Table of Contents

Given the enormous risks of nuclear war, nuclear weapons should be reserved for deterring threats of a scale and type that cannot be deterred or defeated by other means. Since 1945, no threat—including many crises, conflicts, and wars—has caused the United States or any other country or nonstate entity to initiate use of nuclear weapons. The 2018 NPR suggests that the decline in wartime casualties since 1945 is due primarily to the deterrence provided by nuclear weapons.1 There may be truth to this. But the invocation of World War II and the emergence of nuclear deterrence in 1945 invites at least two other pertinent observations.

First, unlike when the United States used nuclear weapons against Japan in 1945, today the adversary also will possess nuclear weapons in any conflict where the United States conceivably would use them.2 In other words, the risks entailed in the United States’ use of nuclear weapons, in either a first or second strike, are greater. These risks may induce greater caution among all parties, including the United States, but they add to the challenge of protecting U.S. allies and partners in potential regional scenarios where Russia, China, or North Korea might threaten them.

Consequently, because nuclear war would carry a clear risk of catastrophic destruction for the United States and probably for its allies or partners, American leaders should only contemplate use when the violence and destructiveness of the aggression that must be defeated is of a similar scale. States with nuclear weapons have been attacked and have lost wars, as the United States and Russia did respectively in Vietnam and Afghanistan.3 Nuclear weapons have not enabled states to compel adversaries to stop supporting terrorism, reverse illegal territorial aggrandizement, respect human rights, or desist from cyber attacks. Curbing these activities is a clear national security priority for the United States, but the U.S. nuclear arsenal cannot reasonably be expected to deter these threatening activities or to compel the countries engaged in them to stop. In short, nuclear weapons cannot solve most security problems.

Russia, China, and North Korea pose the few threats that U.S. nuclear weapons are necessary to deter or, if that fails, possibly help defeat. These threats all have histories and reflect action-reaction dynamics between adversaries, including the United States. The drivers and evolution of these threats look different depending how far back in time one rewinds history and whose perspective is taken to assess what happened before. In any case, the prospective Russian, Chinese, and North Korean threats vary. Sound and credible U.S. deterrence strategies require individual consideration of all three countries’ unique military capabilities and national goals—which are of course debatable. Table 2 presents estimates of the nuclear forces of the United States and its three immediate nuclear concerns.

Table 2: Estimates of U.S. and Potential Adversaries’ Nuclear Forces
  United States Russia China DPRK
Deployed strategic warheads 1,750 1,572 0 0
Nondeployed warheads, strategic and nonstrategic 2,050 2,740 ~low 200s 35
Total warhead stockpile 3,800 4,312 ~low 200s 35
ICBMs 400 302 81 N/A
SLBMs 240 160 48 0
Long-range bombers 66 68 20 0
Nonstrategic nuclear warheads* 230 1,870 87 0
Dual-use theater-range missiles 0 90 ~108 <150**
*Warheads associated with short- and intermediate-range delivery systems but not necessarily deployed day-to-day in peacetime.

** Public estimates vary; this total is based on the approximate number of missile launchers.

Note: “Strategic” refers to warheads on non-forward-based weapons with ranges that are capable of reaching adversaries’ homelands. New START defines their range as greater than 5,500 kilometers.

Source: Data drawn from Hans M. Kristensen and Matt Korda, “United States Nuclear Forces, 2020,” Bulletin of the Atomic Scientists 76, no. 1 (2020), 47; Hans M. Kristensen and Matt Korda, “Russian Nuclear Forces, 2020,” Bulletin of the Atomic Scientists 76, no. 2 (2020): 103–4; Hans M. Kristensen and Matt Korda, “Chinese Nuclear Forces, 2019,” Bulletin of the Atomic Scientists 75, no. 4 (2019), 172; Hans M. Kristensen and Robert Norris, “North Korean Nuclear Capabilities, 2018,” Bulletin of the Atomic Scientists 74, no. 1 (2018), 42; Hans M. Kristensen and Matt Korda, “Status of World Nuclear Forces,” Federation of American Scientists, September 2020,; and Office of the Secretary of Defense, Military and Security Developments Involving the People’s Republic of China, 2020, Annual Report to Congress (Washington, DC: Department of Defense, 2020).


As was the case in the Cold War, Russia drives most U.S. nuclear requirements. Russia’s nuclear capabilities have set the upper bound of threats that U.S. nuclear policy must deter or defeat. If Russia did not have nuclear weapons but all other nuclear-armed states maintained their current levels, U.S. nuclear requirements would be substantially reduced. Alternately, if Russian nuclear forces were comparable in size to China’s, the United States would not need the postures called for in the 2010 and 2018 NPRs. This would be so even if China doubled its stockpile to approximately 500 warheads—as the Department of Defense warns could be the case in ten years.4

Russia’s competitive strategy is to weaken its adversaries through the lowest level of violence necessary (or preferably no violence at all). Among other reasons, Russia seeks to avoid the mobilization of the West’s superior economic and military potential.5 Actors with various affiliations to the Russian state have engaged in political and economic interference and nonkinetic military actions at various levels of conflict. At the same time, Russian leaders welcome the deterring psychological shadow that nuclear weapons cast over any potential conflict with Russia.6

The U.S. nuclear deterrent is relevant only for those threats that involve large-scale armed conflict, as a war between Russia and the North Atlantic Treaty Organization (NATO) would. U.S. nuclear policy will not plausibly compel Russian leaders to curtail coercive or subversive actions against the United States and its allies below any reasonable threshold of armed conflict.7 Scenarios for analyzing nuclear escalation generally posit Russian forces or proxies advancing into a NATO state and/or taking coercive control over it before NATO mustered the resolve and capacity to stop them and compel them to exit.8 A recent RAND Corporation simulation of a NATO-Russian conflict in Estonia highlights NATO’s vulnerability to such a scenario. The report concludes that “NATO lacks the conventional forces required to slow or stop the rapid Russian advance. NSNW [non-strategic nuclear weapons] alone cannot substitute for NATO’s lack of those conventional forces. . . . This problem will not be solved by new means of basing or delivery of low-yield nuclear weapons alone.”9

Russia is modernizing its nuclear forces in order to disabuse the United States that it could initiate limited nuclear operations against Russian forces in Europe and then, through offensive strikes and/or missile defense interceptions, deny Russia the capability to escalate against the U.S. homeland. The Avangard hypersonic glide vehicle (HGV); the Sarmat heavy, MIRV (multiple independent reentry vehicle)-capable silo-based ICBM (intercontinental ballistic missile), which can carry up to ten warheads (or more); the Poseidon autonomous underwater vehicle; the Burevestnik nuclear-powered cruise missile; and the Kinzhal air-launched ballistic missile (ALBM) are all intended to survive U.S. offensive and defensive attacks and ensure an ability to deliver nuclear weapons against the U.S. homeland.

Russian leaders mirror U.S. strategists in thinking about the role of nuclear weapons in deterring or defeating threats.10 According to a 2020 study by the Center for Naval Analyses, Russian leaders “view nuclear use as defensive, forced by exigent circumstances, and in the context of regional or large-scale conflicts.”11 In this study, Michael Kofman and Anya Loukianova Fink state

the purpose of Russia’s escalation management strategy is to deter direct aggression, preclude a conflict from expanding, prevent or preempt the use of highly damaging capabilities against the Russian homeland that could threaten the state or the regime, and terminate hostilities on terms acceptable to Moscow. . . . Only strategic deterrence forces, armed with conventional capabilities (offensive strike and aerospace defense), nonstrategic nuclear weapons, and strategic nuclear weapons, are effective deterrents in regional and large-scale wars.12

Similar to the ways in which an emerging generation of nuclear strategists view U.S. nuclear policy today, Kofman and Fink write,

the Russian military does not believe that limited nuclear use necessarily leads to uncontrolled escalation. The Russian military believes that calibrated use of conventional and nuclear capability is not only possible but may have decisive deterrent effects. This is not an enthusiastically embraced strategy, but an establishment’s answers to wicked problems, in the context of a great-power conflict, which have no easy or ideal solutions.13

In confronting these wicked dilemmas, Russian policymakers echo their U.S. counterparts by saying that if they used nuclear weapons first, it would be to prevent further escalation of the conflict (Russia) or restore deterrence at the lowest level of damage possible (the United States).14 Yet both Russia and NATO seek this outcome on terms favorable to them, not to the adversary, and therein lies the risk of escalation. If nuclear escalation did occur with attacks on their strategic forces, both maintain the option to launch a retaliatory strike with alert nuclear forces before an enemy’s nuclear weapons arrive. Both countries also threaten nuclear use to deter an adversary’s attack—perhaps by cyber means—against critical nuclear command and control infrastructure.15 In all of this, both countries acquire and make plans to use incomparably extensive and destructive nuclear and dual-use arsenals and, for the United States, perhaps missile defenses in order to prevent the other from dominating the imagined escalatory process.

More than NPRs traditionally acknowledge, the United States and its allies need to understand how to reassure Russia that NATO poses no offensive threat to Russian interests while simultaneously projecting sufficient capabilities and political resolve to deter Russian armed aggression. Such reassurance can foster stability—an overarching goal of U.S. and NATO policy. In doing this, the United States and its allies must demonstrate that Russian reductions of coercive rhetoric, actions, and forces will beget reductions in NATO rhetoric, action, and forces, that Russia reasonably could find threatening.16 The United States and its allies also must seek greater clarity on Russia’s nuclear doctrine to better understand how Russian political and military officials think operationally about first use—under what circumstances, against what targets. This awareness can help ascertain how NATO could deter or dissuade Russia from undertaking such escalation. It is cliché, but nonetheless true, that the political cohesion and therefore resolve of NATO is vital in all of this.

The greatest threat to the NATO alliance would be if the strength of its conventional forces, cyber defenses, military resilience, and political cohesion declined relative to Russia.

For deterrence, the greatest threat to the NATO alliance would be if the strength of its conventional forces, cyber defenses, military resilience, and political cohesion declined relative to Russia. Western policymakers also must anticipate the possibility that deterrence could fail, and in that case deploy capabilities and plans best suited to end the postulated Russian aggression at costs acceptable to NATO populations and governments. The abovementioned RAND report does not suggest there are any acceptable “nuclear solutions” to this problem—not greater numbers, different yields, alternative basing, or different targeting.17 However, this does not mean that changes in strategic nuclear capabilities and/or arms control arrangements would not be advisable to counter new Russian forces and deter escalation of conflict in Europe. Chapters 4 and 6, respectively, present recommendations for strategic modernization and new U.S.-Russia arms control arrangements.


China poses numerous challenges to the United States and its allies. Many of them are economic and diplomatic and, therefore not elaborated here. In terms of military activity, the U.S. Defense Department asserts “China calibrates its coercive activities to fall below the threshold of provoking armed conflict with the United States, its allies and partners, or others in the Indo-Pacific region.”18 The United States and its allies and partners would like the Chinese state to stop maltreating Uighurs and dissidents, stop using cyber tools to steal intellectual property, and refrain from taking disputed territories (such as islets) and marine resources. U.S. nuclear weapons cannot reasonably be used to accomplish these goals.

Coupled with a long-term effort to diversify and increase the survivability of its nuclear arsenal, China is pursuing kinetic and nonkinetic capabilities to frustrate U.S. efforts to come to the aid of its allies and Taiwan in a regional crisis.19 The key challenge for U.S. and friendly defense policymakers is to counter these capabilities and related Chinese intentions to prevail in conflicts around its periphery and to exert hegemonic power in Asia.

The integrity of Taiwan is probably the gravest concern. Chinese President Xi Jinping has staked his government’s credibility on doing whatever is necessary to prevent Taiwan from declaring and implementing independence.20 The military challenge of conventionally defending Taiwan from a concerted Chinese attack is becoming more difficult over time.21 The second and perhaps more imminent challenge derives from territorial disputes between China and several U.S. allies—Japan most prominently—over maritime claims. To date, China has been careful to keep its exertions below the level of armed conflict. However, these disputes could stimulate purposeful or accidental military confrontations that could then escalate and involve a wider number of U.S. allies and partners, including Australia, members of the Association of Southeast Asian Nations, and India. 22

U.S. nuclear capabilities and policies contribute to deterring China from initiating or escalating large-scale military conflict with U.S. allies and partners and the forces that would defend them. The complex nature of U.S. alliances in Asia and the varied capabilities of allies and partners to defend themselves complicate U.S. deterrence policy toward China. The lack of a NATO-like decisionmaking structure makes it more difficult to ensure effective political and military coordination among all governments, and ultimately to assure U.S. allies that the United States will not involve them in a war of its own making, and vice versa.

China has traditionally been restrained in its deployment of nuclear forces and in its no-first-use (NFU) doctrine.23 According to former Defense Intelligence Agency (DIA) director Lt. Gen. Robert Ashley, China possesses a nuclear warhead stockpile in the “low couple hundreds.”24 Approximately 180 of these would be paired with intercontinental-range (or “strategic”) delivery systems, and roughly 100 would be paired with short- to intermediate-range weapons capable of hitting regional targets.25 The Department of Defense reports that China’s nuclear warhead stockpile is projected to “at least double in size” in the next ten years.26 (By comparison, Russia is estimated to have 6,490 active and inactive warheads in its stockpile, of which 1,600 are deployed on strategic delivery vehicles, and the United States has 5,800, of which 1,750 are deployed on strategic delivery vehicles.27) See figure 1 for a comparative inventory of U.S., Russian, and Chinese strategic nuclear warheads.

The Department of Defense states that China “almost certainly” does not integrate warheads with delivery systems in peacetime, but believes that China may take steps to adopt a “high alert posture conceptually comparable to the claimed high alert posture kept by portions of U.S. and Russian nuclear forces.”28 However, China’s posture today is consistent with Beijing’s declared intentions not to initiate nuclear war or counterforce operations against the United States.29

In recent decades, China has sought to increase the survivability and effectiveness of its nuclear forces. It has fielded an array of mobile land-based missiles that can range regional and intercontinental targets, and has developed a new generation of SSBNs (ballistic missile submarines) and SLBMs (submarine-launched ballistic missiles). China has also built extensive underground facilities to shelter its mobile land-based nuclear forces from a potential U.S. attack. It is developing and testing new maneuverable hypersonic delivery systems to overcome U.S. missile defenses. Chinese experts explain that the enhancement of China’s nuclear forces—largely to make them more resilient—are driven by advances in U.S. and other states’ capabilities to threaten China’s nuclear deterrent with new non-nuclear kinetic weapons, cyber operations, and ballistic missile defenses.30

U.S. military leaders openly doubt that China’s policies have been or will be as restrained as Beijing claims. In reference to China’s NFU policy, the head of U.S. Strategic Command (USSTRATCOM) remarked before the Senate that he “could drive a truck through that no-first-use policy.”31 The question of NFU in relation to China continues to present a paradox. Some U.S. commentators say that U.S. and allied security would be diminished by adopting NFU, and that the United States should continue to seek a combination of offensive strike capabilities and missile defenses to threaten China’s second-strike deterrent.32 Yet these same U.S. commentators then denounce any suggestion from China’s own experts that their country might be better served by hedging NFU in order to deter a potential U.S. non-nuclear first strike against China’s nuclear forces.

The absence of meaningful sustained dialogue between senior U.S. and Chinese defense officials and military officers regarding issues related to nuclear and missile defense policies impairs both countries’ capacity to avoid conflict or to prevent any potential conflict from escalating. Chinese leaders—especially the top officials of the PLA (People’s Liberation Army) Rocket Force—traditionally have been averse to engaging transparently in such dialogue, even though they may be able to represent Beijing’s genuine strategic perspectives.

China is pursuing nuclear, kinetic, and non-kinetic capabilities to frustrate U.S. efforts to aid allies and Taiwan.

China’s nuclear force modernization could improve stability by reassuring China’s leaders that their second-strike nuclear forces will survive. But some warn that China will seek to become a nuclear peer of the United States and then more assertively project its power in Asia and beyond.33 Here the most practical worry is that China could forcibly take disputed territory or inflict other harm on a U.S. ally or partner and then use explicit or unstated nuclear threats to deter the United States and its ally or partner from using their full capabilities to contest China’s action. Combined with China’s political, economic, and other military influence augmented by increasingly capable conventional forces, the threat of nuclear escalation can create coercive pressure on U.S.-allied relationships, particularly in the case of a crisis involving Taiwan.34 The key policy question for U.S. and allied policymakers is which additional nuclear or non-nuclear capabilities would most effectively deter and, if necessary, counter Chinese escalation of regional conflict.

North Korea

The DPRK poses the threats that are most likely to lead U.S. decisionmakers to first use of nuclear weapons. North Korea could rapidly inflict extensive nuclear and non-nuclear damage on Seoul and other regional targets. U.S. and Republic of Korea conventional weapons alone may not be sufficient to destroy hardened North Korean targets, including some of its nuclear forces. Moreover, U.S. and allied missile defenses could interdict at least some of the relatively small number of nuclear weapons with which North Korea could retaliate after a U.S. first strike. If North Korea were inflicting heavy damage on Seoul with conventional artillery and missile barrages, and intelligence indicated that the North was readying its nuclear weapons for use, the situation could well stimulate the most urgent-ever U.S. leadership considerations of launching nuclear weapons.

U.S. military leaders are confident that they and their allies ultimately can defeat North Korea without recourse to nuclear weapons. The presence of U.S. Forces Korea, other U.S. forces in the region, and the military strength of South Korea provide a conventional force posture that credibly deters a North Korean attack on South Korea. If North Korean leaders are not rational actors there is no reason for thinking that greater U.S. and South Korean forces would deter them more effectively.

South Korean military capabilities and operational thinking have advanced greatly in the past several years. They have focused on modernizing conventional strike and missile defenses to destroy and defend against the North Korean missile threat.35 However, some U.S. analysts argue that preemptive U.S. nuclear strikes on the DPRK could significantly reduce the level of damage that a North Korean conventional attack might inflict on South Korea and local U.S. forces.36 Theater and national missile defense theoretically could blunt any nuclear retaliation from the North. Yet in considering such an operation, and the policies required to enable it, other factors would need to be considered. How would fallout from potential U.S. nuclear strikes affect South Korea, Japan, and other nations in the region? What are the probabilities that regional and national U.S. missile defenses would blunt attacks by surviving North Korean nuclear weapons? Would capabilities to preemptively strike North Korean nuclear forces stimulate its leadership to deploy more nuclear weapons and adopt risky policies to enhance their survivability?

It is important to recognize that additional capabilities the United States might pursue to locate and destroy mobile North Korean nuclear forces could exacerbate crisis and arms race instabilities with China and perhaps Russia. Forward-deploying U.S. nuclear weapons in South Korea could ameliorate some problems but would create or exacerbate others. The U.S. arsenal deployed for the primary mission of deterring Russia and China is more than adequate if the DPRK is deterrable. Again, if North Korean leaders are not rational or deterrable, then there is no way to assess whether and how various U.S. nuclear postures would make a difference and thus little foundation from which to suggest changes to current U.S. nuclear posture.

Unforeseen Threats

National security strategists, including authors of past NPRs, naturally caution that new adversaries and/or threatening capabilities will emerge to threaten the United States. Some analysts and officials then immediately conclude that nuclear weapons—of newer types or in greater numbers—will be needed to deter or defeat these potential new threats.37 The designers and producers of nuclear weapons systems reinforce these impressions for obvious material reasons; so, too, do the congressional delegations from districts where such weapons systems are developed, built, and deployed. This line of thinking, complete with special interests, reflects the traditional U.S. approach to nuclear posture and policy and general planning for future military requirements. Yet Washington must take more care to examine second- and third-order ways in which developing new nuclear weapons would affect competitions with adversaries and other security and diplomatic objectives. Here, the U.S. defense establishment reinforces a similar predilection in Russia. (Judging by the much smaller sizes and varieties of their arsenals, other nuclear-armed states do not fall into similar predilections.)

Prudence does require the United States to maintain an unsurpassed research and development base to be able to anticipate and respond to emerging threats, including through the maintenance and development of nuclear weapons. It must also maintain advanced intelligence gathering and reconnaissance capabilities to detect threats that could require nuclear or other responses. Confidence in detecting and countering emergent threats would be all the greater if arms control and other international security regimes were in place to foster transparency, aid in the early recognition of hostile intent, limit military forces designed for rapid attacks against targets of national importance, establish norms against weapons of mass destruction, and so on.

The combination of research and development, state-of-the-art intelligence collection and analysis, and durable arms control measures offers the most cost-effective and stabilizing way to manage risks of unforeseen existential threats (and perhaps prevent those threats from materializing). Given that the United States will retain a nuclear arsenal scaled to that of its largest nuclear competitor, it is difficult to see circumstances in which additional nuclear capabilities should be retained or developed as a hedge against invisible future threats.

The Central Challenge: Preventing Escalatory Regional Conflict

Preventing and managing escalation is central to all the threats or challenges just discussed. Escalation can occur inadvertently or purposefully. It can transform conventional war into nuclear war, and limited nuclear war into all-out nuclear war.

Inadvertent escalation is not a new problem in the Atomic Age. It was analyzed comprehensively in the 1980s.38 However, the risks may have grown significantly in recent decades as nuclear-armed states have developed and deployed advanced non-nuclear weapon technologies that could target both nuclear and conventional forces and their associated command and control systems.

U.S. nuclear forces and command, control, and communication systems (NC3) capabilities increasingly are entangled with those used to manage conventional military operations. Components include space- and ground-based early warning and reconnaissance systems that Russia or China would be interested in disabling in a conventional or a nuclear conflict. Similarly, Russian and Chinese missiles and their command and control systems may employ either conventional or nuclear warheads. In some cases, nuclear-armed and conventionally armed missiles are co-located. These three countries also assume that adversaries are increasing cyber capabilities to target their command and control systems.

U.S. nuclear forces and command, control, and communication systems capabilities increasingly are entangled with those used to manage conventional military operations.

Early in a conventional conflict, Russia or China could use various means to penetrate and/or attack U.S. command and control systems. Cyber penetrations, for example, could gather intelligence on or simply attack these systems. U.S. military operators and political leaders would naturally be inclined to think the worst and assess that Russia is preparing nuclear attacks on U.S. nuclear capabilities.39 Similarly, the United States, early in a conflict, would plan to conduct cyber and conventional attacks to disable adversary conventional military command-and-control systems. To the extent that such U.S. attacks were directed against Russian, Chinese, or North Korean assets with both conventional and nuclear functions, these adversaries could perceive them to be the early stages of U.S. nuclear attack, even if the United States in fact was not planning to escalate to nuclear use. Russian and Chinese analysts have expressed deep insecurity about such scenarios.40 The United States’ resistance to NFU and its long-standing plans to target these adversaries’ nuclear forces could reinforce adversary perceptions of imminent nuclear attack. Adversaries, in turn, could escalate tensions by targeting U.S. NC3 assets in space in a bid to disrupt such attacks, or issue potentially destabilizing orders to raise the alert level of their nuclear forces. The circular dynamic of perception and possible misperception here adds to the overall risk of inadvertent escalation.

Even before attacks, activities by American, Russian, or Chinese units known to operate both conventionally and nuclear-armed missiles could be perceived as preparations for nuclear attacks. After known dual-capable missiles were fired, the targeted state may be unable to determine whether the released weapons are carrying nuclear or conventional warheads. Leaders of states facing such attacks would thus have to choose to either launch their own nuclear weapons in the possibly incorrect belief that the adversary is initiating or escalating nuclear war, or to hold their fire even in the face of an actual nuclear attack. The first would risk creating nuclear war from a situation where neither side intended it. The second would run the risk of being wrong and then hobbled in an escalatory contest. The second option is obviously the saner one, especially for large states like the United States, Russia, and China, but this situation would be unprecedented.

In an ideal situation, this cautious thinking could strengthen deterrence. If military and civilian leaders of the United States, Russia, and China understand the risks of inadvertent escalation, they could be more inclined to dampen crises, avoid initiation of armed conflict, or deescalate before nuclear weapons are unleashed. (It is more difficult to assess North Korean intentions and thinking in this regard.)41

New nuclear warheads and delivery systems, however, cannot unilaterally solve the risks of inadvertent escalation. The relevant actors across multiple agencies and departments in the United States, Russia, and China must first fully understand the problem and then create venues for bilateral or trilateral dialogue on it. Such dialogue—if more sustained and detailed than has been the case heretofore between the United States and Russia and, especially, the United States and China—could help clarify whether and how alternative force postures, procurement decisions, and confidence-building measures could mitigate dangers of inadvertent escalation.

Purposeful escalation is most likely to occur when a state is failing to achieve its objectives in conventional war. It could then conclude that it must employ nuclear weapons to compel the adversary to reverse course. The same escalatory dynamic can occur after nuclear weapons have been employed when one or both (or more) adversaries decide to increase nuclear attacks to compel the other(s) to desist from further warfare. For instance, if Russia or North Korea were losing a conventional conflict with the United States and its allies, leaders in Moscow or Pyongyang would at least contemplate employing nuclear weapons to reverse the U.S. advantage or at least to deter the United States from pressing on to inflict greater loss. If Russia or North Korea did unleash nuclear weapons on U.S. forces, allies, or the U.S. homeland, U.S. policymakers would face the excruciating judgment of whether nuclear reprisals would be necessary and likely to reverse the escalatory dynamic.

Such scenarios with China are more difficult to predict insofar as Beijing insists it would not be the first to use nuclear weapons and has not to date deployed nuclear forces well suited for engaging the United States in limited nuclear war. Indeed, Beijing has strengthened a suite of kinetic and nonkinetic capabilities to deny the United States and its allies and partners from being able to defeat it in a regional conventional conflict. In any case, to avoid situations of nuclear escalation, realistic and prudent nuclear policy requires serious planning for diplomatic signaling and non-nuclear military options to pursue war termination.42 If nuclear weapons are exchanged even in limited numbers, none of the belligerents would be likely to accomplish its favored outcome. For belligerents as well as the rest of the world, preventing further nuclear escalation would be better than the alternative.43

The more difficult scenario would arise if Russia or China gained significant territorial or other advantages early in a conventional conflict with the United States and its allies. If these potential adversaries managed to negate U.S. and allied conventional capabilities to reverse such losses, the United States would then consider nuclear first use to compel them to stop the fighting (and more quixotically return to the status quo ante). Indeed, some theorists regard the threat of U.S. first use in these scenarios as a critical component of deterrence.44 This is why deploying capabilities and operational plans to conduct limited nuclear war has become central to U.S. (and Russian) policymakers in recent years, even if grave doubts remain that nuclear escalation would be controllable.

Reflecting the salience of regional war scenarios, a recent State Department analysis notes the relative unlikelihood that Russian, Chinese, or North Korean leaders would be able and willing to conduct “bolt-out-of-the-blue” nuclear strikes on the United States or its allies. Political conditions, U.S. intelligence and warning capabilities, and surviving U.S. nuclear and other forces practically guarantee that such strikes would result in devastating U.S. countermeasures of the exact sort all three countries seek to avoid.45 The greater risk, which warrants the most attention by nuclear experts, is escalation from a regional conventional war. To prevent such scenarios or to achieve satisfactory outcomes in them, U.S. and allied policymakers must answer four key questions:

  • What, if any, diplomatic initiatives could ameliorate sources of potential conflict?
  • What conventional, cyber, and information warfare capabilities would deny adversaries their hoped-for advantageous conventional options?
  • What sorts of nuclear operations—against which targets, with what numbers, types and yields of weapons, and using what delivery systems—would predictably cause the adversary to cease further aggression and would be consistent with the law of armed conflict and other interests? Relatedly, what capabilities are redundant for these purposes?
  • What sorts of nuclear operations would probably cause the adversary to retaliate against allied and U.S. territory and interests in ways that would leave allies and/or the United States worse off than if they had not initiated nuclear use?

Clearly, there is a world of difference between the dangers implied in the third and fourth questions. Only with North Korea does escalation dominance with nuclear dimensions still theoretically stand a chance of succeeding on tolerable terms, given the limited nature of the North Korean nuclear capabilities to date and the overwhelming U.S. and allied military advantages that Pyongyang faces.

The current debate about limited nuclear war repeats decades of failed attempts by the United States and the Soviet Union (and then Russia) to resolve the unresolvable dilemma of “how to plan a nuclear attack that [is] large enough to terrify the enemy but small enough to be recognized unambiguously as a limited strike, so that, if the enemy retaliated, he’d keep his strike limited too.”46 And, if neither side is willing to back down after the first round of limited nuclear attacks, how plausibly could they move up the escalation ladder without destroying each other and much of the rest of the world?47 Eventually, both countries learned that it was impossible to win a nuclear war against an adversary that could maintain a survivable second-strike capability. They may now need to relearn this as emerging nuclear and non-nuclear capabilities may tempt one or the other to undertake preemptive counterforce strikes.

The United States and China heretofore have avoided such contests for escalation dominance but may be verging toward one. The United States is increasingly concerned that the geography of the West Pacific may align with China in possible regional conflicts. Exchanging attacks on each other’s homeland with large numbers of nuclear weapons would be suicidal and therefore not credible as a deterrent. However, either could be tempted to deploy suites of lower-yield, shorter-range weapons that conceivably would make the adversary stop fighting (that is, deescalate) before mutual suicide is achieved. This temptation to find nuclear solutions to the overall challenge will grow, even if such solutions may be Pyrrhic.


1 2018 NPR, 17.

2 The 2018 NPR states (p. 21) that “the United States will not use or threaten or threaten to use nuclear weapons against non-nuclear weapons states that are party to the NPT and in compliance with their nuclear non-proliferation obligations.” The states that meet these conditions are all nuclear-weapon states, with the exception of Iran. Had the Trump administration not reneged on the Joint Comprehensive Plan of Action and Iran continued complying with its terms, there would be little reason to contemplate U.S. use of nuclear weapons against Iran.

3 Examples include the 1969 Sino-Soviet border war, the 1973 Yom Kippur War, the 1982 Falklands War, the 1983 bombings of the Multinational Force barracks and the U.S. Embassy in Beirut, the 1999 Indo-Pakistani Kargil War, the 2001 attacks on the World Trade Center and Pentagon, and the Indian and Pakistani air strikes during the 2019 Balakot incident.

4 Office of the Secretary of Defense, Military and Security Developments Involving the People’s Republic of China, Annual Report to Congress (Washington, DC: Department of Defense, 2020), ix, (hereinafter China Military Power Report).

5 Dmitri Adamsky, “Cross-Domain Coercion: The Current Russian Art of Strategy,” Proliferation Papers no. 54 (Paris: Institut Francais des Relations Internationales, November 2015), 24,

6 Steve Pifer, “Putin’s Nuclear Saber-Rattling: What Is He Compensating For?,” Brookings, June 17, 2015, For more on Russian military exercises involving nuclear-capable forces, see Franz-Stefan Gady, “Russia Kicks Off Annual Nuclear Forces Readiness Exercise,” Diplomat, October 15, 2019, (Grom-2019); Dave Johnson, “VOSTOK 2018: Ten Years of Russian Strategic Exercises and Warfare Preparation,” NATO Review, December 20, 2018, (Vostok-2018); and Keir Giles, “Russia Hit Multiple Targets With Zapad-2017,” Carnegie Endowment for International Peace, January 25, 2018, (Zapad-2017).

7 2018 NPR, 9; and Leo Michel and Matti Pesu, “Strategic Deterrence Redux: Nuclear Weapons and European Security,” Finnish Institute of International Affairs, September 2019,

8 See Michael Kofman, “Fixing NATO Deterrence in the East Or: How I Learned to Stop Worrying and Love NATO’s Crushing Defeat by Russia,” War on the Rocks, May 12, 2016,

9 Paul K. Davis et al., Exploring the Role Nuclear Weaposn Could Play in Deterring Russian Threats to the Baltic States (Santa Monica: RAND Corporation, 2019), 81,

10 Nikolai Sokov, “Russia Clarifies Its Nuclear Deterrence Policy,” Vienna Center for Disarmament and Nonproliferation, June 3, 2020,

11 Michael Kofman and Anya Loukianova Fink, “Escalation Management and Nuclear Employment in Russian Military Strategy,” War on the Rocks, June 23, 2020,

12 Ibid.

13 As the authors of the April 2020 U.S. State Department paper on the W76-2 Low-Yield Option put it, “there is no such thing as a safe nuclear war or a low-risk nuclear strike, regardless of its magnitude.” Bureau of Arms Control, Verification and Compliance, U.S. Department of State, “Strengthening Deterrence and Reducing Nuclear Risks: The Supplemental Low-Yield U.S. Submarine-Launched Warhead,” Arms Control and International Security Papers 1, no. 4 (April 24, 2020), 6, Regarding views of emerging U.S. strategists, see Elbridge Colby, “If You Want Peace, Prepare for Nuclear War,” Foreign Affairs, November/December 2018,, and Elbridge Colby, “Against the Great Powers: Reflections on Balancing Nuclear and Conventional Power,” Texas National Security Review 2, no. 1 (November 2018): 145–52; Keir A. Lieber and Daryl G. Press, “The New Era of Counterforce: Technological Change and the Future of Deterrence,” International Security 41, no. 4 (Spring 2017): 9–49, and Keir A. Lieber and Daryl G. Press, The Myth of the Nuclear Revolution: Power Politics in the Atomic Age (Ithaca: Cornell University Press, 2020).

14 Olya Oliker, “New Document Consolidates Russia’s Nuclear Policy in One Place,” Russia Matters, June 4, 2020,; and 2018 NPR, 23.

15 See, for example, Dmitri Trenin’s analysis of Russia’s new “Foundations of State Policy of the Russian Federation in the Area of Nuclear Deterrence” document: “Finally, the Kremlin policy guidelines provide for the deployment of nuclear weapons in response to an attack against the critical national infrastructure that is responsible for controlling and employing nuclear weapons. This new provision may refer, for example, to cyberattacks that can disable Russian strategic command and control systems.” Dmitri Trenin, “Decoding Russia’s Official Nuclear Deterrence Paper,” Carnegie Moscow Center, May 6, 2020,

16 It is impossible to alleviate paranoid perceptions of threat.

17 See Paul K. Davis et al., Exploring the Role Nuclear Weapons Could Play in Deterring Russian Threats to the Baltic States (Santa Monica, CA: RAND Corporation, 2019), “The insights derived from the research highlight the reality that, even if NATO makes significant efforts to modernize its nonstrategic nuclear weapons, it would have much stronger military incentives to end a future war than Russia would. That is, Russia would still enjoy escalation dominance,” 2.

18 China Military Power Report, 70.

19 China Military Power Report, 72.

20 See James Griffiths, “China Ready to Fight ‘Bloody Battle’ Against Enemies, Xi Says in Speech,” CNN, March 20, 2018,; James Griffiths, “Xi Jinping Warns Taiwan Independence Is ‘a Dead End,’” CNN, January 2, 2019,

21 Eric Heginbotham, “An Interactive Look at the U.S.-China Military Scorecard,” RAND Corporation,

22 ASEAN is composed of ten members (Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, the Philippines, Singapore, Thailand, and Vietnam) and two observer states (East Timor and Papua New Guinea).

23 China’s pledge consists of two commitments: “China will never use nuclear weapons first at any time nor under any circumstances, and China unconditionally undertakes not to use or threaten to use nuclear weapons against any nonnuclear-weapon state or in nuclear-weapon-free zones.” The Department of Defense states that while “Some PLA officers have written publicly of the need to spell out conditions under which China might need to use nuclear weapons first; for example, if an enemy’s conventional attack threatened the survival of China’s nuclear force or of the regime itself.” In fact, “There has been no indication that national leaders are willing to attach such nuances and caveats publicly to China’s existing NFU policy as affirmed by recent statements by the PRC Foreign Ministry.” See China Military Power Report, 86.

24 Rebeccah L. Heinrichs, “Transcript: The Arms Control Landscape ft. DIA Lt. Gen. Robert P. Ashley, Jr.,” Hudson Institute, May 31, 2019,

25 Kristensen and Korda, “Chinese Nuclear Forces, 2019,” 171–78.

26 China Military Power Report, 85.

27 See Hans M. Kristensen and Matt Korda, “Russian Nuclear Forces, 2019,” Bulletin of the Atomic Scientists 75, no. 2 (2019): 73–84, DOI: 10.1080/00963402.2019.1580891; and Kristensen and Korda, “United States Nuclear Forces, 2020.”

28 “Although China almost certainly keeps the majority of its nuclear force on a peacetime status—with separated launchers, missiles, and warheads—nuclear and conventional PLARF brigades conduct “combat readiness duty” and “high alert duty,” which apparently includes assigning a missile battalion to be ready to launch, and rotating to standby positions as much as monthly for unspecified periods of time.” China Military Power Report, 88.

29 China Military Power Report.

30 Wu Riqiang, “Living With Uncertainty: Modeling China’s Nuclear Survivability,” International Security 44, no. 4 (2020): 84–118.

31 “United States Northern Command and United States Strategic Command,” United States Senate Committee on Armed Services, February 13, 2020,

32 Caitlin Talmadge, “Would China Go Nuclear? Assessing the Risk of Chinese Nuclear Escalation in a Conventional War With the United States,” International Security 41, no. 4 (2017): 50–92; and John Speed Meyers, “Mainland Strikes and U.S. Military Strategy Towards China: Historical Cases, Interviews, and a Scenario-Based Survey of American National Security Elites,” (Santa Monica, CA: RAND Corporation, 2019), 16–17.

33 David E. Sanger and William J. Broad, “A New Superpower Competition Between Beijing and Washington: China’s Nuclear Buildup,” New York Times, June 30, 2020, updated July 14, 2020,

34 See Roberts and Perkovich, “U.S. Nuclear Weapons in the Twenty-First Century”: “Asymmetry of stake couldn’t have been more perfectly summarized than by General Yao Yunzhu (ph) . . . a member of the PLA, a participant in our unofficial dialogues on nuclear issues. She said, let’s review this point on Taiwan. What interest will be at stake for you Americans in a conflict on Taiwan? Well, your commitment to democracy in Taiwan, your historic commitment to the Republic of China, your standing with your allies in East Asia, your global standing with your allies and your credibility. You’d have a lot at risk, a lot at stake. Where would we, China, have a stake? Sovereignty and culmination of the recovery after the century of humiliation. Who’s got more at stake, she said? Well, obviously we will. So our Chinese steps to escalate will look more credible to you Americans than your threats to escalate. You will have to contemplate the need to back down.” Brad Roberts and George Perkovich, “U.S. Nuclear Weapons in the Twenty-First Century,” Carnegie Endowment for International Peace, June 28, 2016, transcript, pp. 4–5,

35 Adam Mount, Conventional Deterrence of North Korea (Washington, DC: Federation of American Scientists, 2019), 20–21,

36 “If the United States has a credible damage limitation option, the Kim regime is more likely to calculate that crossing the nuclear threshold would be a strategy for suicide, not survival, because North Korea would lack a reliable second-strike capability to deter regime change.” Vince Manzo and John K. Warden, “Want to Avoid Nuclear War? Reject Mutual Vulnerability With North Korea,” War on the Rocks, August 29, 2017,

37 “The current and prospective security environment is marked both by increasing uncertainty and increasingly severe threats to the United States and its allies. These developments have become blatantly evident since the 1994, 2001, and 2010 NPRs. . . . These developments mandate a fundamental reorientation of US nuclear policy priorities from that established explicitly in the 2010 NPR. The United States must now reestablish deterrence, assurance, and defense as the priority goals for US nuclear policy.” Keith Payne and John S. Foster, New Nuclear Review for a New Age, National Institute for Public Policy (Fairfax, VA: National Institute Press, 2017), 50–61,

38 Barry R. Posen, Inadvertent Escalation: Conventional War and Nuclear Risks (Ithaca, NY: Cornell University Press, 1991). See discussion in chapter 1.

39 James M. Acton, “Escalation Through Entanglement: How the Vulnerability of Command-and-Control Systems Raises the Risk of Inadvertent Nuclear War,” International Security 43 no. 1 (2018): 56–99.

40 James M. Acton et al., “Entanglement: Chinese and Russia Perspectives on Non-Nuclear Weapons and Nuclear Risks,” Carnegie Endowment for International Peace, November 8, 2017,

41 Thomas C. Schelling, Arms and Influence (New Haven, CT: Yale University Press, 1966), 94.

42 For example, John Warden suggests the United States and its allies “think through the details of potential settlements, including the tradeoffs and concessions required,” and incorporate them into deterrence planning. Andrew Coe and Victor Utgoff similarly posit a need for the U.S. government to consider how allies, adversaries, and the domestic population will view the conclusion and aftermath of a nuclear conflict and assess important assumptions regarding how the United States will act in such a future war. See John Warden, “Limited Nuclear War: The 21st Century Challenge for the U.S.,” Livermore Papers on Global Security No. 4, July 2018, pp. 42–47,; and Andrew Coe and Victor Utgoff, “Restraining Nuclear War” (Alexandria, VA: Institute for Defense Analyses, June 2011), 8,

43 In Exploring the Role Nuclear Weapons Could Play in Deterring Russian Threats to the Baltic States, wargame participants considered the merits of a Russia pressing a conventional campaign even in the face of NATO first use: “The Russians, surrounding or occupying the Baltic capitals and not having responded with nuclear weapons to NATO’s first use, could shift the blame to NATO by labeling the Alliance as having engaged in a reckless provocation to which Russia wisely did not respond. The participants considered this response by Russia to be a definite possibility but did not rule out that Russia would find it necessary to respond with a nuclear attack.” Davis et al., Exploring the Role Nuclear Weapons Could Play in Deterring Russian Threats to the Baltic States, 95.

44 For a thoughtful and partially clarifying analysis of this challenge, see Bureau of Arms Control, Verification and Compliance, “Strengthening Deterrence and Reducing Nuclear Risks.”

45 Timothy McDonnell, “Bolts from the Blue, Monsters Under the Bed, and the Pursuit of Absolute Security,” War on the Rocks, October 17, 2017,

46 Fred Kaplan, The Bomb: Presidents, Generals, and the Secret History of Nuclear War (New York: Simon & Schuster, 2020), 120.

47 USSTRATCOM commander General John Hyten vividly recounted in July 2018 the proclivity of large arsenals and attendant operational doctrines to produce escalation. General Hyten described to an audience a “big exercise” that his command conducted that February. “I just want you to ask in your own head, how do you think it ends? It ends the same way every time. It does. It ends bad . . . meaning it ends with global nuclear war.” Hyten continued by describing how the Chairman of the Joint Chiefs of Staff, General Joseph Dunford, and the entire Joint Staff observed the exercise from the National Airborne Operations Center, the airborne command and control base. “As soon as the NAOC landed after the exercise was over,” Hyten recalled, Dunford “called me, like within seconds. And I’ll just say General Dunford wasn’t happy with the way the exercise went. He said we should provide the President more options, not fewer options. And the way the process was driving down, we were providing very few options. [So] the goal is to provide more options to the President to give him options to de-escalate a conflict, not just escalate a conflict. To get us off that escalation ladder. . . . I don’t know how many times I’ve said I don’t want on the escalation ladder; I want off the escalation ladder. That’s the point. And for whatever reason, the whole structure of the command was about the escalation ladder.” See General John Hyten, “The Mitchell Institute Triad Conference,” U.S. Strategic Command, July 17, 2018,