremely cautious about doing those specific things.

That caution creates space below the threshold for border clashes, proxy wars, limited territorial grabs, cyberattacks, sabotage, even localized conventional fighting.

Pakistan’s behavior at Kargil is usually read as nuclear weapons giving its leadership some confidence India would hesitate before dramatically widening the war, even though Pakistan’s broader nuclear signaling ultimately failed to deliver the political outcome it wanted.

The mushroom cloud becomes the ceiling nobody wants to hit.

But there’s still a lot of unpleasant, survivable territory sitting underneath it.

NATO describes its deterrence posture as a deliberate blend of conventional, missile-defense, and nuclear capability, and says the circumstances where nuclear weapons would actually be used as “extremely remote.”

Here’s why I think this distinction should reshape how Europe thinks about its own rearmament right now. If conventional deterrence were pointless in the shadow of nuclear weapons, none of that Cold War fence-building would have mattered.

Korea would have stayed contained by luck alone. It didn’t. It stayed contained because both governments actively, deliberately built limits into the fighting, restricted their own pilots, chose targets carefully, and managed the escalation on purpose.

Fences don’t build themselves.

Somebody has to want to build one badly enough to pay for it, and that’s exactly the muscle a fatalistic Europe risks losing.

The nightmare scenario, and where the actual danger hides

Picture Russian forces crossing into Estonia.

Estonian troops fire (as they have publicly said they will shoot marked or unmarked invaders on sight). Article 5 works as intended. Other NATO forces engage. British or German aircraft strike Russian armored formations sitting inside Estonian territory.

Nobody has yet launched a nuclear weapon. Now the real decisions begin.

Does NATO strike artillery firing from inside Russia itself?

Does it hit Russian air-defense batteries across the border?

Does it reach for airfields launching attacks against NATO forces?

Does it strike command centers hundreds of kilometers deep in Russian territory?

Remember, Russia now has one of the strongest deep strike complexes on the planet, as we’ve seen used against Ukraine almost daily.

Does Moscow respond by hitting bases in Germany or Britain? At what specific point does either side start reading the other’s conventional operations as preparation for something far larger?

That’s where the danger lives, in the accumulation of individual decisions, not in the first gunshot.

A major new RAND Europe study published this month makes the problem a bit uglier, examining how a conventional NATO-Russia conflict could accidentally climb toward nuclear escalation even when neither side initially intends to use nuclear weapons at all.

The researchers identified several especially dangerous pathways: Conventional strikes might hit command, control, and intelligence systems that also happen to support Russia’s nuclear forces.

NATO could attack Russian weapons capable of carrying either conventional or nuclear warheads, leaving Moscow to wonder whether its nuclear capability was being deliberately dismantled piece by piece.

Missile-defense operations could interact badly with Russian fears about a disarming first strike. And conventional operations could eventually start looking, from Moscow’s side, like a genuine threat to the survival of the Russian regime or Russian society itself.

Imagine NATO detects a Russian radar supporting conventional operations in Estonia and destroys it with a long-range precision weapon, an entirely routine wartime targeting decision. Except that same radar might also feed Russia’s strategic nuclear warning architecture.

Now Moscow has a real question sitting in front of it: Was NATO simply destroying a battlefield sensor, or was NATO quietly starting to blind Russia’s nuclear early-warning systems ahead of something bigger?

Intent and perception have just come apart. The same problem runs through dual-use systems like Iskander launchers, which can fire either conventional or nuclear missiles from the identical platform.

NATO might strike a launcher purely because it’s raining conventional missiles down on NATO forces. Russian commanders watching several of those same dual-capable launchers disappear one after another may read a completely different story into it.

Russia has, deliberately, made this line blurrier still.

Putin revised Russia’s public nuclear doctrine in November 2024. The older 2020 language reserved nuclear use against a conventional attack only when “the very existence of the state” was threatened. The revised version drops that bar, allowing possible nuclear use when conventional aggression creates a merely “critical threat” to Russian or Belarusian sovereignty or territorial integrity.

That’s supposed to sound alarming, because ambiguity is itself a deterrence tool, and Moscow wants NATO commanders unsure exactly where the real line sits.

Still, treat that doctrine as political signaling rather than an automated trigger.

Carnegie’s Nicole Grajewski makes this point: Putin isn’t going to sit in the Kremlin during a war with a highlighter, find paragraph seventeen, and announce that subsection C means it’s nukin’ time.

Actual nuclear use would remain a real political decision, shaped by the specific circumstances of the moment.

A terrifying decision. But still a decision, made by a person, weighing consequences, not a mechanism that fires itself the instant a threshold gets crossed on paper.

NATO needs to get deadly serious about Article 5

There’s a second, more realistic nightmare sitting underneath the first one, and this is what I think will happen in the next twelve months.

Putin knows he can’t take on the whole alliance at once. Thirty-two countries, a combined military and economic weight Russia has no realistic path to matching, especially while still in Ukraine, and an Article 5 guarantee that, on paper, turns any attack on one member into an attack on all of them.

So instead of testing whether NATO would fight a real war, the smarter Putin play is testing whether NATO would fight at all, over something engineered to look too small to bother with.

I’m not talking about sabotage or cyber.

Picture a hundred unmarked Russian troops crossing into a stretch of forest in Lithuania, just inside the border with Belarus.

No uniforms that clearly identify them.

No obvious command structure visible from a satellite photo.

They occupy roughly ten square kilometers of woodland with zero strategic value; no town, no infrastructure, nothing worth defending on any military planner’s map, chosen specifically because it sits a few hundred meters on the Lithuanian side of an ambiguous, heavily forested border.

Lithuania, understandably furious, invokes Article 5.

Now watch what happens next, because this is the actual test Moscow is running…

How many European governments want to risk a shooting war with a nuclear-armed power over ten square kilometers of pine trees nobody’s ever heard of?

How many parliaments, presented with grainy footage of unmarked men in an empty forest, vote to authorize real military force in response?

I’d put real money on hesitation from at least a few capitals, and I’d put even more money on hesitation from Washington specifically.

Trump has, on the record, questioned whether the Baltic states have “fulfilled their obligations” to warrant American protection, and told a campaign crowd in 2024 that he’d encourage Russia to do “whatever the hell it wants” to NATO members he judged delinquent on defense spending.

But even if we had a sane president in the White House, I would bet even money that most American leaders would ask the question, at least in private: “Why send US soldiers to fight and die for ten square kilometers of land that’s strategically worthless?”

It’s a cost/benefit analysis, where the cost could be huge and the benefit seemingly small.

But here’s why that hesitation, if it comes, would be absolutely catastrophic: Article 5 has never actually been about the specific patch of ground it gets invoked over. It’s a credibility mechanism. The instant that promise gets treated as negotiable, even ONE SINGLE TIME, even over ten worthless square kilometers, it stops functioning as a deterrent for every square kilometer that comes after it.

Putin only needs NATO’s own governments to prove once, in public, that the guarantee comes with an asterisk attached. *

That’s a different danger than the Estonia scenario above. The first nightmare is about losing control once real fighting starts: an escalation problem.

This one is about the alliance losing its nerve before real fighting ever has the chance to start: a credibility problem.

But they’re connected… The fear of escalating into nuclear war is exactly what Putin is counting on to destroy NATO credibility.

And it’s why conventional rearmament, as important as it is, can’t be the whole answer on its own. You can build every tank and stockpile every artillery shell in Europe, and none of it matters if the political will to actually use them evaporates the moment Moscow hands you a scenario small enough to make hesitation look politically reasonable.

This is Putin’s trap. And shame on us if we fall for it.

The hardware question and the willpower question must get solved together, or building the hardware alone just buys Russia a more convincing bluff to call.

Still, the fatalistic version of a NATO-Russia war skips over almost every one of the choices sitting between the first gunshot and any nuclear exchange.

Russia attacks Estonia.

Somebody has to decide whether NATO confines its own response strictly to NATO territory.

Somebody decides whether Russian territory gets struck at all.

Somebody decides how deep those strikes reach.

Somebody decides which specific Russian military systems become fair targets.

Somebody decides whether nuclear command infrastructure stays off limits entirely.

Somebody decides whether Russia introduces tactical nuclear weapons of its own.

Somebody, on the other side, decides how NATO answers that.

Somebody decides whether the whole thing keeps climbing or stops right there.

None of those choices are safe. But my point is, none of them are automatic either.

Now, before people start calling me a warmonger, none of this should read as “a NATO-Russia war would be fine.”

To be fair, it would probably be the most dangerous conventional conflict in human history, every long-range strike carrying real danger and political weight, every attack on Russian air defense or command infrastructure getting scrutinized for nuclear implications, every missile warning raising fresh questions about payload and intent.

But the historical record still pushes back hard against the opposite claim, that nuclear war becomes inevitable the instant nuclear powers exchange conventional fire.

Russia maintains hundreds of thousands of conventional troops and keeps building tanks, artillery, aircraft, drones, and air-defense systems.

They’re doing that because they understand, correctly, that plenty of real military contests can happen beneath the strategic nuclear threshold, and conventional strength is what determines who wins or loses those contests.

That is exactly why Europe can’t afford to treat conventional preparation as a wasted gesture standing pointlessly in the nuclear shadow. Every euro spent on ammunition stockpiles, air defense, and standing brigades is a euro spent building the same kind of fence that kept Korea, the Ussuri River, and Kargil from becoming something far worse.

Nuclear deterrence was never a magical force field that physically prevented American and Russian soldiers from shooting at each other, and it won’t be one for Europe and Russia either.

What nuclear weapons actually changed was the price of losing control once the shooting starts.

Treating nuclear war as automatic hands nuclear threats far more political power than they’ve ever earned and lets Moscow win an argument it hasn’t won on the actual battlefield.

Europe’s job right now isn’t to assume that argument is already lost. It’s to build the conventional force strong enough to make sure Russia never gets to find out whether it was right.

Слава Україні!