Nuclear Anniversary

Today--August 6, 2025--marks 80 years since the first nuclear weapon was used in anger, and August 9 will mark 80 years since the last one was deployed against an enemy (I expect and hope). Although the destruction of Hiroshima and Nagasaki at the end of World War II is tragic, the fact that we as a species have kept a nuclear taboo is worth celebrating.

We'll meet again; don't know where, don't know when...

Ever since the nuclear club expanded to two countries with the Soviets' first test in 1949, the logic of "mutually-assured destruction" has been the primary limiter to the use of nuclear weapons. No nuclear-armed country could risk using their nuclear weapons against another nuclear-armed country without the high risk (or "assurance") that they would be destroyed too by the inevitable counterattack. Nuclear attacks against third parties were discouraged by treaties and alliances (e.g. the "nuclear umbrella" the US extended to its allies) and general fears of escalation - that using nuclear weapons anywhere risked further use until it rose to the level of a complete nuclear exchange (destrction again). That taboo now seems more fragile than it had for decades: arms-reduction treaties are expiring and swaggering heads of government around the world are more willing to talk publicly about using those weapons.

This post will cover a few nuclear topics: the "assurance" of destruction (i.e. deterrence) under various stressors, the types of destruction (counterforce versus countervalue), and different types of nuclear weapons and the risk of escalation (e.g. "tactical" versus "strategic" weapons). I'm writing from the perspective of the United States, although this logic applies to any nuclear power.

Assured Destruction

The logic of mutually-assured destruction creates deterrence: a nuclear-armed state fears to use its weapons because the risk of destruction is too high. To induce that fear, the United States must be able to deliver its nuclear weapons to sensitive targets in its enemy's territory with minimal notice. This often means that some portion of the arsenal must be survivable or second-strike capable: even after a sucessful nuclear "first strike" by an enemy, the US must retain sufficient weapons to destroy that enemy. Without second-strike capability, the US must "launch on warning": when it receives the warning that it's under attack, it must launch its arsenal before the first enemy missiles land. This drastically increases the chances of a misunderstanding or error leading to a nuclear war, so if everyone involved can be confident that the second-strike weapons will survive in sufficient numbers, the situation is much more stable.

As technology has developed, the US has come to rely on a "triad" of nuclear delivery systems for large ("strategic") nuclear warheads:

1. Land-based bombers: currently the B-52 and B-2, soon to be joined by the B-21)
2. Land-based missiles: LGM-30 Minuteman III intercontinental ballistic missiles (ICBMs), to be replaced with the LGM-35 Sentinel
3. Sea-based missiles: nuclear-missile submarines (SSBNs, currently the Ohio class and eventually the Columbia class) armed with the UGM-133 Trident II missile

The strategic bombers have been a delivery system ever since the B-29 "Enola Gay" dropped the first weapon on Hiroshima. B-52s are now primarily armed with air-launched cruise missiles to improve their chances of getting through enemy defenses, and the AGM-181 "long-range standoff weapon" is intended for use by both the B-52 and its low-observable comrades the B-2 and B-21. Previously the "stealth" bombers had been limited to gravity bombs which require them to drop directly over their targets. Bombers are the most flexible of the various nuclear-employment options, are useful in non-nuclear roles, and often can be recalled until very late in their attack - which is extremely helpful in a false-positive situation or the adversaries decide to stop escalating partway through a nuclear exchange. During the Cold War, the US maintained some bombers on permanent airborne alert, where some bombers would always be in the air and therefore could not be destroyed before they took off during a nuclear first strike. After the fall of the Soviet Union, this was discontinued due to its high cost.

The land-based missiles are the most vulnerable to a first strike in that they are fixed in known locations. That is part of the reason why they are sited in remote areas and heavily armored underground. However, they are extremely difficult to stop and reach their targets much faster than bombers, so they provide the bulk of the US nuclear deterrent. An adversary would have to target each silo, probably with multiple warheads to ensure that the silo cannot launch, to succeed in a first strike. Still, the fixed airfields for bombers and silo locations for ICBMs are why the US has used the "launch on warning" posture in the past.

Submarine-based missiles provide unique first-strike and second-strike capabilities. Because they can maneuever close to their targets and fire on depressed trajectories, they could deliver nuclear missiles with negligible warning to an adversary (this doesn't increase US deterrence, but it is something that affects others' calculus on how survivable their nuclear arsenals are). As a second-strike system, submarines are hard to find and mobile, which gives them a huge survivability advantage over land-based systems. Every few years, I see a new comment about wave-measuring satellites + AI or quantum-computing measurement of local gravity or other high-tech systems that might make submarines much more vulnerable. Because of this, I disagree with some people who want to move exclusively to submarine-based strategic weapons: they are near-invulnerable right now, but having a diversity of systems means that they don't have the same weaknesses. I could see a situation in the future where the President wakes up to find that a patient adversary has invested in [amazing sensing technology] and managed to locate and sink all of the 14 nuclear-missile submarines at sea; that would be an extremely vulnerable time, where e.g. that adversary could threaten to nuke the docks where the other SSBNs are tied up if they sense them starting to move.

As long as mutally-assured destruction is the governing logic, the most-stable situation is where the US (and honestly, its adversaries if they already have nuclear weapons) are secure enough in their second-strike capabilities that they can wait as long as possible before launching. The "perfect" nuclear deterrence weapon is not useful as a first-strike weapon (ineffective against enemy second-strike weapons), invulnverable against enemy first-strike weapons, and reliable and destructive enough that the enemy knows that if used, they would be hit hard. The only potential addition to the triad that I've heard of recently is orbit-based weapons, which have the disadvantages of ICBMs (known location), submarine-based missiles (induce fear of first strikes in adversaries), and would also be nearly impossible to maintain (because they're in space) and not provide any significant advantage over existing systems.

Targets of Destruction

A major question of mutually-assured destruction is "what are we threatening to destroy?" The two main options are "counterforce" and "countervalue". Counterforce is aiming nuclear weapons at military targets, primarily an opponent's nuclear weapons. Countervalue is aiming nuclear weapons at civilian and military targets, including major population centers. During World War II, countervalue strikes included the firebombing of Dresden and Tokyo, as well as the nuclear attacks on Hiroshima and Nagasaki. All of those cities had military value, but civilian casualties were deliberately courted. Because of the destructiveness of nuclear weapons, even counterforce strikes tend to involve the destruction of cities; counterforce may seem less omnicidal, but ask yourself "how many military bases and headquarters are next to (or in!) major cities?"

Daniel Ellsberg (of Pentagon Papers fame) in his 2017 book The Doomsday Machine asked a question I hadn't considered: if we are in the opening minutes of a strategic nuclear exchange, what do we want from our enemy? For them to stop! The irony of this realization is that it reverses some targeting priorities: typically command-and-control systems are among the first targets in counterforce (or countervalue) to stop the enemy from launching any more weapons. However, those same systems are the only way the enemy might have to reverse any attack orders (e.g. recall bombers and/or submarines, halt any missiles not yet launched)! I don't have a complete answer on this, but there's an argument for attacking enemy delivery systems (maybe even with conventional weapons if in the middle of a city) while keeping command-and-control intact so that there's someone to negotiate with, and retaining the monster of countervalue as a bargaining chip. On the other hand, if the President (or surviving government leader) has just experienced tens of millions of Americans dying or about to die, is there even a possibility that they would not want revenge? The movie "Fail Safe" (spoiler ahead) posits an accidental US nuclear strike on Moscow (bombers erroneously given the "go" order and they cannot trust any of the orders to abort), and to avoid a general nuclear exchange the US President orders a nuclear strike on New York City.

Because of the firepower of nuclear weapons, all strategic weapons end up being countervalue weapons.

Weapons of Destruction

On the way to mutually-assured destruction is the "escalation ladder": in theory nuclear-armed countries can fight each other conventionally without resorting to nukes (historically happened with India and Pakistan) or even fight with "tactical" or "theater" nuclear weapons (smaller, typically aimed at military formations or bases) without resorting to strategic weapons and a general exchange. I am skeptical of our ability to contain the use of nuclear weapons once they start, regardless of "tactical" or "strategic". The fact that the weapons haven't been used creates a taboo, and once it's broken we have no experience on where we might stop.

One telling example is the Proud Prophet" wargame run by the US in 1983. The intent was to use actual decision-makers to work through a realistic scenario of the USSR invading Europe. One of the features of US doctrine at the time was "demonstration" nuclear attacks, where an outnumbered NATO would use a small number of tactical nuclear weapons on the battlefield to a) destroy Soviet forces, and b) more importantly, create fear and a desire to stop in the mind of the Soviets. The result of the wargame was that the nuclear strikes actually intensified the determination of the people playing the Soviet side, and "tit-for-tat" nuclear strikes escalated from tactical to theater to strategic weapons and a general nuclear exchange. Many people involved were surprised and terrified by how the situation had gotten out of everyone's control, and later US doctrine emphasized surging more conventional forces and speeding up the ability to reinforce them. A great write-up of this appears in the December 2024 New York Times Magazine.

The idea of "tactical" weapons is that they provide an intermediate step between conventional and strategic weapons, but because they have never been used either, the risks of escalation are unacceptable. Increasing options with conventional forces is much safer. In addition, the primary US tactical weapon in Europe at this point is the B-61 gravity bomb, used on (among others) the F-35 fighter. Because the B-61 is unpowered, it has a limited range and so would be risky for aircraft to deliver compared to an air-launched cruise missile (conventional or nuclear).

A related idea is putting conventional warheads on ICBMs; although this has the advantage of increasing flexibility, it takes a previously nuclear system and makes it dual-use. The fear is that an adversary might see an ICBM incoming to their territory and not know which warhead it has: are they under nuclear attack? This uncertainty makes miscalculations and escalations more likely and outweighs the advantages.

For a funny viewing on the advantages of optionality, Dr. Bret Devereaux pointed me to this clip from "Yes, Prime Minister". In it, several Soviet attacks are posited which all fall below the threshold of nuclear retaliation; so if there aren't sufficient conventional forces, the Prime Minister has no options at all.

Conclusions

Nuclear weapons are never a happy topic, but the impending anniversary on August 9 is a victory for humanity. Unfortunately, I don't see nuclear weapons going away soon, so all we can do is remain clear-eyed about the risks and avoid miscalculations for as long as we can. And remember to celebrate Stanisav Petrov Day on September 26.