Nuclear Perils in a New Era

Mad Momentum Redux? The Rise and Fall of Nuclear Arms Control

Back to table of contents
Steven E. Miller and Alexey Arbatov
Promoting Dialogue on Arms Control and Disarmament

Mad Momentum Redux? The Rise and Fall of Nuclear Arms Control1

In a September 1967 speech in San Francisco that attracted little notice at the time, Robert McNamara—then U.S. secretary of defense and one of the Cold War’s most formidable strategic thinkers—took note of the primacy of technological progress in determining the state’s policy-making: “There is a kind of mad momentum intrinsic to the development of all new nuclear weaponry. If a weapon system works and works well, there is strong pressure from many directions to procure and deploy the weapon out of all proportion to the prudent level required.”2

True, the enormous destructive power and technical complexity of nuclear arms had made critical political decisions hostage to the weapons’ technical characteristics. Actually with regard to nuclear war, Carl von Clausewitz’s classical postulate—that war is the continuation of politics by other means—might have been recast to say that war is the continuation of the technical characteristics of weapon systems that determine doctrines, operational plans, and the contingencies of their employment. In the same speech, McNamara pointed out that “actions—or even realistically potential actions—on each side . . . trigger reactions on the other side. It is precisely this action-reaction phenomenon that fuels the arms race.”3 He also recognized something that seldom, if ever, had been acknowledged: “If we had more accurate information about planned Soviet strategic forces, we simply would not have needed to build as large a nuclear force as we have today.”4

Coming from a top American official, these insights signified a revolution in the strategic mentality of the time. A half-century later, they remain relevant. First, McNamara proposed a conceptual breakthrough out of the “mad momentum” of the arms race: “We do not want a nuclear arms race with the Soviet Union, primarily because the action-reaction phenomenon makes it foolish and futile. . . . Both of our nations would benefit from a properly safeguarded agreement first to limit and later to reduce both our offensive and defensive strategic nuclear forces.”5 This prompted the start of negotiations on strategic arms between the two nuclear superpowers two years later, which would usher in forty years of diplomatic interaction between the United States and the Soviet Union (later Russia) that produced nine major treaties and agreements on nuclear forces. The quantities and aggregate destructive power of nuclear weapons were steeply reduced, the probability of nuclear war was drastically lowered, and the unprecedented transparency and predictability of nuclear forces that McNamara desired was ensured.

Second, McNamara’s ideas have contemporary relevance because legacy Cold War–era arms control is collapsing, and an uncontrolled nuclear arms race is threatening to return.

Third, the principal nuclear powers’ current generation of leaders, political elites, and military officials has an inadequate understanding of the history of the nuclear arms race and nuclear arms control, and therefore an insufficient appreciation of the dangers of the vicious circle of the arms race and the international crises it provoked. Russian President Vladimir Putin recently expressed the hope that “no new crises of the Cuban type happen in the world,” adding that “if anybody over there want it, they are welcome.”6

The world’s ability to muddle through the next phase of international tensions without a major crisis, and to prevent such a crisis from escalating to nuclear Armageddon, is in doubt.


  • 1An earlier version of this essay was first published in Survival 61 (3) (June–July 2019): 7–38. Copyright © The International Institute for Strategic Studies, reprinted by permission of Taylor & Francis Ltd., on behalf of The International Institute for Strategic Studies.
  • 2Quoted in Robert S. McNamara, The Essence of Security: Reflections in Office (New York: Harper & Row, 1968), 166.
  • 3Ibid., 58–59.
  • 4Ibid., 58.
  • 5Ibid., 62.
  • 6Putin predostereg SSHA protiv novogo Karibskogo Krizisa” [Putin Warned the United States Against a New Cuban Crisis], RIA Novosti, February 20, 2019.

Falling Dominoes

The evidence of arms-control disintegration is obvious and is nowadays broadly discussed among states, within the world’s professional community, and by the mass media. Still, the array of emerging systemic crises is worth examining.

The United States’ and Russia’s withdrawal from the 1987 Intermediate-Range Nuclear Forces (INF) Treaty was the last turning point. Given the U.S. renunciation of the 1972 Anti-Ballistic Missile (ABM) Treaty in 2002, this removed the remaining cornerstone of the nuclear-arms-reduction regime launched by the 1991 Strategic Arms Reduction Treaty (START I). Ten years have passed since Russia and the United States have discussed any option for the START follow-on agreement—the longest pause in strategic arms talks for fifty years. Although both parties fulfilled their reduction obligations under the current New START by the February 2018 deadline (albeit with a number of reservations from Russia), and the extension has now been successfully implemented, the treaty will expire in 2026. The chances for successful negotiations on a new agreement after the abrogation of the INF Treaty, and given deep disagreements between the two parties on ballistic-missile defense (BMD) and other important issues, are bleak indeed.

Against this background of the apparent abandonment of bilateral nuclear arms control, the United States and Russia are entering a new cycle of the arms race. Unprecedentedly, it will include competition not only in offensive nuclear weaponry but also in offensive and defensive non-nuclear strategic and medium-range weapons, as well as in the development of space weapons and cyber warfare.

Russia has been modernizing its strategic triad for more than a decade, deploying and developing two new intercontinental-ballistic-missile (ICBM) systems (the SS-27 Mod 2/3 Yars and SS-29 Sarmat), one submarine-launched ballistic-missile (SLBM) system (the SS-N-32 Bulava-30), two heavy-bomber systems (the Tu-160M Blackjack and PAK DA), and long-range nuclear as well as conventional and easily convertible to nuclear air-, ground-, and sea-launched cruise missiles (the Kh 102/101 [AS-23A/B], 9M729 [SSC-8], and 3M14 [SS-N-30], respectively).

Russia is also developing and deploying a new generation of nuclear and dual-purpose weapon systems unveiled in Putin’s March 1, 2018, address: the Avangard strategic nuclear boost-glide hypersonic system; Poseidon long-range, high-speed, nuclear-propelled and nuclear-armed heavy torpedoes; Burevestnik nuclear-powered intercontinental nuclear cruise missiles; Kinzhal air-launched hypersonic middle-range missiles; and a number of other sub-strategic nuclear and dual-purpose systems.7 Given the demise of the INF Treaty, intermediate-range land-based Kalibr-type cruise missiles and hypersonic missiles may be deployed. (Indeed, the U.S. government has alleged that Russia has already deployed a ground-launched cruise missile similar to the Kalibr 3M14.)

The United States, for its part, is developing strategic systems for limited nuclear strikes. These include Trident-2 SLBMs with low-yield W76-2 warheads, B61-21 variable-yield gravity bombs for heavy bombers and tactical-strike aircraft, long-range stand-off air-launched nuclear cruise missiles, and nuclear sea-based cruise missiles. The U.S. withdrawal from the INF Treaty has lent further momentum to the development of land-based medium-range cruise, ballistic, and hypersonic systems. In the longer term, beginning in the mid-2020s, the United States plans to modernize its whole strategic triad, replacing heavy bombers, ICBMs, and nuclear submarines with SLBMs.8

Unlike the Cold War version, the new nuclear arms race will be multilateral, involving states such as China, India, Pakistan, Israel, and North Korea as well as the United States and Russia. The intensification of the arms race would undoubtedly undermine the nuclear non-proliferation regime. The review conference of the Nuclear Nonproliferation Treaty (NPT) in 2015 ended in failure, and the next conference, scheduled for 2021 due to the COVID-19 pandemic, is likely to fail as well. The nuclear-weapons states have reneged on their obligation under the NPT’s Article VI to “undertake to pursue negotiations in good faith on effective measures relating to cessation of the nuclear arms race at an early date and to nuclear disarmament.” Further aggravations include the U.S. withdrawal from the 2015 multilateral nuclear deal on the Iranian atomic program, the deadlock over the concept of a zone free of weapons of mass destruction in the Middle East, and the deep split between nuclear and non-nuclear NPT states over the Treaty on the Prohibition of Nuclear Weapons approved by the UN General Assembly on July 6, 2017.9 The probable degradation of NPT norms will prevent the treaty from effectively addressing the challenges of the significant future growth of the world’s atomic energy and trade in nuclear materials and technologies. As a consequence, the line between peaceful and military use of nuclear energy through the nuclear fuel cycle will become even blurrier.

The new cycle of the arms race among nuclear-weapons states will probably encourage a new round of nuclear proliferation: Iran and Saudi Arabia could well join the nuclear club, as could Brazil, Egypt, Japan, Nigeria, South Korea, Taiwan, and Turkey, among others. This would eventually seal the fate of the Comprehensive Nuclear-Test-Ban Treaty (CTBT), which for twenty-three years has not entered into legal force because of the refusal of the United States and several other nations to ratify it. Under the thunder of nuclear explosions, the Fissile Material Cut-Off Treaty, on which negotiations have been stalled for more than a quarter-century, will die a quiet death. Increased production of weapons-grade uranium and plutonium, and nuclear-arms proliferation in the unstable regions of the world, will sooner or later afford international terrorists with access to nuclear explosives. This could end current civilization if a war between nuclear states does not do so earlier.

McNamara must have been pleasantly surprised at the fifty years of successful nuclear arms control and non-proliferation that followed his 1967 speech. But the impending implosion of his hopes and ideas would have deeply depressed him.


The Political Roots of the Crisis

The present confrontations between Russia and the West and the United States and China are exacerbating the crises of arms control and whipping up the arms race, but the roots of the crisis run deeper. Traditional nuclear arms control emerged on the basis of a predominantly bipolar world order, a more or less symmetrical balance of power between the United States and the Soviet Union, and a relatively simple delineation between nuclear and conventional weapon systems. Strategic circumstances have profoundly changed during the last fifty years, and arms control has largely failed to adapt to the changes.

The collapse of the Soviet Union accelerated the emergence of a multipolar world order. Other power centers—China and the EU globally, and India, Iran, Japan, Pakistan, Turkey, and others regionally—started playing increasingly important international roles. With few exceptions, nuclear arms control did not figure prominently in their external interests and security concepts. In addition, the transition from confrontation to cooperation among the great powers during the 1990s brought the probability of war between them close to zero. This development redirected the international security agenda to ethnic and religious conflicts, international terrorism, nuclear proliferation, and illegal arms and drugs trafficking. For a time, the unprecedented improvement in relations between Russia and the West actually encouraged monumental breakthroughs in nuclear arms control: huge Cold War–era stockpiles were reduced by roughly an order of magnitude in weapons numbers and, to an even greater extent, in aggregate destructive power.10 This was accomplished by unilateral reductions on the part of France, Russia, the United Kingdom, and the United States, and even more through the INF Treaty, START I (1991), START II (1993), the START III Framework Agreement (1997), the Strategic Offensive Reductions Treaty (SORT) in 2002, and finally New START (2010). The deeper trend over the course of these agreements was worrying, however, as the substantial achievements of 1987–1997 gave way to complacency.

Despite Moscow’s regular appeals to turn the bilateral arms-reduction process into a multilateral one, occasionally joined by Washington, the other seven nuclear states declined. They have routinely asserted that Russia and the United States still possessed 90 percent of the global nuclear arsenal and called for more substantial reductions as a precondition for their participation in multilateral disarmament efforts. Some multilateral agreements were achieved: indefinite extension of the NPT (1995), the signing of the CTBT (1996), and the adoption of the Additional Protocol to the NPT, which expanded International Atomic Energy Agency (IAEA) safeguards (1997). Nevertheless, the two leading powers failed to elaborate consistent and equitable principles of multilateral nuclear-arms limitations (parity, strategic stability, national or aggregate quotas for the third nuclear-weapons states), to propose a sensible sequence by which third states could join the process, or to put forward a practical substantive agenda for negotiations on classes and types of weapon systems and realistic verification methods.

In the context of ongoing proliferation of medium- and long-range ballistic and cruise missiles, the two leading powers proved unable to adapt existing arms-control treaties (in particular, this relates to the ABM and INF treaties and the New START follow-on) to the new military environment. Instead, the probable emergence of a multilateral nuclear and missile world became a convenient argument for the abrogation of such agreements. Another mistake was the common assumption that improved political and economic relations between states made arms control irrelevant. In reality, there was a large gap between merely ceasing to be enemies and becoming allies, and arms control remained useful, if not crucial, for narrowing that gap. The nuclear arms race between Russia and the United States actually stopped during the 1990s and 2000s.11 But other types of technological developments tangibly affected military capabilities of states and non-state entities. As a result, traditional demarcation lines between nuclear and conventional arms, offensive and defensive weapons, and global and regional systems were eroded.

When START I was about to expire in 2009, it fell to the Barack Obama and Dmitry Medvedev administrations to hastily work out New START (the Prague Treaty), which effectively legalized the strategic nuclear force levels set by SORT seven years earlier. Having been a useful stopgap measure, the treaty failed to address new weapons developments and was quite relaxed in its traditional limits, including the counting rules and verification regime.12 There has been no material progress since. The effective hiatus of aggressive arms control after 1997, and still more after 2010, has led to the wholesale disintegration of the arms-control system and the beginning of a new cycle of the arms race.


  • 10Calculations are based on SIPRI Yearbook 2017: Armaments, Disarmament and International Security (Oxford: Oxford University Press, 2017), 648–717; and SIPRI Yearbook 1990: World Armaments and Disarmament (Oxford: Oxford University Press, 1991), 3–51.
  • 11Examples of small-scale and slow remaining modernization programs of the time are the United States’ refitting its missile submarines with Trident-2 instead of Trident-1 missiles, and Russian deployment of the SS-25 Topol ground-mobile ICBM as a replacement for older ICBMs, SLBMs, strategic submarines, and bombers.
  • 12SORT in 2002 fixed ceilings of 1,700–2,200 warheads per party, while New START reduced them to 1,550 warheads in 2010. However, its new counting rules provided substantial “discounts.” For example, strategic bombers counted as one delivery vehicle and one warhead, while each could carry 12–20 nuclear cruise missiles and gravity bombs. Submarines in overhaul and other de-alerted weapons were not counted under the main ceilings. Hence, the actual force loading by the START I counting rules is closer to 2,000 warheads per party.

Technological Drivers of Disintegration

Trends in military technology have been blurring the fundamental arms- control delineation between nuclear and conventional offensive systems. The development of high-precision, long-range non-nuclear air- and sea-launched cruise missiles relying on advanced electronics and command-control and information systems, increasingly based in space, have been especially consequential. Such weapons were effectively demonstrated in the wars in Iraq (1990, 2003), Kosovo (1999), Libya (2011), and Syria (2014–2018).13 New long-range, precision-guided offensive arms are degrading the nuclear threshold in several ways. First, most of them are using dual-purpose delivery systems such that the other side would not know whether it was under conventional or nuclear attack before actual explosive impact.14 Second, many of the weapons are able to hit the nuclear forces and command-and-control information systems of the opponent, potentially prompting nuclear retaliation or pre-emption. The threats posed by U.S. long-range cruise missiles to Russian strategic missiles may be exaggerated, as Russia’s silo launchers and underground command centers are super-hardened.15 The fact remains that the American weapons can hit early-warning radars, light shelters for mobile ICBMs, missile submarines at bases and heavy bombers at airfields, as well as command-and-control and communications sites that are not hardened.

A much larger potential threat to strategic targets may come from prospective boost-glide weapons. The United States has been developing and testing several systems of this type.16 Recently, Russia overtook the United States with its analogue, called the Avangard, and in 2019 started deploying it on modified SS-19-X-Mod4 ICBMs and later possibly on new heavy SS-29 ICBMs (Sarmat), which are to replace SS-18 (Satan) missiles.17 The United States and Russia are not the only countries to develop high-precision long-range conventional (including boost-glide hypersonic) weapons. China is working on its project at an accelerated pace, India is developing these weapons as well, and other countries are likely to follow.

The concepts and systems for limited nuclear strikes (“tailored options”) are also blurring the nuclear-conventional threshold. The United States is apparently associating such options with strategic/tactical gravity bombs (B61-21), W76-2 low-yield warheads for a portion of its Trident-2 SLBMs, the nuclear long-range air-launched cruise-missile system, and new medium-range sea-based (nuclear) cruise-missile systems.18 Russia is vaguer on the notion of limited nuclear warfare, but some unofficial sources relate it to nuclear platforms, such as the boost-glide Avangard and various sub-strategic systems.19

Another avenue of technical development is dissolving the border separating defense and offense. In 2007, the United States initiated deployment of a global missile-defense system with regional segments in the Euro-Atlantic and the Pacific. Over Russia’s objections, the United States refused to develop a joint system, or to accept binding obligations not to calibrate its missile-defense system to intercept Russian missiles. Starting in 2011, Russia initiated an air-space defense program that includes missile defense.20 Some senior Russian military and military-industrial authorities have indicated that the U.S. system’s defensive capability against Russian strategic forces is negligible because the Russian ICBMs and SLBMs are sufficiently numerous, survivable, and equipped with effective BMD penetration aids.21 Many Russian and American defense and security experts share this view.22 Nevertheless, Russia’s political leadership has continued to insist that U.S. BMD is undercutting Russian nuclear deterrence and bilateral strategic stability. While such claims are to a significant degree politically motivated, the open-ended nature of the U.S. BMD program and an American rejection of any technical or strategic limitations or predictability regime for that program could raise legitimate strategic concerns.

The upshot is that new U.S. defensive programs are conceptually erasing the strategic demarcation that McNamara established between “offensive” BMD (intended to negate an opponent’s second-strike capability) and “defensive” BMD (intended to protect strategic retaliatory forces against a disarming strike, or defend against a third nuclear state’s attack).23 This line is also being weakened at the operational level. For Russia, the U.S. Aegis and Aegis Ashore BMD Standard-3 (SM-3) interceptor launchers on ships, and at land bases in Romania and Poland, are indistinguishable from the universal Mk-41 launchers for ship-based Tomahawk cruise missiles. Hence, Russia is able to claim that the United States is in violation of the INF Treaty, which prohibited deployment of land-based, long-range cruise missiles and their launchers.

The development of BMD systems with anti-satellite capabilities is also destabilizing. In the United States, the most advanced system of this class is a modified version of the Aegis Mk7 naval anti-missile/anti-satellite system equipped with SM-3 missiles and a self-guided kinetic warhead, tested against a satellite in 2008.24 Russia envisions anti-satellite capability for the S-500 surface-to-air missile complexes, as well as for the long-range kinetic-kill Nudol missile interceptor (an analogue of the U.S. ground-based interceptor system in Alaska and California) for the new A-235 Moscow BMD system.25 China has also joined the anti-satellite arms race, having tested its system in 2008, and India conducted its first test in 2019. Talks between Moscow and Washington on space weapons were conducted in the late 1970s and 1980s and in a multilateral format in the 2000s but failed. Substantial dialogue on cyber-warfare capabilities, which will have an undoubted but as yet unclear effect on strategic stability, has not developed further than preliminary consultations.

Yet another victim of technological developments is the delineation between global and regional offensive and defensive weapons. This has never been ironclad—recall the Cold War debates about Soviet missiles in Cuba, and American forward-based missiles and strike aircraft—but now is creating growing strategic problems. U.S. regional BMD in Europe and in Asia, aimed at Iranian and North Korean missiles, is perceived in Russia and China as intended to intercept their strategic ICBMs and SLBMs at boost phase, thus degrading their respective deterrents. Accordingly, the two powers are developing a range of missile systems to penetrate these defenses. The United States is planning to counter these programs with new nuclear arms of its own, as proclaimed in the U.S. Department of Defense’s Nuclear Posture Review of 2018.

Russia and China perceive the United States’ employment of long-range, precision-guided conventional systems (subsonic and, in the future, hypersonic) against hostile regional states and terrorists as implicitly threatening non-nuclear counterforce strikes. Russia is countering with its air-space defense and long-range dual-purpose offense programs, China with analogous conventional offensive systems. The U.S. Nuclear Posture Review interpreted these measures as new threats, to be deterred by, among other means, threatened nuclear retaliation.26

Russia’s concern about third states’ medium-range missiles in Asia within reach of its territory has prompted its criticism of the 1987 INF Treaty at a high official level, since it bans only Russian and American weapons of that variety.27 In 2018, Washington echoed this argument, claiming the right to deploy medium-range missiles to counter comparable Chinese weapons. Against that political background, mutual accusations of treaty violations gained momentum and contributed to its collapse.

In light of the growing impact of such systems on strategic stability, leaving them out of arms-control agreements would diminish the effects of nuclear disarmament. Including such weapons in agreements would create tough problems in terms of definitions, counting rules, and verification, all the more so given that conventional systems have been and most probably will be extensively used by the United States, Russia, and other powers in local military operations.

A sharp turn in global politics dealt an especially devastating blow to the nuclear arms-control system. After 2012, Moscow embarked on strengthening Russia’s control of the post-Soviet space (Georgia, Ukraine) and its projection of force beyond it (Syria, Venezuela), modernizing its conventional forces, and energizing the implementation of a nuclear modernization program that had begun earlier. In response, the United States and its allies imposed economic sanctions and revived the strategy of isolation, containment, and arms build-up against Moscow. A fierce propaganda fight broke out, amped up by hacker sabotage operations. Military competition between Russia and the United States intensified in Eastern Europe, the Baltic and Black Sea areas, and the Arctic and Asia-Pacific regions.


  • 13This applies to U.S. systems such as the Tomahawk sea-launched cruise missile (BGM-109) and air-launched cruise missiles (AGM-84, AGM-158B JASSM-ER). Russian non-nuclear cruise missiles are the Kalibr 3M-54 and 3M-14 sea-launched cruise missiles, and the Kh-55SM, Kh-555, and Kh-101-type air-launched cruise missiles.
  • 14These are heavy and medium bombers, tactical-strike aircraft, ships, and attack submarines with missiles capable of carrying both nuclear and conventional warheads: the Kalibr and Tomahawk sea-based cruise missiles (some of which will again be armed with nuclear warheads), air-launched cruise missiles of the Kh101/102 type or the AGM-158, and Iskander-type ground-launched tactical ballistic and cruise missiles.
  • 15Alexey Arbatov and Vladimir Dvorkin, eds., Nuclear Proliferation: New Technologies, Weapons, Treaties (Moscow: Carnegie Moscow Center, 2009), 85–103; and Alexey Arbatov, Vladimir Dvorkin, and Natalia Bubnova, eds., Missile Defense: Confrontation and Cooperation (Moscow: Carnegie Moscow Center, 2013), 183–225.
  • 16James M. Acton, Silver Bullet? Asking the Right Questions About Conventional Prompt Global Strike (Washington, D.C.: Carnegie Endowment for International Peace, 2013), 33–63.
  • 17“Statement of the President of Russia to the Federal Assembly.”
  • 18Nuclear Posture Review, xii.
  • 19See N. Boitzov, “Terminologiia v Voennoi Doktrine” [Terminology in the Military Doctrine], Nezavisimoe voennoe obozrenie, no. 40, October 3, 2014; and D. Akhmerov and M. Valeev, “Aerostat–Drug ‘Sarmata’” [Balloon–A Friend of ‘Sarmat’], Voenno–Promyshlennyi Kur’er, no. 39, October 12–18, 2016, 6, available at http://www.vpk
  • 20Expanded Meeting of the Defense Ministry Board,” Kremlin, December 19, 2014.
  • 21Ibid.
  • 22Vladimir Dvorkin and Vladimir Pyriev, “The US/NATO Program and Strategic Stability,” in Arbatov, Dvorkin, and Bubnova, eds., Missile Defense, 183–203.
  • 23McNamara, The Essence of Security, 63–66.
  • 24Vladimir Dvorkin, “Space Weapons Programs,” in Alexey Arbatov and Vladimir Dvorkin, eds., Outer Space: Weapons, Diplomacy, and Security (Washington, D.C.: Carnegie Endowment for International Peace, 2010), 30–45.
  • 25A. Mardasov, “‘Nudol’: Ubiitza Amerikanskich MBR i Sputnikov” [‘Nudol’: A Killer of the US ICBMs and Satellites], Free Press.
  • 26Nuclear Posture Review, 21.
  • 27Meeting of the Valdai International Discussion Club, Kremlin; and “INF Treaty Can’t Last Endlessly, Ivanov Said,” RIA Novosti.

Lessons of the Cold War Arms Race

The expectation at the top level of the Harry Truman administration was that it would take a generation or more for the Soviet Union to break the U.S. nuclear monopoly established at the end of the Second World War.28 With some help from espionage, however, it happened only four years later. The U.S. attempt to restore its preponderance with a thermonuclear-weapon test at the Pacific Eniwetok Atoll on October 31, 1952, was thwarted even sooner—by the Soviet hydrogen-bomb test on August 12, 1953. As Daniel Ellsberg, once one of McNamara’s “whiz kids,” has noted, during the ensuing decades the crash production of fission and then fusion nuclear weapons went on, apparently without any rational justification. The United States was simply matching growing production rates with an ever-expanding target list, while the Soviet Union was just catching up.29 This course produced insane levels of destructive overkill on both sides.

The United States’ plan for the actual use of nuclear weapons, set out in the Strategic Air Command’s first Single Integrated Operational Plan (SIOP-62), called for quickly following any armed conflict with the Soviet Union with massive airstrikes, conducted by 1,850 heavy and medium bombers, that would drop 4,700 atomic and hydrogen bombs on cities and military installations across the Soviet Union, China, and their allies.30 The Pentagon’s estimates were that this attack would have resulted in 800 million casualties in the targeted and adjacent neutral countries.31 That amounted to approximately one-third of the global population at the time.

The U.S. nuclear build-up peaked around 1965 at about 34,600 warheads, and by the end of the 1980s this had declined to 24,700 weapons. The Soviet stockpile was consistently rising and, according to the highest available assessment, reached a plateau of 46,100 warheads by the end of the 1980s. The cumulative destructive power of the U.S. arsenal was at its maximum of 19,000 megatons in 1960, while that of the Soviet Union peaked at 19,700 megatons in 1975. Taken together, the two superpowers accumulated the maximum destruction potential of 26,000 megatons (equivalent to 1.3m Hiroshima bombs) in 1973–1974.32 William Perry, U.S. secretary of defense in 1993–1996, wrote in 2015:

When I look back on those years, I see a historically all-too-familiar irrational, impassioned thinking that . . . drove the frenzied debates on nuclear strategy, drove the huge additions in destructiveness we made to our nuclear forces, and brought us to the brink of blundering into a nuclear war . . . Even before the nuclear arms buildups of the 1970s and 1980s, our nuclear forces were more than enough to blow up the world. Yet we obsessively claimed inadequacies in our nuclear forces. We fantasized about a “window of vulnerability.” Both governments—ours and that of the Soviet Union—spread fear among our peoples. We acted as if the world had not changed with the emergence of the nuclear age, the age in which the world changed as never before.33

The rivalry of the two superpowers in new nuclear-delivery vehicles had four distinct but overlapping rounds. In the late 1940s and 1950s, it involved bombers and medium-range missiles; in the 1960s and early 1970s, strategic land- and sea-based ballistic missiles; in the 1970s and early 1980s, ballistic missiles with multiple individually targeted re-entry vehicles (MIRVs); and in the 1980s, medium-range cruise missiles and strategic ballistic missiles with enhanced hard-target kill capability (that is, against hardened ICBM silo launchers and command centers). Until the end of the 1980s, arms-race cycles went through intensive build-ups of new generations of delivery systems that fully or partially replaced the preceding ones.

Various weapon systems affected the probability of nuclear war in different ways. Some, such as sea-based long-range ballistic missiles and ICBMs in hardened silos and on ground-mobile launchers, lowered that probability insofar as they provided for survivable retaliatory capability. Others, such as ICBMs and SLBMs with enhanced counterforce (that is, disarming-strike) capability, increased the threat of first strike or preemption. McNamara in San Francisco nonetheless recognized the crowning irony of nuclear weapons:

While thermonuclear power is almost inconceivably awesome and represents virtually unlimited potential destructiveness, it has proven to be a limited diplomatic instrument . . . There is a strong psychological tendency to regard superior nuclear forces as a simple and unfailing solution to security . . . What must be understood is that our nuclear strategic forces play a vital and absolutely necessary role in our security and that of our allies, but it is an intrinsically limited role.34

McNamara’s paradox is the fundamental lesson of the seventy-year nuclear arms race. Just two or three decades ago, in the late 1980s and 1990s, the notion was commonly accepted in the Euro-Atlantic and post-Soviet political and strategic communities, but now is increasingly questioned. Although today the number and megatonnage of the weapons in the U.S. and Russian arsenals are many times lower than they were in McNamara’s time, the current nuclear balance still reflects massive overkill: about 1,600 megatons, or some 80,000 Hiroshimas.35 Furthermore, modern nations have far lower tolerances for war casualties and face considerably greater economic and social fragility than they did during the Cold War. Even though societies are weaker and more risk-averse, the United States and Russia consider their existing destructive potentials insufficient for effective deterrence. Thus, policy-makers do not appear to have learned McNamara’s lesson. They seem more focused on technological breakthroughs for the sake of gaining a decisive theoretical military advantage than on actual improvements in national and international security by enhancing arms-control systems and regimes.

The history of the nuclear arms race is full of examples of initial strategic breakthroughs that have led to serious damage to national security. Having been the first to create nuclear weapons, the United States assumed it would enjoy long-term world dominance. At the time, this implied using the nuclear threat to contain Joseph Stalin’s communist expansion and, if necessary, dropping atomic bombs on major Soviet cities. As John Newhouse has recounted, after warily considering the so-called “Baruch Plan”—which called for the transfer of atomic weapons and technology to the IAEA, a UN body—in 1946, the Truman administration adopted an “anti-Soviet line . . . combined with a conviction that Soviet science would always lag well behind America’s, whose security . . . must lie in doing whatever it took to preserve a long lead in advanced weapons over the enemy.”36

Only three years later, of course, the United States lost its nuclear monopoly and in another ten years—after Soviet development of long-range bombers and ICBMs—the United States was once and forever deprived of its traditional invulnerability to conflicts and wars sourced beyond the two oceans that surrounded its territory. Thirty years after Hiroshima, the People’s Republic of China could target the United States with nuclear missiles, and in another fifty years North Korea could do so. The creation of the atomic bomb may have been inevitable. But clearly Truman and other American officials of his time did not foresee the long-term outcome of the nuclear arms race and proliferation, and had they done so they would have been horrified.

A less dramatic but still instructive example is the U.S. initiative in the deployment of MIRVed sea- and land-based strategic missiles. The United States began to develop them in the mid-1960s to trump any robust ABM defense that the Soviets might deploy in the future. In 1969, however, the U.S.–Soviet Strategic Arms Limitation Talks (SALT) generated the prospect of stringent mutual limits on ABM systems, which materialized with the ABM Treaty in 1972. Once that treaty came into view, there was no longer any urgent need to deploy MIRVed missiles. But McNamara’s successors proceeded with the deployment of MIRVed Minuteman-3 ICBMs in 1970 and Poseidon SLBMs in 1971 in order to gain superiority over the Soviet Union in nuclear warheads after missile launchers had been limited by the SALT I agreement. This move was seen as enabling the expansion of the United States’ target list in the Soviet Union and returning to a counterforce strategy (that is, attacking the strategic military forces of the opponent), which was officially declared in 1974 with the “retargeting doctrine” of Secretary of Defense James Schlesinger.37 Once again, Washington was establishing a lead in advanced military technology. And once again, the other side caught up swiftly, deploying one new MIRVed SLBM and three MIRVed ICBM systems. In the late 1970s, this provoked a panic in the United States with respect to the “window of vulnerability” of its land-based missile force that cast fatal doubt on the SALT II Treaty and lasted throughout the 1980s.

A more recent and revealingly analogous case involves conventional high-precision, long-range systems. Initially, the technology was incorporated into dual-purpose air-launched cruise missiles (ALCMs) and sea-launched Tomahawk cruise missiles (SLCMs), developed and deployed by the United States starting in the mid-1970s. The Soviet Union followed suit in the early 1980s, but, due to inadequate guidance systems, only with missiles carrying nuclear warheads. American conventional SLCMs were mass-produced and extensively used in local conflicts, but eventually were integrated into the U.S. strategic doctrine and began to affect the nuclear balance as an instrument of “conventional deterrence” against nuclear opponents—namely, Russia and China.38 In this area, American superiority continued for much longer—about thirty years—but eventually Russia caught up, and by 2010 started mass production of conventional precision-guided SLCMs (the Kalibr 3M14 [SS-N-30A] and ALCMs [Kh-555 AS-22 and Kh-101 AS-23A]). Their number by 2018 increased thirty-fold, and they were effectively demonstrated in Syria after 2015.39 Current Russian military doctrine postulates: “In the context of implementing the missions of strategic deterrence by use of force the Russian Federation envisions employment of high precision weapons.”40

In the meantime, since old cruise missiles are subsonic, with long flight times and limited range, the United States initiated a program dubbed “prompt conventional global strike” to develop boost-glide weapons capable of hitting any target in the world with precision-guided conventional warheads within sixty minutes after launch.41 Supposedly, such arms were intended to counter terrorists and rogue states, but Moscow, keeping in mind the U.S. concept of strategic “conventional deterrence,” suspected that this qualitatively new American capability would also be a strategic threat to Russia. Speaking at the Valdai Discussion Club in 2015, Putin said: “A strategy already exists for a so-called first disarming strike, including with the use of long-range, high-precision non-nuclear weapons, the effect of which may be compared to that of nuclear arms.”42

In tests of boost-glide systems in 2010–2011, the United States seemed to take the lead over Russia. By 2018, however, Russia conducted a series of successful tests of the boost-glide Avangard system and commenced deployment of two missile regiments in 2019. Obviously impressed by Russian advances in cruise missiles and hypersonic systems, the Pentagon, in the 2018 Nuclear Posture Review, for the first time officially expressed concern over this threat: “Extreme circumstances could include significant non-nuclear attacks. Significant non-nuclear strategic attacks include, but are not limited to, attacks on the US, allied, or partner civilian population or infrastructure, and attacks on US or allied nuclear forces, their command and control, or warning and attack assessment capabilities.”43 It is uncertain whether Avangard gliders carry nuclear or conventional warheads, whether they can be MIRVed, whether their accuracy is sufficient for non-nuclear strikes, and whether Russia will keep its advantage in boost-glide hypersonic systems for a sustained period. But an emerging U.S. vulnerability to conventional missile attacks would represent a major strategic shift.

The Soviet Union also experienced comparable “boomerang effects” of the arms race. Its launch of the first artificial satellite Sputnik in 1957 demonstrated its primacy in space and intercontinental-missile technology. Prompted by Soviet leader Nikita Khrushchev’s bravado (“we are forging missiles like sausages” and “we will bury you”), John F. Kennedy campaigned in part on the existence of a “missile gap” favoring the Soviet Union (it turned out to be illusory), and was obliged to initiate a crash missile build-up.44 From 1961 to 1967, the United States’ strategic-missile force increased forty times over and achieved a 4:1 superiority over the Soviet Union’s force.45 Khrushchev’s desperate attempt in 1962 to curtail the United States’ growing superiority by deploying medium-range missiles to Cuba provoked a crisis in which only sheer luck and some timely political acumen saved the world from catastrophe. The crisis ended with Moscow’s withdrawal of the missiles and, while American officials had confidentially indicated to their Soviet counterparts that the United States would remove medium-range nuclear missiles from Turkey at a later date, Moscow’s international humiliation. Khrushchev’s successors invested immense resources to close a missile gap they now perceived to favor the United States and thus achieve strategic parity in the 1970s.

Another example was the development of BMD systems. The Soviet Union made an early start in 1953 and initially got ahead of the United States, achieving the first successful intercept of a medium-range missile in 1961.46 Once again, Khrushchev could not refrain from reckless boasting: “We can without missing hit a fly in the outer space.”47 But the American BMD program, started in 1958, had outpaced the Soviet one by 1963. Since the late 1960s, American BMD programs—the Safeguard system from 1969–1972, Reagan’s Strategic Defense Initiative (SDI, or “Star Wars”) after 1983, and European BMD since 2007—have plagued Moscow. As noted, the Soviet BMD program in the mid-1960s incentivized the development of the American MIRVed systems, leading to two massive and costly arms-race cycles in the 1970s and 1980s involving fivefold increases in strategic-warhead numbers and the destabilization of the nuclear balance.

Perhaps the most striking example was the deployment of the Soviet RSD-10 Pioneer (SS-20) land-based, intermediate-range ballistic missiles (IRBMs), which started in 1976. Allegedly, it was designed to replace obsolete SS-4 and SS-5 missiles and maintain regional balance vis-à-vis U.S. forward-based nuclear-attack aircraft and French and British nuclear forces. In reality, the Soviet deployment was hugely excessive, reflecting the absence of any rational civilian control over the military-industrial complex. The total number of SS-4 and SS-5 missiles was about 700, but the new ground-mobile SS-20 IRBMs were MIRVed, and constituted a radical qualitative improvement and manifold build-up of nuclear forces in terms of aggregate warheads.48 As revealed later, the planned total deployment of missiles of this type was 650, of which two-thirds were to be located in Europe and one-third in Asia. Of those, 405 (carrying 1,215 warheads in total) were already deployed by 1987.

Responding in 1979, the United States and NATO decided to bring 108 Pershing-2 IRBMs and 464 BGM-109G U.S. ground-launched cruise missiles to Europe. According to Oleg Grinevsky, a patriarch of Soviet diplomacy, the Soviet Foreign Ministry (in particular, Deputy Minister Georgy Kornienko) timidly proposed stopping or limiting the SS-20 build-up so as to preclude the U.S. deployment. But Marshal Dmitry Ustinov, the defense minister, and Marshal Nikolay Ogarkov, head of the General Staff, consolidating the hardline position of the Communist Party Politburo under Yuri Andropov, flatly refused.49 The U.S. deployment started in 1983, and led to a second dangerous crisis in superpower relations and a breakdown of arms-control negotiations in Geneva. But Moscow’s view of the situation soon changed dramatically: while Soviet missiles could not reach American territory, those of the United States could easily cover all of the Soviet Union’s European territory. Worse still, as seen from the Kremlin, Pershing-2 missiles were capable of striking targets with high-precision, ground-penetrating warheads and, in an ominously short (seven-minute) flight time, destroying hardened underground national command centers. Furthermore, ground-launched cruise missiles, with their low trajectory, could not be tracked by radars and therefore afforded almost zero warning time—and, according to Soviet military estimates, might destroy up to 65 percent of other military and civilian targets across the European part of the Soviet Union.50

The upshot was that Moscow’s attempt to redress the theatre nuclear balance with NATO turned into a major blunder that deeply undercut its security. As a matter of damage control, Mikhail Gorbachev, the new Soviet leader, in 1987 was compelled to agree to the INF Treaty, based on the principle of “double global zero.” It effectively required the elimination of 1,846 Soviet medium- and shorter-range deployed and reserve missiles—1,000 missiles more than corresponding U.S. missile cuts and covering three times as many nuclear warheads. The medium-range-missile saga of the 1980s is of particular relevance with the collapse of the INF Treaty.


  • 28Daniel Ellsberg, The Doomsday Machine: Confessions of a Nuclear War Planner (New York: Bloomsbury, 2017), 266–268.
  • 29Ibid., 270.
  • 30Fred Kaplan, The Wizards of Armageddon (New York: Simon & Schuster, 1983), 269.
  • 31Ellsberg, The Doomsday Machine, 100–104.
  • 32Thomas B. Cochran, William M. Arkin, and Milton M. Hoeing, Nuclear Weapons Databook: Vol. IV–Soviet Nuclear Weapons (New York: Harper & Row, 1989), 22–27, 42–43.
  • 33William Perry, My Journey at the Nuclear Brink (Palo Alto, Calif.: Stanford University Press, 2015), 55.
  • 34McNamara, The Essence of Security, 59–60.
  • 35K. Sivkov, “Razoruzhen i ochen’ opasen” [Disarmed and Very Dangerous], Voenno–Promyshlennyi Kur’er, no. 11, March 22–28, 2017, 1–4.
  • 36John Newhouse, War and Peace in the Nuclear Age (New York: Alfred A. Knopf, 1989), 69.
  • 37See Ronald L. Tammen, MIRV and the Arms Race (New York: Praeger, 1973), 114; and “Third Annual Report to the Congress on United States Foreign Policy,” February 9, 1972, in Public Papers of the Presidents of the United States, Richard M. Nixon, 1972 (Washington, D.C.: U.S. Government Printing Office, 1974), 307.
  • 38Robert Einhorn and Steven Pifer, Meeting U.S. Deterrence Requirements: Toward a Sustainable National Consensus (Washington, D.C.: Brookings Institution, 2017).
  • 39“Statement of the President of Russia to the Federal Assembly.”
  • 40Kremlin, “The Military Doctrine of the Russian Federation.”
  • 41See Acton, Silver Bullet?
  • 42Meeting of the Valdai International Discussion Club, President of Russia, October 22, 2015.
  • 43Nuclear Posture Review, 21.
  • 44L. Glazkova, Mozhet li povtorit’sya Karibskii krizis? [Can the Cuban Crisis Be Repeated?].
  • 45McNamara, The Essence of Security, 57.
  • 46Pavel Podvig, “The Development of Soviet and Russian Ballistic Missile Defense in the 20th Century,” in Arbatov, Dvorkin, and Bubnova, eds., Missile Defense, 33–51.
  • 47Cited by A. Kislyakov, PRO “‘Verbu,’ Mukhu, ‘Kaktus’ i ‘Krota’” [ABM, ‘Verba’, Mukha, ‘Kaktus’, and ‘Krota’], RIA Novosti.
  • 48O. Grinevskiy, Perelom. Ot Brezhneva k Gorbachevu [The Turning Point: From Brezhnev to Gorbachev] (Moscow: Olimpia, 2004), 13.
  • 49Ibid., 18–23.
  • 50Ibid., 23.

Ignoring the Lessons

Vasily Klyuchevsky, a Russian historian who lived in the nineteenth century, is supposed to have observed: “History does not teach anybody anything—it just punishes for not learning its lessons.” It looks as though the nuclear powers are on the verge of once again living up to this grim insight.

The main novelty of the current U.S. nuclear strategy and weapons programs is the concept of a limited or selective nuclear war, which originated in the 1960s with massive deployments of tactical nuclear arms in Europe and Asia. From the early 1970s, the United States promoted various options for selective and limited strategic strikes against Soviet military targets.51 In the 2018 Nuclear Posture Review, this concept once again took on a central role and was addressed to Russia:

Recent Russian statements on this evolving nuclear weapons doctrine appear to lower the threshold for Moscow’s first-use of nuclear weapons. Russia demonstrates its perception of the advantage these systems provide through numerous exercises and statements. Correcting this mistaken Russian perception is a strategic imperative . . . To address these types of challenges and preserve deterrence stability, the United States will enhance the flexibility and range of its tailored deterrence options.52

As noted, this concept would rely on the full range of sea- and air-launched nuclear and dual-purpose systems, and possibly medium-range land-based ones, though the United States has not as yet indicated that it will deploy INF-covered land-based systems in Europe.53

As for Russia, it played with this idea in 2003, when an official Ministry of Defense document announced plans for the “de-escalation of aggression . . . [by] the threat to deliver or by the actual delivery of strikes of various intensity using conventional and (or) nuclear weapons.” Thus, the document assumed the possibility of “dosed combat employment of selected components of the Strategic Deterrence Force.”54 Current Russian military doctrine and other official documents make no mention of such concepts, but they have been frequently discussed in professional military circles, including those associated with governmental institutions, which stressed “the limited nature of a first nuclear strike, which is designed not to harden, but rather to sober up an aggressor, to force it to halt its attack and move to negotiations.”55

In an address to the Russian Federal Assembly on March 1, 2018, Putin said: “Any use of nuclear weapons against Russia or its allies, weapons of small, medium or any yield at all, will be considered as a nuclear attack on this country. Retaliation will be immediate, with all the attendant consequences.”56 This statement does not appear to countenance the concept of limited nuclear response, though it also does not negate it. Russian military doctrine postulates: “The Russian Federation shall reserve the right to use nuclear weapons in response to the use of nuclear and other types of weapons of mass destruction against it and/or its allies, as well as in the event of aggression against the Russian Federation with the use of conventional weapons when the very existence of the state is in jeopardy.” The purpose of a nuclear strike is defined as “the infliction of the unacceptable level of damage on an aggressor in any conditions.”57 These formulations too do not embrace the notion of limited nuclear war, but do not exclude them either. It is not clear when and how exactly the “existence of the state” can be considered in jeopardy, and what “level of damage” to the enemy might be interpreted as sufficient.

Moscow has often followed the U.S. example by adapting its strategy and doctrine to fit its technology. No matter how much the deterrence doctrine is used to justify supposedly limited nuclear capabilities, they actually lower the nuclear threshold and increase the likelihood of any armed clash between the superpowers escalating into a nuclear conflict with a subsequent exchange of mass nuclear strikes. Having retained more sub-strategic nuclear arms than the rest of the world combined, Russia could be shifting its emphasis to conventional or dual-purpose systems.58 Nevertheless, if the United States is really concerned about neutralizing Moscow’s suspected concept of limited nuclear use, the best way to do so would be to flatly deny such a possibility instead of responding in kind. Still better would be a joint U.S.–Russia declaration excluding any nuclear first strike or first use, as voiced in the 1970s and 1980s with respect to “winning and fighting nuclear war,” especially if it were substantiated by a follow-on to START and a radical reduction in sub-strategic nuclear forces.

The same boomerang dialectics may arise with advanced hypersonic weapon systems. Russia’s program has been justified by the need to penetrate the American BMD system on the U.S. continent, in Europe, in Asia, and on surface ships. Putin declared the last successful test of the boost-glide Avangard in December 2018 as “a New Year’s present to the country” and even compared it with the Sputnik launch of 1957.59 Describing its unique qualities, he said: “It flies to its target like a meteorite, as a burning ball, fireball . . . As you understand nobody in the world has anything comparable . . . Sometime probably there will be, but in the meantime our guys will invent something else.”60 On cue, the United States has accelerated its hypersonic-development program.61 The future strategic importance of the new weapon systems remains uncertain. It will be defined by their cost and scale of deployment, accuracy and class of warhead (nuclear or conventional), resistance of command-and-control and navigation assets to countermeasures, and the availability of opposing tracking and intercept systems. From a strategic perspective, such a system might be needed if the United States could create a BMD system capable of defending against 1,500 Russian ballistic missiles’ nuclear warheads, or at least a few hundred of those surviving a counterforce strike. But this is impossible in the foreseeable future, and the expansion of U.S. BMD, envisioned by the ballistic-missile-defense review of 2019, does not imply anything like SDI’s notional capabilities.62 (In fact, the Soviet Union initiated development of a nuclear boost-glide system called Albatross in the mid-1980s as a countermeasure to SDI.) Hence, Avangard, like a number of other advanced arms programs that Putin announced in 2018, may look exciting to Russia as a technological achievement, but is obviously excessive as a response to the United States’ BMD systems. If deployed at limited scale, hypersonic arms will not tangibly affect the strategic balance. But if both sides were to deploy them in large numbers, with nuclear or highly accurate conventional warheads, they could disrupt Moscow’s nuclear deterrence strategy and Russia’s national security.

At the Valdai Discussion Club in Sochi in October 2018, Putin formulated the main concept of the Russian nuclear doctrine:

Our concept is based on a launch-on-warning strike. . . . This means that we are prepared and will use nuclear weapons only when we know for certain that some potential aggressor is attacking Russia, our territory . . . A missile attack early warning system . . . monitors the globe, warning about the launch of any strategic missile . . . and identifying the area from which it was launched. Second, the system tracks the trajectory of a missile flight. Third, it locates a nuclear warhead impact zone. Only when we know for certain—and this takes a few seconds to understand—that Russia is being attacked we will deliver a retaliatory strike.63

This launch-on-warning concept is extremely controversial, leaving supreme national command authority only a few minutes for a decision, which may be triggered by a technical mishap, strategic miscalculation, or psychological stress. Some fifty years ago, Herbert York warned about “a state of affairs in which the determination of whether or not doomsday has arrived will be made either by an automatic device . . . or by a pre-programmed President who, whether he knows it or not, will be carrying out orders written years before by some operations analyst.”64

Hypersonic systems are prone to making the situation still more dangerous. Launched to fly at an altitude of 50–60 kilometers, their trajectory goes largely under the BMD radars’ beams with broadly changing azimuths, which makes their flight path unpredictable and precludes interception at a pre-programmed rendezvous point. Moscow emphasizes this very characteristic in its BMD penetration strategy. At the same time, however, the characteristic precludes confirmation of a missile attack by tracking radars after the launch of hypersonic boosters is detected by early-warning satellites 60–90 seconds after start. As long as there are no space-based infrared systems for tracking hypersonic gliders, after their booster’s launch is detected by satellites, the next time a hypersonic glider will be seen is three to four minutes before impact, which does not leave time for authorization of a launch-on-warning strike.65 While the air-defense challenges presented by hypersonic systems may be addressable through the deployment of different sensors and other technical innovations, this remedy would take time to develop and its feasibility remains uncertain.

If the United States and Russia broadly introduce hypersonic arms, both nations will face this problem. But, according to Putin, launch on warning amounts to Russia’s main deterrence concept. About half of its strategic warheads are deployed on silo-based ICBMs (including the forthcoming Sarmat heavy missiles and Avangard boosters). They are the primary weapon systems for launch on warning due to both their vulnerability to counterforce strike and their high (“hair-trigger”) launch readiness. For the United States, the concept is secondary since only a quarter of its force (by actual loading) is deployed on silo-based ICBMs. Thus, Moscow, having initiated the hypersonic arms race, may in the future face the threat of a disarming strategic strike and would have to consider several fraught options. One would be to sustain “the infliction of the unacceptable level of damage on an aggressor in any conditions,” envisioned by the current military doctrine, without launch on warning.66 This would imply mammoth costs in relocating the strategic force in sufficient numbers to highly survivable ground-mobile, sea- and air-basing modes, along with their command-and-control complexes.

Another option might be to retain the launch-on-warning concept, under which retaliation should be authorized upon receiving information from early-warning satellites. This would mean neglecting the history of satellites’ false alarms over the course of their decades of service. In addition, the reliability of space systems could become compromised by growing anti-satellite capability or cyber warfare.

The third option would be to reduce “the unacceptable level of damage on an aggressor” and rely primarily on ground-mobile and sea-based systems, while gradually phasing out silo-based ICBMs. This move would save a lot of money and might be facilitated by lowering the overall force numbers under the follow-on START. This would be in line with the rational strategic program elaborated in 1998 by a Russian version of a “blue ribbon” military-civilian panel commissioned by then-Minister of Defense Marshal Igor Sergeyev and headed by Nikolay Laverov, vice-president of the Russian Academy of Sciences.67 Rational considerations would seem to dictate this option, but times have dramatically changed since 1998.

Given the parlous state of U.S.–Russia relations, the arms-control crisis, and the nature and ideology of Moscow’s decision-making system, the first or second alternative, or some combination of the two, seems more likely in the foreseeable future if hypersonic systems become a key element of the arms race.


Reflections on Disarmament

One lesson from the last half-century of arms control is that shifts in the military balance make the sides periodically alternate their stances on the limitation or prohibition of certain weapon systems. Arms-control negotiators have frequently joked that Moscow and Washington have the same positions on all arms-control issues, just at different times. An impor­tant moment in strategic arms control occurred at a June 1967 meeting in Glassboro, New Jersey, between American President Lyndon Johnson and Soviet Prime Minister Alexei Kosygin. McNamara urged Kosygin to appreciate the destabilizing effect of ABM systems. Kosygin categorically rejected this viewpoint, indignantly asserting: “Defense is moral, offense is immoral.”68 By that time, the Soviet Union had decided to deploy the Moscow area A-35 Galosh missile-defense system, while McNamara was dragging his feet on the proposed U.S. Nike-X BMD. In just two years, Moscow would embrace McNamara’s philosophy, regarding BMD as destabilizing and sticking to this position for the ensuing half-century, while Washington would adopt Kosygin’s position, during the 1980s and thereafter.

The reason for the flip is obvious: each party is trying to limit arms in which the opponent is superior and maximize its own military advantages. But in the course of the arms race, the sides regularly match or overtake each other and correspondingly change arms-control priorities. For example, Russia has for many years emphasized the threat of U.S. precision-guided, long-range conventional systems, portrayed by Putin as weapons of “the first global disarming strike.”69 After Russia recently built up its conventional cruise-missile capability and achieved a breakthrough in hypersonic systems, this threat all but disappeared from the Russian list of strategic concerns. Likewise, Russian deployment of the new-generation ground-mobile conventional A-235 Nudol and S-500 BMD systems may change its attitude toward missile defense. One conclusion is that it is not worthwhile to ideologically demonize the other side’s advantages in arms programs or differences in negotiating positions. These asymmetries regularly alternate and require clear-headed professional assessments rather than shrilly politicized pronouncements on various “gaps.” Another important lesson is that arms-control treaties, even if concluded in a tense international environment, have usually enhanced mutual security and facilitated détente. The ABM Treaty and SALT I agreement of 1972 were concluded despite the opposition of Soviet hardliners in the Politburo soon after the escalation of the U.S. bombing of Vietnam and its mining of Haiphong Harbor, which damaged Soviet ships. These agreements stimulated broader progress in nuclear-arms limitation, reduction, and elimination, enhancing international security, improving U.S.–Soviet relations, and helping to end the Vietnam War.

Conversely, the breakdown of arms-control negotiations or refusal to ratify agreements has always damaged security and never helped resolve other international problems. Washington’s rejection of SALT II ratification due to the Soviet invasion of Afghanistan in 1979 hindered strategic arms control, and in no way facilitated peace in Afghanistan or great-power cooperation on international security. Likewise, the Russian political elite’s indignation over NATO expansion and the use of force in Yugoslavia prevented timely ratification of START II and the conclusion of the treaty on the basis of the START III Framework Agreement of 1997. This was counterproductive for arms control and did not alleviate mutual grievances and mistrust in NATO–Russia relations. Finally, Moscow’s refusal to start negotiations on the START follow-on after 2012 contributed to the ensuing crisis of arms control and to new strategic tensions between Russia and the West.

Possibly the most important lesson from the history of arms control is that it is very difficult to build disarmament agreements, but quite easy to destroy them. Doing the latter has never enhanced national or international security and has invariably compromised it. For instance, the United States denounced the ABM Treaty in 2002, citing the missile threat of rogue states. Eighteen years later, the United States has 44 strategic ground-based, mid-course defense-system interceptors in Alaska and California, increasing to 64 by 2023.70 Yet, under the 1974 protocol to the ABM Treaty of 1972, each side was permitted 100 interceptor missiles, which the United States could base in North Dakota. The treaty did not envision any restrictions on technical characteristics of interceptors (that is, as to range, guidance system, or warhead type), while the location, if necessary, could be easily renegotiated as an amendment to the treaty. The U.S. Standard-3 Aegis-type interceptors in Europe and Asia or on surface ships, for use against medium-range ballistic missiles, could come under the documents included in the 1997 agreement on the delineation of strategic and theatre missile-defense systems.71 So the ABM Treaty could easily have been preserved with light amendments that would have permitted the United States to do everything it has done since 2002, or is planning to do in the foreseeable future.

U.S. withdrawal from the ABM Treaty did not alleviate overall strategic tensions, and in fact made them worse. Missiles and missile technologies have proliferated. North Korea withdrew from the NPT in 2003, started nuclear tests in 2006, and had been testing missiles of ever-growing range up to 2018. Iran agreed to curtail its nuclear program in 2015 not because of U.S. BMD development but for unrelated reasons and continues to develop and test missiles. After New START in 2010, U.S.–Russia strategic negotiations stopped, the main objection on Moscow’s side being the absence of the ABM Treaty and cooperative development of defense systems by the two nations. In 2018, Russia unveiled a package of new offensive programs to counter the U.S. BMD, which is seen in Moscow as an open-ended program. China is emulating Russia on this score.

Another example is Russia’s “suspension” of its participation in the Treaty on Conventional Armed Forces in Europe (CFE) in 2007 and “final suspension” in 2015. Initially, Moscow justified these steps as a means of applying pressure on NATO to ratify the 1999 CFE Adaptation Agreement. But in 2011, NATO states responded by also ceasing their adherence to the terms of the treaty. Presently, there is no functioning conventional-forces-limitation regime in Europe. Russia has been building up its forces in its western and southern military districts, as in Crimea, South Ossetia, and Abkhazia. On the other side of their borders, NATO has undertaken defensive deployments in the Baltic states, Poland, and Romania, and U.S. military units and heavy arms and equipment are returning to the continent. Substantial NATO superiority over Russia in all military and economic dimensions, alongside American logistics and power-projection capabilities, make the prospects for Russian security along its western borders quite precarious. Moscow would probably have felt more comfortable if NATO forces in Eastern Europe had been tangibly limited by CFE national and territorial quotas, and open to confidence-building and transparency regimes.

Still greater near-term threats may emerge after the collapse of the INF Treaty and eventual expiration of START without a follow-up treaty. The loss of their stabilizing effects cannot be offset by any medium-range or strategic-weapons program on either side. Possible deployment of new U.S. medium-range missiles in Europe and Asia occasioned by the end of the INF Treaty would, due to their short flight time and low trajectory, render a Russian deterrent based on launch on warning unconvincing, as there would be no time for its implementation. According to a statement by one respected Russian military commander, this might force Russia to move to the highly risky concept of a preemptive nuclear strike.72 If the United States were to follow suit by adopting a similar concept, crisis stability would be practically impossible to maintain.

The revolutionary impact of military technological progress is not something new, but quite a logical and historically regular phenomenon. However, the progress of U.S.–Soviet/Russian arms control talks during the last fifty years, despite some setbacks and exemptions, did manage to impose deep and stabilizing reductions on their medium range and strategic forces.

Dealing with new threats to strategic stability in the follow-on START treaty requires that the long-range (i.e., more than 600 km)73 air-launched nuclear and conventional cruise and hypersonic missiles and nuclear gravity bombs are included under a common warhead ceiling, and that they be counted according to the actual loading of the heavy bombers. In the past, air-launched missiles were counted under warheads ceilings in the 1991 START I and in the 1993 START II treaties.74 Limits on strategic delivery vehicles and warheads should also cap the innovative weapon systems: ground-based intercontinental cruise missiles and long-range autonomous underwater drones, as well as land- and sea-based boost-glide hypersonic systems with ranges defined similar to what was in the SALT and START treaties (e.g., land-based missiles with ranges greater than 5,500 km and sea-based missiles with ranges greater than 600 km).75 Such weapons should be limited regardless of whether their warheads are nuclear or conventional. Most of these new systems can be verified using the methods and means of START/INF regimes.

In this way, the most destabilizing long-range strategic systems, which are blurring a clear line between conventional and nuclear warfare, would become subject to verifiable arms control (including conventional missiles and low-yield nuclear bombs). Indirectly, their numbers would be limited, since under common ceilings they would “compete” with the number of proven and reliable nuclear-tipped strategic ballistic missiles. The latter would also have to be reduced to allow for ground- and air-launched cruise missiles, hypersonic boost-glide and ram-jet missiles, and underwater nuclear drones under the overall limit. In fact, even under New START the above change in counting rules would require at least a 30 percent cut in land- and sea-based ballistic missiles’ warheads—no less than proposed by President Obama in 2016.

The proposed model of START follow-on would not address a number of potentially destabilizing weapon systems and technologies: anti-missile defense, space arms, cyber-warfare, directed-energy weapons, tactical nuclear weapons, and a great variety of drones with artificial intelligence. Those systems and technologies cannot be addressed immediately, either technically or diplomatically. However, it does not mean that there is no sense in addressing weapons and technologies that may be immediately managed by arms control under the follow-on START treaty for the sake of salvaging strategic stability. Eventually the exotic weapons might be taken care of by future negotiations, provided that the first steps outlined above are urgently taken to prevent the final collapse of the arms control regimes.

The roots of the present crisis of arms control are not in the technical complexity of the current strategic relationships, as intricate as they are, nor in the turmoil in the world order, as chaotic as it is. The core of the problem is rather the distinct failure on the part of the new generation of political elites on both sides to appreciate the high strategic importance and priority of arms control. Indeed, the current state leaders and defense and foreign policy cadres came to positions of influence in the beginning of the new century (or even later) and inherited “for free” the legacy of the arms control system built during the preceding decades. Hence, they have been taking it for granted and treating it as a pawn in the game of foreign and domestic politics. They have a very vague idea of the world without such a system and do not know (or believe in myths) about dangerous crises and wasteful cycles of the nuclear arms race of the Cold War times.

There is no certainty that they would accept the solutions presented above or any other reasonable and practical proposals for getting out of the deepening dead ends. However, it is absolutely certain that continuation of the present course of actions of the major states is leading the world to an uncontrolled, multifaceted, and multilateral arms race and eventually to catastrophe. In the last few years there has been a proliferation of well-intended studies on various substitutes for formal arms control in the absence of the INF Treaty or START.76 All of the options are considerably less effective than existing arms-control treaties with respect to preserving strategic stability and predictability, and managing the arms race. Furthermore, if the present political elites of leading nations lack the will or knowledge to sustain formal arms control, they are still less likely to manage the strategic environment by dubious surrogates. While thinking about a bleak future for arms control and entertaining other purportedly tolerable arrangements may be intellectually exciting, that vocation could turn counterproductive. Politically, it would service the illusion that living without formal arms control might not be so bad and that the damage from its disintegration could be limited. Instead, they should be providing politicians with a realistic picture of the myriad future dangers of a world without arms control.

The problem of saving the effect of the INF Treaty could be quickly fixed by agreeing on a moratorium on the deployment of the intermediate-range missiles in Europe with short-notice, and on-site inspections at Russian Iskander/Novator (9M729) missile bases and U.S. Aegis Ashore bases in Romania and Poland in order to remove mutual suspicions.77 Negotiating a START follow-on would be more challenging, but possible during the next five years if there were firm political directives from the Kremlin and the White House. After all, New START was negotiated in just one year.

McNamara finished his luminous San Francisco speech with these words: “In the end, the root of man’s security does not lie in his weaponry, it lies in his mind. What the world requires in its third decade of the Atomic Age is not a new race towards armament, but a new race towards reasonableness. We had all better run that race.”78 Those words have never been as relevant as they are now—in the eighth decade of the Atomic Age.


  • 68Newhouse, War and Peace in the Nuclear Age, 205.
  • 69Transcript of the meeting of the Valdai Discussion Club, Sochi, October 24, 2014.
  • 70“Missile Defense Review.”
  • 71This agreement permitted tests of BMD interceptors against missile targets with a speed of no more than 5 km/sec and range of 3,500 km. For the future, it was permitted to develop land- and air-based interceptors with a speed of up to 5.5 km/sec, and sea-based interceptors of up to 4.5 km/sec. Such BMD subsystems were exempted from the limitations of the ABM Treaty.
  • 72In an interview, Colonel General Viktor Esin, former chief of staff of the Russian Federation Strategic Missile Forces, said: “If the Americans begin to deploy their missiles in Europe, we will have no choice but to abandon the doctrine of launch-on-warning and move to a doctrine of preemptive strike.” See “Interview with Colonel General Viktor Esin,” Zvezda Weekly, November 8, 2018.
  • 73The 600 km range was set to define strategic ALCMs and heavy bombers equipped with such weapons for SALT II (1979) and START I treaties.
  • 74Only the 2010 New START Treaty adopted liberal count rules for each bomber: one delivery vehicle = one warhead, although in reality, it could carry up to 20 missiles.
  • 75Such criteria were set in the SALT I (1979) and START I (1991) treaties.
  • 76Sergey Karaganov, “On the New Nuclear World: How to Strengthen Deterrence and Maintain Peace,” Russia in Global Politics 15 (2) (March–April 2017); Andrey Kortunov, “The End of the Bilateral Era: How the US Withdrawal from the INF Treaty Changes the World Order,” Carnegie Moscow Center, October 23, 2018; and Vincent Manzo, “Nuclear Arms Control Without a Treaty? Risks and Options After the New START: CNA’s Strategy, Policy, Plans, and Programs Division (SP3),” Deterrence and Arms Control Paper no. 1, April 2019.
  • 77On October 26, 2020, about a year and a half after this essay was published in the journal Survival, President Putin made the same proposal to NATO: on-site inspections in the Kaliningrad region of Russia to verify the absence of 9M729 missiles’ deployment, and in Poland and Romania to make sure that Tomahawk missiles are not installed in Aegis Ashore BMD launchers. In addition, Putin promised to refrain from deployment of 9M729 missiles in the European part of Russia, conditional on the nondeployment of U.S. INF-types of missiles in Europe; see If this had been proposed in June 2019, it would have been more difficult for the United States to abrogate the INF Treaty in August. “Too little, too late” is a plague of arms control.
  • 78McNamara, The Essence of Security, 67.