Rethinking the US strategic triad: When it comes to nuclear platforms, how many are enough?

By Stephen J. Cimbala, Lawrence J. Korb | December 20, 2023

Artist's concept shows future ICBM blasting into skyArtist's concept of the next generation of US land-based intercontinental ballistic missile, initially known as the Ground Based Strategic Deterrent and now dubbed Sentinel. Credit: Northrop Grumman

US armed forces have undergone a virtual revolution in military technology over the past several decades, along with a rethinking of a number of concepts about military strategy. The conceptualization of the US nuclear strategic military force structure, however, is still very much on autopilot. The traditional “triad” of strategic launch systems—intercontinental ballistic missiles (ICBMs), submarine-launched ballistic missiles (SLBMs), and heavy bombers—dates from the 1960s. The Obama, Trump, and Biden administrations all endorsed modernization plans for the triad that replaced each category of launchers with an updated version of its predecessor.[1] As a recent report by the Congressional Commission on the Strategic Posture of this nation pointed out, the United States is still on course to spend more than a trillion dollars rebooting the strategic nuclear triad and supporting command, control, and communications systems and infrastructure.

By all accounts, it seems clear that political and bureaucratic inertia have as much to do with nuclear modernization as does any master strategy for nuclear deterrence. The assumption behind the present triad of nuclear delivery systems is that a three-sided force structure of land-based, sea-based, and airborne launch systems has a synergistic complexity and diversity that complicates the calculations of any prospective nuclear adversary. And to some extent, this is so. But present modernization plans for the US strategic nuclear triad may not have sufficiently interrogated the range of available options made possible by new technology, or adequately considered whether a more streamlined and less expensive force structure would suffice for deterrence.

For example, the present American ICBM force is based in silos that are potentially first-strike vulnerable, unless the missiles are launched on warning of attack. The growing interest of nuclear weapons states in hypersonic attack weapons only adds to the time pressures on those who must make a launch decision. Unless more effective antimissile defenses can be developed, fast-flying attackers equipped with defense-avoidance technology can plausibly threaten prompt destruction of most or all US land-based missiles. Under these conditions, the United States effectively operates with a nuclear dyad instead of a triad, unless it makes the fateful decision to become more reliant on nuclear preemption in a crisis, even if a warning of attack is ambiguous.

Some critics have called in the past for elimination of the US land-based missile force on account of its first-strike vulnerability and potential to trigger a mistaken preemption that could lead to global nuclear war. Some have responded by arguing that elimination of the ICBM force would make it possible for Russia to destroy the remaining US strategic nuclear deterrent using only a fraction of its deployed nuclear strategic launchers, imposing a counterforce defeat that would forestall any American decision to retaliate.

This concern is understandable but misplaced.

Even if the United States eliminated its ICBM force, Russia would still have to contend with the remaining American ballistic missile submarine and bomber forces, which are sufficient to impose unacceptable retaliation against Russian military forces and society. Pessimistic analyses of US survivability without ICBMs emphasize that the numbers of American bomber bases and submarine pens and supporting infrastructure and command-control systems would in theory require only tens of Russian warheads to disable. But if so, that would only be the case for a “bolt from the blue” attack not preceded by any nuclear crisis, a most unlikely contingency. More likely is the existence of a preceding political crisis in which both states have already alerted and mobilized delivery systems and nuclear command, control, and communications systems above normal peacetime levels.

Moreover, the ICBM force could be adapted in four ways to fit the needs of 21st century deterrence.[2] First, some or all of the ICBM force could be based on mobile platforms, either road or rail. Second, ICBMs could be denuclearized and turned into a strategic conventional prompt global strike system. Third, the ICBM force could be supported by strategic antimissile defenses that are collocated with silos (so-called terminal or simple-novel defenses) or by more advanced defense systems of the future offering “left of launch” intercepts in the boost phase or even prior to launch. Fourth, the United States could decide to get by with a nuclear strategic dyad of submarine-launched ballistic missiles and nuclear-capable bombers.

In the first option, the United States could reduce the total number of ICBMs from 400 to 300, with 150 in a mobile basing mode, possibly with two warheads each. The gain in survivability would more than compensate for the fewer numbers of launchers compared to present plans.

Another reason to cancel the Sentinel missile: the rising cost of its nuclear warhead

In the second option, the entire ICBM force would become a prompt global strike force, but this option could also be combined with option one. In that case, the l50 silo-based ICBM launchers would carry conventional weapons instead of nuclear charges. This would provide policy makers with the capability for selective prompt intercontinental and transoceanic non-nuclear strikes without necessarily escalating into a massive nuclear war. One problem with this approach is that Russia might believe that any ICBM headed its way would be assumed automatically to carry nuclear weapons and so would be responded to. However, the United States would probably not be attacking Russia with conventional long range ballistic missiles, and these missiles could be used for a variety of strikes against other hard targets outside of Russia.[3]

A third option for modernizing the ICBM force would defend silo-based ICBMs with so-called simple-novel endo-atmospheric defenses designed to intercept attacking warheads in their terminal stage of flight. Assuming that New START or similar arms control agreements remain in place with respect to US, Russian and possibly Chinese strategic nuclear warheads and launchers, terminal defenses would not have to be perfect or near perfect to be effective.  For example, with New START limits in place, US defenses would need to raise the “attack price” for destroying a silo-based ICBM from 2-3 warheads per silo to 5-6 warheads. Above that level, fratricide among attacker warheads targeted on the same silo could confound attacker predictions about the number of ICBMs destroyed.[4] Of course, given the assumption of an unlimited arms race, terminal defenses would be easily overwhelmed, even if supplemented by a midcourse layer of ballistic missile defenses like the currently deployed ground-based midcourse defense system. Only defenses based on new physical principles can change this equation, and some research and development toward “left of launch” intercept prior to or during the earliest stages of ICBM launch, based on combinations of directed energy weapons, electronic warfare, and unpiloted autonomous vehicles is already in progress.

This brings us to the fourth option for ICBM modernization: elimination of the ICBM force and relying on a strategic nuclear dyad instead of a triad. This would in theory remove or greatly reduce the problem of survivability from the US strategic nuclear deterrent, provided that sufficient intelligence and warning for submarine-based and bomber forces were available. Under this assumption, US planners could assign additional numbers of weapons to ballistic missile submarines and bombers to maintain parity with rivals. They might also consider modification of additional ballistic missile submarines, so some of their launchers were delegated to launching cruise instead of ballistic missiles, armed either with conventional or nuclear weapons. In a variant of this option, instead of putting all of the sea-based strategic weapons on Columbia-class nuclear missile submarines, the United States might deploy purpose-built cruise missile firing submarines that were less expensive and also equipped with conventional as well as nuclear options. Under either option, the sea based strategic forces would then resemble the US heavy bomber B-21 force, capable of delivering conventional as well as nuclear strikes on assigned targets.

Related to this fourth option is the alternative of deep underground basing for the ICBM force. Given the inertia of American domestic politics and strategic policy planning, attempting to do away with the ICBM force is a daunting challenge.  On the other hand, some or all of the strategic land-based missile force could be based underground in tunnels accessible by postattack delivery systems and command-control networks. Deeply buried ICBMs would not require active defenses or launch on warning for survivability. They would have to be able to survive a first strike together with their assigned personnel and equipment, remain connected to military and political leadership, and be capable of timely movement to positions for retaliation once authorized to do so.  Some environmental issues would be involved in creating infrastructure for deep underground basing, and critics would be quick to note the “mineshaft gap” and its giggle factor as it appeared in the dark comedy Dr. Strangelove. On the other hand, deep underground basing ensures that no conceivable breakthrough in anti-submarine warfare could reduce the survivability of the US ballistic missile submarine force and leave the United States without a robust response to nuclear attack.

But any analysis must also ask, regardless of the US plan for nuclear modernization, how much is enough?[5] How many warheads deployed on intercontinental and transoceanic launchers are necessary for deterrence, arms control, and crisis stability? There is no “school solution” to this question, since it depends in part upon the decisions made by Russia, China, and possibly other nuclear weapons states.  It also depends upon assumptions about what is necessary for deterring, or for fighting and surviving, a nuclear war.

Congress will hold a hearing about the Sentinel missile's exploding budget, but is it too little, too late?

Since the 1960s, US nuclear war plans have never been based on a one-variant “wargasm” model of nuclear response to attack. Instead, policy makers and military planners have demanded options for responding selectively with strikes against enemy military forces and other war related targets and, to the extent possible, avoiding unnecessary damage to civilian populations and society. On the other hand, strategic war plans have not infrequently been driven by the numbers of weapons available for targeting and by models of nuclear damage that may have underestimated the effects of nuclear war.[6]

Regardless of the assumptions built into the models of policy makers and strategic theorists, history provides many examples of generals and politicians who embarked on war with unrealistic expectations of short wars, favorable outcomes, and minimal costs.  Pushing the nuclear button takes the world into terra incognito, regardless of the forces deployed and the plans developed beforehand.

This discussion unavoidably assumes a more or less linear evolution in science and technology over the next half decade or so. But drastic changes in technology and thinking may call for a revised prescription for nuclear modernization. It is now accepted military wisdom that the future domains of military deterrence and conflict will be land, sea, air, space, and cyber—with cyber wrapped around and integrated with all the others. In addition, we can anticipate a military revolution with respect to advanced technologies for artificial intelligence, advanced conventional weapons based on new principles, autonomous vehicles (including drones and UAVs for all military services), and for want of better terminology, brain warfare or mindwar.

Arguably, nuclear deterrence has always been mostly about mindwar, albeit if deterrence should fail, then mindwar mixed with heavy metal and mass destruction. The containment of nuclear weapons spread will be a major challenge but a necessary requirement for strategic stability. Global nuclear disarmament may not be achievable by deliberate political agreement among nuclear weapons states and others. But it is not inconceivable that better conventional weapons, including non-nuclear antimissile defenses, may become more competitive with prospective weapons for attack. Indeed, there is already in progress a competition between new technologies for hypersonic attack weapons and those for preclusive antimissile and air defenses. By mid-century, nuclear weapons could remain deployed in the arsenals of major powers but without their status as mainstays of deterrence, passing that baton to advanced non-nuclear weapons offering deterrence by denial preferentially to deterrence by threat of regional or global annihilation.


[1] Congressional Commission on the Strategic Posture of the United States, America’s Strategic Posture(Washington, D.C.: October, 2023),

[2] See Lauren Caston, et. al., The Future of the U.S. Intercontinental Ballistic Missile Force (Santa Monica, Calif: RAND Corporation, 2014),, for a related assessment of options for ICBM basing.  See also: Steven A. Pomeroy, An Untaken Road: Strategy, Technology, and the Hidden History of America’s Mobile ICBMs (Annapolis, Md.: Naval Institute Press, 2016).  Earlier studies included: Office of Technology Assessment, MX Missile Basing (Washington, D.C. U.S. Government Printing Office, September 1981, NTIS order #PB 82-108077.

[3] One issue is whether a sufficient number of appropriate targets exist for long range conventional PGS that could not be attacked in other ways. Another problem for mixing conventional and nuclear warheads on similar launchers is the verification of arms control agreements—although a rule of thumb could be used to count any weapons deployed on intercontinental launchers as nuclear for arms control purposes.

[4] Assessment methodologies are discussed in: Stephen Weiner, “Systems and Technology,” Ch. 3 and Ashton B. Carter, “BMD Applications: Performance and Limitations,” Ch. 4, in Ashton B. Carter and David N. Schwartz, eds. Ballistic Missile Defense (Washington, D.C.: Brookings Institution, 1984).

[5] An interesting discussion, although mostly focused on less than strategic nuclear weapons, appears in: Edward Geist, Qualities Precede Quantities: Deciding How Much Is Enough for U.S. Nuclear Forces (Santa Monica, Calif.: RAND Corporation, September 2023)

[6] Fred Kaplan, The Bomb: Presidents, Generals, and the Secret History of Nuclear War (New York: Simon and Schuster, 2020).

Together, we make the world safer.

The Bulletin elevates expert voices above the noise. But as an independent nonprofit organization, our operations depend on the support of readers like you. Help us continue to deliver quality journalism that holds leaders accountable. Your support of our work at any level is important. In return, we promise our coverage will be understandable, influential, vigilant, solution-oriented, and fair-minded. Together we can make a difference.

Get alerts about this thread
Notify of
Inline Feedbacks
View all comments


Receive Email