The Most Important Nuclear Weapon is the One We Don’t Build
The United States and China may have found a small opening in an increasingly dangerous nuclear landscape.
American and Chinese officials have held preliminary discussions about allowing reciprocal visits to nuclear facilities, with Chinese officials visiting U.S. nuclear laboratories and potentially testing sites while American officials receive similar access in China. According to reporting by CNN, subsequently confirmed by Reuters, the idea was raised by China during informal discussions earlier this year.
That detail matters.
For years, Washington has attempted to bring China into arms-control negotiations alongside the United States and Russia. Beijing has resisted, pointing out an obvious disparity: its nuclear arsenal remains considerably smaller than those maintained by the two Cold War superpowers.
That disparity is real. The Stockholm International Peace Research Institute estimates China possesses approximately 620 nuclear warheads. The United States and Russia together still possess roughly 86 percent of the world’s nuclear weapons.
China’s arsenal is nevertheless growing quickly, and that expansion deserves serious attention.
But the answer to China’s nuclear buildup cannot simply be an American nuclear buildup.
That is how arms races work.
One country builds because it fears another. The other responds because the first is building. Each describes its weapons as defensive and the other’s as threatening. Each new missile becomes justification for another missile, another submarine, another bomber, another warhead.
Eventually, everyone becomes less secure while claiming to have increased security.
Reciprocal nuclear-site visits offer another possibility.
They will not eliminate nuclear weapons. They will not resolve tensions over Taiwan. They will not immediately produce a sweeping trilateral arms-control treaty.
They could, however, do something enormously valuable: reduce uncertainty.
And uncertainty may be one of the most dangerous weapons in the nuclear arsenal.
When countries cannot see what their adversaries are doing, governments prepare for worst-case scenarios. A missile test can look like preparation for an attack. An exercise can look like mobilization. Construction at a nuclear facility can generate intelligence assessments, speculation and countermeasures.
Transparency does not eliminate distrust. It gives countries information with which to manage it.
That is why inspections, notifications, hotlines and verification systems have played such important roles in nuclear arms control. The machinery of peace is rarely dramatic. It consists of scientists looking at instruments, diplomats exchanging data and inspectors walking through facilities.
None of it makes for an impressive military parade.
But it can prevent a misunderstanding from becoming a catastrophe.
That infrastructure is especially important now.
On February 5, 2026, New START, the last treaty limiting the strategic nuclear arsenals of the United States and Russia, expired. For the first time in decades, the world’s two largest nuclear powers are operating without a legally binding bilateral agreement limiting their nuclear arsenals.
The disappearance of those guardrails comes just as China is expanding its forces and the United States faces growing pressure to respond.
SIPRI warns that nuclear advocates in the United States are already calling for additional American deployments specifically in response to China’s expansion.
This is the moment when political leaders must decide what kind of security they actually want.
One approach measures security primarily in weapons.
If China has more missiles, we need more missiles. If China constructs silos, we must expand our arsenal. If China develops another generation of warheads, America must develop another generation of warheads.
There is no natural endpoint to that logic.
Peace requires a different calculation.
What if we measured security by how difficult it is to accidentally start a nuclear war?
What if we measured it by how frequently military officials communicate?
By how much each government knows about the other’s nuclear doctrine?
By whether missile launches and exercises are announced in advance?
By whether independent instruments can distinguish an underground weapons test from an earthquake?
By whether inspectors can enter facilities rather than intelligence agencies having to guess what happens inside them?
Those things are security infrastructure too.
And compared with nuclear weapons, they are remarkably inexpensive.
The United States is preparing to spend enormous sums modernizing its nuclear arsenal over the coming decades. Every dollar devoted to weapons represents a political decision about what we believe keeps us safe and what kinds of threats deserve national resources.
A reciprocal inspection regime between China and the United States would cost a fraction of another weapons system.
Its value could be immense.
China’s suggestion should therefore neither be romanticized nor dismissed.
There are legitimate reasons for caution. Reciprocal visits can become carefully choreographed performances. Governments can hide facilities, restrict inspectors or use negotiations to delay other actions. Verification must be rigorous enough to produce meaningful information rather than diplomatic theater.
But skepticism is not an argument against diplomacy.
It is an argument for good diplomacy.
Nor should Washington insist that the first step must be a treaty imposing identical numerical limits on three profoundly unequal arsenals. China has long argued that it should not accept the same numerical restrictions as countries possessing substantially larger stockpiles.
Washington does not have to accept China’s position completely to recognize that a successful negotiation might begin somewhere else.
Start with visits.
Then expand military hotlines.
Create stronger notification requirements for missile tests and major exercises.
Develop shared procedures for investigating suspected nuclear tests.
Increase exchanges between nuclear scientists.
Discuss doctrines governing when nuclear weapons might be used.
Build mechanisms to prevent cyberattacks, false warnings or technical failures from triggering escalation.
And continue talking.
We know this kind of work is possible because much of the architecture that prevented U.S.-Soviet competition from becoming nuclear war was constructed gradually.
The Cuban Missile Crisis brought the world terrifyingly close to catastrophe before Washington and Moscow began building some of their most basic mechanisms for crisis communication and nuclear risk reduction.
We should not require another Cuban Missile Crisis to rediscover their value.
The emerging U.S.-China nuclear relationship presents a choice.
We can wait until another confrontation—a crisis over Taiwan, a misunderstood missile launch, a military exercise interpreted as preparation for attack—forces both countries to confront how close they have come to disaster.
Or we can construct the guardrails before reaching the cliff.
For decades, advocates of nuclear deterrence have told us that peace depends upon maintaining the ability to destroy an adversary.
But deterrence has never functioned through weapons alone.
It has also depended upon communication, restraint, information, verification and the ability of adversaries to understand one another well enough not to mistake fear for aggression.
Those mechanisms rarely receive the attention given to missiles and bombers because there is nothing spectacular about preventing something from happening.
There is no mushroom cloud from the war that never begins.
No parade commemorating the missile that was never launched.
No monument to the misunderstanding that was resolved by a telephone call.
Yet those may be among the most important victories nations can achieve.
If China and the United States are genuinely willing to open portions of their nuclear complexes to one another, they should pursue the opportunity aggressively and carefully.
Not because reciprocal visits will end the nuclear arms race.
Because they offer something increasingly rare in international politics:
a way to become safer without building another weapon.
