In a little under three decades, nuclear weapons have gone from center stage to a sideshow in U.S. defense strategy. Since the 1990s, the United States has drastically reduced its stockpile and concentrated on its conventional and irregular warfare capabilities. Nuclear weapons policy has focused overwhelmingly on stemming proliferation to countries such as Iran and North Korea, and prominent political and national security figures have even called for abolishing nuclear weapons altogether. What was once the core of the country’s Cold War strategy has been reduced to an afterthought.

Immediately after the Cold War, when the United States enjoyed unprecedented global power, this approach seemed reason­able. Washington didn’t need much of a nuclear strategy against Iraq or Serbia. But now, great-power competition has returned. Russia wants to upend the post–Cold War status quo in Europe. A rising China seeks ascendancy, first over Asia and ultimately beyond. To accomplish this, each country has developed military forces ideally suited to fight and defeat the United States in a future war. And modern, mobile nuclear capabilities are a key part of their strategies.

These capabilities could allow Russia or China to pressure or attack U.S. allies and to block any efforts by the United States to fight back. This should cause great alarm among U.S. policymakers: American grand strategy is rooted in a network of alliances designed to maintain favorable regional balances of power and protect U.S. access and trade across the globe. These alliances work as long as they can be credibly defended against outside challengers. But if Russia and China can win wars against the United States in Europe and Asia, respectively, then these revisionist states will press their advantage—with painful and possibly disastrous consequences for U.S. interests in the world. 

Washington’s task is clear. It must demonstrate to Moscow and Beijing that any attempt to use force against U.S. friends and allies would likely fail and would certainly result in costs and risks well out of proportion to whatever they might gain. This requires conventional military power, but it also means having the right strategy and weapons to fight a limited nuclear war and come out on top.

In a little under three decades, nuclear weapons have gone from center stage to a sideshow in U.S. defense strategy.

For the first time in a generation, then, getting U.S. defense strategy right means getting nuclear strategy right. This requires more than just modernizing the current arsenal of immensely destructive strategic nuclear weapons and their delivery systems. This arsenal, designed to inflict unimaginable damage in an apocalyptic war, is necessary to deter the gravest forms of attack. But threatening to use such weapons in a limited war in defense of allies thousands of miles from U.S. shores is just too extreme to be convincing and therefore unlikely to work. 

Instead, the United States needs weapons systems that can bridge the wide gulf between conventional and all-out nuclear war. In particular, Washington should step up its efforts to develop low-yield tactical nuclear weapons and associated strategies that could help blunt or defeat a Russian or Chinese attack on U.S. allies without provoking a nuclear apocalypse. Demonstrating to potential opponents that the United States has this ability is the best way to avoid ever having to put it into practice. 


During the Cold War, nuclear weapons formed the centerpiece of U.S. strategy. Initially, when the United States enjoyed vast nuclear superiority over the Soviet Union, it relied on the threat of an immediate and decisive nuclear attack to deter aggression in Europe. By the early 1960s, U.S. strategic forces dwarfed the Soviet Union’s. NATO’s defenses in Western Europe bristled with nuclear weapons, while conventional forces largely played second fiddle. As the Soviet nuclear arsenal ballooned and the United States’ advantage faded, however, Washington decided that this strategy was no longer enough to credibly defend Western Europe. As a result, it reinvigorated its conventional forces and devised strategies for limited nuclear use designed to blunt a Soviet invasion and persuade Moscow to end any war short of nuclear Armageddon. Thus, although Washington continued investing in strategic nuclear forces, it also developed tactical nuclear weapons and capabilities designed to offset the Warsaw Pact’s much larger conventional forces. Thankfully, these strategies never had to be put to use, probably because they were credible enough to dissuade the Soviet Union from risking a major offensive—a testament to their value for deterrence.

After the collapse of the Soviet Union, the United States turned its focus to the rogue states that now posed the main, if far more modest, threat to its interests. U.S. conventional forces demonstrated their ability to quickly defeat such foes, whether Saddam Hussein’s army in Iraq in 1990–91, Serbian forces in 1998–99, or the Taliban government in Afghanistan in 2001. If nuclear strategizing had seemed morbidly excessive during the Cold War, it seemed positively absurd in this world of U.S. dominance. 

Accordingly, Washington’s emphasis shifted to conventional forces that could be used for preventive attacks and regime change abroad. The United States dramat­ically downsized its nuclear forces and reduced their role in its defense strategy. Concerns about nuclear weapons now focused on fears about their acquisition by rogue states or terrorists. As a result, successive administrations worked to contain proliferation and to delegitimize the use of nuclear weapons except in the narrowest of circumstances. This approach was appealing: given the United States’ unrivaled conventional military might, pushing nuclear weapons out of the picture seemed like it would only solidify U.S. power. 

Moreover, the strategy enjoyed support from across the political spectrum. It was no surprise that doves applauded getting rid of the weapons they so loathed, but even hawks welcomed the shift. Nuclear weapons, after all, tend to raise the threshold for military action. Thus, President George H. W. Bush cut over 5,000 warheads from the stockpile in 1992. Every administration after him—Democratic and Republican—continued the drawdown. All in all, the U.S. nuclear arsenal has shrunk to a fraction of its Cold War size.


But if this approach once made sense, it no longer does. Russia and China have made impressive strides toward building militaries that can take on the United States and its allies over key strategic interests. Gone are the days when the United States could easily swat away a Chinese attack on Taiwan or when it did not even have to contemplate a Russian assault on the Baltics.

The problem is not just that Russia’s and China’s increasingly sophisticated and powerful conventional militaries are well poised to strike U.S. allies and partners (think Poland or the Baltics in Europe and Japan or Taiwan in Asia). It is also that any future confrontation with Russia or China could go nuclear. First, in a harder-fought, more uncertain struggle, each combatant may be tempted to reach for the nuclear saber to up the ante and test the other side’s resolve, or even just to keep fighting. Second, should Moscow seize the Baltics or Beijing invade Taiwan, both U.S. foes are likely to threaten to use or actually use nuclear weapons to close the door on U.S. counterattacks, or to drastically curtail their effectiveness. In fact, this forms a central pillar of their theories of victory—the potential playbooks they could use to take on the United States and come out the better for it.

If downsizing the U.S. nuclear arsenal once made sense, it no longer does.

This threat is not a figment of the imagination. Russia has spent much of its limited money building a modern and varied nuclear weapons arsenal. Much of this arsenal is designed to attack specific military targets rather than to wipe out major cities in one fell swoop. For instance, Russia fields a substantial number of naval nuclear weapons, including antiship cruise missiles, nuclear torpedoes, and nuclear depth charges. As Russian exercises and military journals suggest, the idea behind Moscow’s nuclear strategy is to use tailored nuclear weapons to settle a war on Russia’s terms, gambling that going nuclear will intimidate the United States into backing down—a strategy known as “escalate to de-escalate.” 

If Russia wished to challenge NATO, it could deploy “little green men”—soldiers or intelligence officers in disguise or unmarked uniforms—to Poland or the Baltics in an attempt to sow confusion and shape opinion in Moscow’s favor, as it did in Crimea in 2014. It could then send in lethal conventional forces, which could rapidly seize ground, dig in, and set up a formidable defensive position. Threatened or real nuclear attacks designed to knock back any conventional counterattack that U.S. and NATO forces might launch in defense of their allies would seal the deal. Moscow could, for example, hit key U.S. bases in western Europe or U.S. flotillas in the Atlantic. Washington would be left with a simple choice: a settlement or a major nuclear war. 

China has been more restrained than Russia in its nuclear buildup, but it is also developing modern, nuclear-capable forces that could be used in a regional conflict, such as the DF-21 and DF-26 ballistic missiles. These are just the type of weapons China would need to checkmate the United States in Asia. In the event that it wanted to force the Taiwan question or dictate the terms of a settlement of territorial disputes with Japan, Beijing could rely on its newfound wealth and power to politically isolate one of these states. If the situation escalated, China’s conventional forces could try to seize Taiwan or the disputed territories and prepare to block an effective response from U.S. and allied troops. If this didn’t prove enough, China’s increasingly accurate and flexible nuclear forces could hit U.S. air and naval bases in the western Pacific, testing how far the United States would be willing to go in defense of its allies and partners. The bottom line is that if the United States wants to sustain its alliance architecture in Europe and Asia, it must adapt its strategy to face an opponent prepared to escalate with nuclear weapons.

A missile is test-launched from a U.S. Navy ballistic missile submarine off the coast of California, March 2018.
U.S. Navy/Ronald Gutridge/REUTERS


Above all, this requires jettisoning the outdated assumptions that continue to shape current debates on U.S. nuclear strategy. On one side are the doves, who argue that nuclear war simply cannot be limited or controlled and that the specter of nuclear devastation is enough to deter a major war. The key, as they see it, is to make sure that no one thinks otherwise and to avoid rocking the boat lest things get out of hand. In the meantime, all the United States needs to deter Russia or China is a relatively small arsenal of nuclear weapons with little purpose other than to destroy highly valued but unprotected targets such as cities. This threat is enough, the argument goes, pro­vided that all parties maintain powerful but carefully constrained conventional forces and avoid unnecessary skirmishes. 

This line of reasoning has influential supporters. In 2012, a study group chaired by James Cartwright, the former vice chairman of the Joint Chiefs of Staff, concluded that “there is no conceivable situation in the contemporary world” in which a nuclear attack would be in the United States’ or Russia’s interest. The group’s report urged the United States to reduce its nuclear arsenal substantially and eliminate its tactical nuclear weapons altogether. In the same vein, a letter signed this year by former Secretary of Defense William Perry and other heavy­weights contended, “It is unlikely that there is such a thing as a limited nuclear war; preparing for one is folly.” 

Unfortunately, this view ignores the incentives that U.S. foes would face in a war and the evidence about how they would likely behave. Russia and, to a lesser extent, China field increasingly accurate, lower-yield nuclear weapons that would add little in an all-out nuclear conflagration but would be useful in a limited nuclear exchange. It appears that they believe that limited nuclear escalation is possible—and that it may even represent their winning move against the United States. 

This shouldn’t come as a surprise to Washington. The risks of nuclear brink­manship may be enormous, but so is the payoff from gaining a nuclear advantage over an opponent. Nuclear weapons are, after all, the ultimate trump card: if you can convince your enemy that you have a way to play the card and are actually prepared to go through with it, nothing is more powerful. And the best way to do that is to have palatable options for the limited and effective use of nuclear weapons. Americans should know: they perfected this approach against the Soviet Union during the Cold War. The doves’ strategy, however, would leave the United States without any means to do this, encouraging adversaries to exploit this gap and making war—including nuclear war—more likely. 

Yet some super-hawk thinking would also lead U.S. policymakers astray. For many hawks, the solution is for the United States to develop forces of all kinds able to hobble Russia’s or China’s nuclear arsenal, while setting up massive missile defenses to block any retaliation. If the United States perfected this approach, it could carry out a disarming first strike against an adversary. The long shadow of this threat alone would discourage Russia or China from mounting an attack on U.S. friends or allies. 

The problem with this approach is that it is simply too difficult to pull off and is therefore an obvious bluff. Destroying or blocking all Russian or Chinese nuclear forces would be a mind-boggling challenge. And in a nuclear war, you have to be perfect or just shy of it: allowing even a handful of thermonuclear weapons through U.S. defenses would mean staggering death and destruction. This human cost would be completely out of proportion with whatever interests prompted the United States to engage.

In order to fully disarm Russia or China, the United States would have to not only destroy or disable large numbers of widely dispersed mobile missile launchers, submarines, and aircraft but also do so concurrently, at most within hours, to prevent a counterstrike. This would involve finding and fixing mobile targets, tracking them if they moved, destroying them, and confirming their demise—a task the United States has found extremely difficult even against much weaker opponents, such as Iraq. 

Meanwhile, U.S. defenses would have to keep any enemy missiles from reaching their targets—yet U.S. missile defenses have struggled against primitive ballistic and cruise missiles, let alone advanced Russian and Chinese projectiles. This is a function of the inherent difficulty of defending against incoming ballistic missiles traveling at several times the speed of sound, not to mention dealing with stealthy cruise missiles and underwater torpedoes. As James Winnefeld, then the vice chairman of the Joint Chiefs of Staff, put it in 2015, “Missile defense against these high-end threats is too hard and too expensive and too strategically destabilizing to even try.” Put simply, there is no plausible scenario in which the super-hawk approach makes sense. And patent bluffing is not a wise long-term strategy. 


Ultimately, the logic of deterrence dictates that the United States’ defense strategy for its new great-power rivals must balance two competing demands: whatever actions Washington threatens must be potent enough to coerce the opponent but not so apocalyptic as to be implausible. For the United States, striking this balance is not easy. A country trying to defend its home territory may be able to convince opponents that it will risk nuclear annihilation to avoid foreign occupation. But for Washington, which is trying to help defend far-flung allies against foreign aggression, such threats are far less credible. As one U.S. official quoted former Secretary of State Henry Kissinger as saying, “Great powers don’t commit suicide for their allies.”

The good news is that the United States can protect its allies without going after its opponents’ entire nuclear arsenals or marching on Moscow or Beijing. Instead, American forces must be able to blunt any invasion of allied territory by quickly attacking the conventional and tactical nuclear forces that Russia or China would use to seize and hold on to that territory. Once the United States had successfully done so, Russia or China might decide to end the conflict there—an outcome that Washington could accept. If they decided, however, to press on even after U.S. forces had warded off an initial offensive, the burden of escalation would rest squarely on their shoulders. 

Consider the case of China: instead of being able to quickly seize Taiwan and create facts on the ground, Chinese leaders would face a choice between backing down and risking a major, prolonged war with the United States—not to mention U.S. allies galvanized into action by large-scale Chinese aggression in East Asia. Once the path to a quick invasion of territory was blocked, any escalatory actions that China might turn to would, in effect, be self-defeating, as they would set off a unified response by the United States and its allies.

Washington will have to invest in modern tactical nuclear warheads and delivery systems designed for a regional military fight.

U.S. conventional forces would still do most of the work of blocking the adversary’s advance by delaying, degrading, and ideally halting any invading forces. Accordingly, preparing combat-ready conventional forces to fight alongside allied militaries must be a central pillar of U.S. strategy. But American nuclear forces, especially those designed for a limited war, would have an equally important role to play. For one, Russia or China might decide to escalate to the nuclear level, forcing the United States to respond in kind or risk defeat. Moreover, if the United States’ conventional edge further erodes in the coming decades, particularly in East Asia, it may have to rely on its nuclear forces to halt Chinese conventional forces.

To be able to pull of such a strategy, Washington will have to invest in modern tactical nuclear warheads and delivery systems designed for a regional military fight. As it exists today, the U.S. arsenal consists mostly of strategic weapons, built for waging a large-scale nuclear war against an enemy’s strategic forces, leadership targets, and the like. Almost all U.S. tactical nuclear weapons have been dismantled. The few that remain are of only limited use in a war against Russia or China.

The Pentagon’s 2018 Nuclear Posture Review recognized this gap. It committed to modernizing its air-delivered tactical bombs and developing low-yield nuclear warheads for submarine-launched ballistic missiles. But the United States should go further and specifically develop or adapt a modest number of nuclear weapons and delivery systems that could damage key Russian or Chinese conventional targets, especially those needed for an invasion of the Baltics or Taiwan: entrenched ground forces, maneuver troops, naval flotillas, and invasion fleets. The new weapons would need lower yields than most of those in the current arsenal, which have been optimized to destroy hardened silos sheltering enemy missiles, not to stop conventional forces.

These weapons would not replace U.S. conventional forces. They would, however, help offset any advantages that Russia and China derive from their own nuclear arsenals. Risking a confrontation with a similarly well-equipped United States would mean courting defeat or near-suicidal escalation.


Because there is no effective deterrence without effective communication, Washington also has to change the way it talks about its nuclear strategy. In recent decades, the U.S. government has tended to stress that nuclear war is uncontrollable. There is obviously great merit to this point, since crossing the nuclear threshold would indeed be tremendously perilous. But fixating too much on the uncontrollability of nuclear war actually invites escalation. Opponents may quite reasonably conclude that Washington is so convinced that any limited nuclear operations will escalate to Armageddon that it would never dare cross the threshold except for its own survival—which would leave U.S. allies out in the cold. 

Accordingly, U.S. officials need to change their line. They should continue to stress that a nuclear war could quickly spin out of control, with calamitous effects. Yet they should also demonstrate—by deed, in the exercises the military holds, the training it undertakes, and the capabilities it develops, and by word, in the official statements Washington issues—that the United States is prepared to conduct limited, effective nuclear operations. This would signal to Russia and China that the United States has the will and the way to frustrate any nuclear brinkmanship.

Such a nuclear strategy is compatible with arms control. After all, the goal of arms control is not disarmament but strategic stability. In practice, this means ensuring that all sides have confidence in their own ability to launch an effective retaliatory nuclear strike, while leaving ample room for cooperative steps to reduce the risk that an accident or a miscalculation could lead to war. 

For decades, the dominant thinking in U.S. nuclear policy has been to reduce, minimize, and eliminate. This approach may have been defensible in the 1990s and the early years of this century—but the world has changed. The United States now faces great-power competitors that believe they could successfully take on the United States, hoping to exploit Washington’s fear of the nuclear precipice. Disabusing them of any such notion is the best deterrent against such a scenario. Perhaps paradoxically, then, the best way to avoid a nuclear war is to be ready to fight a limited one. 

To critics, this approach will smack of Cold War thinking. But when it comes to defense strategy, that may not be a bad thing. After all, Cold War thinking enabled the United States and its allies to deter major aggression for 45 years, even though their conventional forces in Europe were consistently out­numbered. The United States should consider itself lucky if it achieves such a result over the next half century. A certain kind of Cold War thinking may be just what Washington and its allies need.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now
  • ELBRIDGE COLBY is Director of the Defense Program at the Center for a New American Security. He served as U.S. Deputy Assistant Secretary of Defense for Strategy and Force Development in 2017–18.
  • More By Elbridge Colby