The Necessity of Testing Our Nuclear Weapons

Secretary of War Pete Hegseth’s recent announcement that the U.S. will resume testing nuclear weapons has alarmed some nuclear-arms experts. It shouldn’t. President Trump had already announced earlier this year that a resumption of testing “on an equal basis” with Russia and China, something the United States hasn’t done since 1992, would be necessary.

However, it is not certain that Russia and/or China would resume nuclear testing in response to U.S. testing. Nor is it certain that Russia or China would gain more than the United States if it resumed nuclear testing.

In a recent Foreign Affairs article, former Los Alamos Lab director Sigfried Hecker walked through the decision to halt nuclear testing in the early 1990s. He noted that once the political decision was made to halt testing before conducting a series of tests designed to explore the impact of plutonium aging on weapons performance, “the United States missed the chance to examine concerns about age degradation of its weapons. China and France, by contrast, each conducted at least six nuclear tests before signing the [nuclear Comprehensive Test Ban Treaty]. Consequently, both were better prepared for a test ban than the United States.”

Instead of making new warheads over the last few decades, the United States has “extended the life” of existing warheads—refurbishing warhead components as necessary, while still keeping the same plutonium within the warhead itself. As a consequence, not only are U.S. nuclear warheads decades old, but so is the plutonium that serves as the fuel of a nuclear weapon and makes a nuclear detonation possible.

“Extending the life of existing [plutonium] pits with confidence is difficult because of how plutonium degrades,” Dr. Hecker points out. “The aging characteristics are now under intense study at the weapons labs, but without testing, the implosion characteristics are difficult to assess, and the [warhead] lifetimes are uncertain.” Indeed, he adds, “scientists do not know what the acceptable lifetimes are for existing plutonium pits.”

Many experts, including former National Nuclear Security Administration chief Jill Hruby, argue that there’s “no scientific reason” to test, due to advances in modelling and simulation that simulate nuclear explosions through various supercomputer models and because of various sub-critical experiments that test aspects of a nuclear detonation virtually, without a mushroom cloud.

There are indeed many parts of a nuclear weapon that can be tested virtually, and the United States should continue to engage in both sub-critical nuclear experiments and computer-based modelling and simulation efforts.

But these are not substitutes for actual explosive, yield-producing nuclear testing.

While it is true that during the Cold War the United States conducted hundreds nuclear tests, all of them used comparatively “new” warheads. Never has the United States tested weapons that were built decades prior to a test.

None of the Cold War tests used plutonium pits that were decades old. As such, the data that goes into contemporary computer-based modelling and simulation experiments are based upon nuclear detonations that used “fresh” plutonium and new warheads using data that was collected in the pre-digital age.

For these reasons, it is reasonable to question the absolute certainty of such computer-based modelling and simulation efforts.

In many ways, the United States’ strategic arsenal is like a 1980 Cadillac inherited from our grandfather. The Cadillac has been kept in the garage for decades and never driven. It has new tires, sparkplugs, and a new battery—but the fuel in the engine is 45 years old. If someone had to drive that Cadillac from New York to Los Angeles, would that person not first want to take it for a test drive across town or even to neighboring Connecticut to ensure that it works properly before driving it over the Rocky Mountains and across the desert?

Any sane person would do so. And the stakes involved in driving from New York to Los Angeles are far lower than the stakes of convincing America’s adversaries that the U.S. strategic deterrent is credible.

And it is not just the existing arsenal that may need to be tested to ensure that it will perform according to military requirements. The United States should, by the 2030s, be fielding entirely new warheads fueled by entirely new plutonium pits. Will the United States need to conduct a nuclear explosive test on these weapons? Perhaps.

To continue with the analogy, let’s say that Cadillac for the first time in 40 years builds a new class of luxury sedan. While Cadillac would surely do modelling and simulation experiments on the design, it would also have humans test-drive the vehicle to ensure it performed as expected before shipping them out to dealerships. It is inconceivable that they would NOT test the new sedans before putting them on the showroom floor.

Testing isn’t something the United States should fear. Ensuring that a weapon—particularly one that is designed to deter a catastrophic strategic attack on the American homeland—performs according to its military specifications is necessary to ensure that it works and that everyone—Americans, allies, and our enemies—has confidence that it will perform.

The stakes of ensuring that a nuclear war doesn’t break out are too high NOT to do so.

Robert Peters is a Senior Research Fellow for Strategic Deterrence in The Heritage Foundation’s Allison Center for National Security.