Why America Stands to Lose If It Resumes Nuclear Testing

People close to former President Donald Trump, including those who could serve in a second administration of his, are once again floating the unhelpful idea of the United States resuming nuclear weapons testing. Although the Trump administration did not conduct a nuclear test explosion while in office, senior national security officials did consider the idea in May 2020, according to The Washington Post. Had the administration proceeded with the idea, it would have been the first U.S. nuclear test since 1992 and would likely have encouraged other countries to do the same.

Yet the idea is not dead—far from it. In a Foreign Affairs essay, Robert O’Brien, Trump’s national security adviser from 2019 to 2021, made the startling suggestion that the United States should resume the practice of exploding nuclear weapons underneath the Nevada desert. “The United States has to maintain technical and numerical superiority to the combined Chinese and Russian nuclear stockpiles,” O’Brien wrote. “To do so, Washington must test new nuclear weapons for reliability and safety in the real world for the first time since 1992—not just by using computer models.”

The United States has not conducted a nuclear explosion in more than 30 years. Nor has Russia or China. The United States, Russia, and China are among the 187 countries that have signed the 1996 Comprehensive Nuclear Test Ban Treaty (CTBT) which prohibits all nuclear explosions, of any size. The treaty is administered by the Comprehensive Nuclear Test Ban Treaty Organization, which maintains a vast international monitoring system of seismic, infrasound, hydroacoustic, and radionuclide stations. Thanks to the treaty, there is a strong international norm against nuclear testing. Since nuclear detonations by India and Pakistan in 1998, no other country has conducted an explosion, with the lamentable exception of North Korea.

Still, the treaty itself has never entered into force. Although the United States signed it in 1996, Republicans in the Senate voted against ratifying it. China, which also signed, has refused to ratify it until the United States does. And Russia, which ratified the treaty in 2000, withdrew its ratification in November 2023. After nearly 30 years in limbo, the treaty’s future is very uncertain.

O’Brien’s proposal, then, comes at an awkward moment. There are real questions about whether Russia might resume nuclear explosions. Russia, China and the United States all keep their nuclear test sites in at least a partial state of readiness. Satellite images show new buildings, ongoing work on tunnels, and a steady stream of people and equipment at all three sites. The big problem is the possibility that one of the big three nuclear powers will resume detonating nuclear weapons. Once one does, the other two are likely to follow suit. And make no mistake: the resumption of nuclear weapons testing would be bad for the United States.

PLAYING CATCH-UP

In his essay, O’Brien argued that nuclear testing—not just computer modeling—is essential for maintaining U.S. technical superiority over Russia and China. But a world in which China and Russia can resume nuclear explosions is one in which they will catch up with the United States, not fall behind. Plus, the purpose of testing nuclear weapons is not to increase their reliability and safety, as O’Brien wrote, but to collect data to validate computer models.

The United States has conducted more nuclear detonations (1,149) than Russia (969) and China (45) combined. Even with all these tests, the United States almost never took a nuclear weapon out of the stockpile and then tested it by detonating it in the desert; instead, the tests were more like experiments. Only in the last decade or so of nuclear testing did the United States test a limited number of weapons out of the stockpile. Even then, however, a single explosion doesn’t establish anything like statistical reliability that a weapon would work as intended. The outcome of a test is better described as confidence, a sense that everything works as it should—confidence in the computer models, confidence in the people who design the bombs, and confidence in the processes for making them.

The United States has conducted more nuclear detonations than Russia and China combined.

When the United States stopped testing in 1992 and signed the CTBT in 1996, it was much better positioned than Russia or China to maintain its confidence in its nuclear arsenal without nuclear explosions. The United States had conducted more explosions. It had also gotten more out of each test, thanks to its technological advantages—such as fiber optic cables that could relay data effectively without attenuation from the effects of a nuclear explosion.

The United States had another advantage. In 1996, its supercomputers were vastly superior to those in Russia and China. When negotiating the CTBT, both Russia and China had such poor data and lousy supercomputers that they were reluctant to agree to a comprehensive ban on nuclear explosions; instead, they preferred an exception for very small explosions. The United States held firm on a “zero yield” treaty. Then, both countries asked the Clinton administration to ease restrictions on supercomputer exports as a condition of signing the treaty. Russia got a few restrictions lifted, but China got nothing.

SMARTER WITHOUT TESTING

The United States combined supercomputing and the wealth of data from the tests it had conducted into something it called “science-based stockpile stewardship,” a massive investment in science, surveillance, facilities, and computing that has allowed the country to maintain its nuclear arsenal without testing.

Under this approach, the United States conducts much closer surveillance of the nuclear weapons in its stockpile than it ever did during the Cold War. It has also invested in better understanding how thermonuclear weapons really work. During the Cold War, American scientists had a relatively poor understanding of why nuclear weapons behaved the way they did, meaning that models of how the weapons would perform had to be fine-tuned in an ad hoc way—for example, multiplying a variable by two, even if no one could quite say why it worked that way. Many, although not all, of these adjustments have in recent years been replaced by a more complete scientific understanding of the principles behind thermonuclear explosions.

The United States has spent billions of dollars on the infrastructure needed to monitor and understand its nuclear weapons without live testing. The United States has an underground laboratory—the Principal Underground Laboratory for Subcritical Experimentation, or PULSE—that houses machines that conduct no-yield experiments involving small amounts of plutonium. It also maintains a facility to X-ray bombs as they detonate without plutonium and another one to study fusion processes in thermonuclear weapons. (The latter, the National Ignition Facility, cost $3.5 billion to build.) As a result, the United States understands how nuclear weapons work much better today than it ever did when it was conducting nuclear explosions.

Russia and China made some similar investments, although not at the same scale and seemingly without the same result. Russia is thought to simply remanufacture each nuclear weapon about every ten years. Given some of the reliability problems that the Russian military has experienced in Ukraine, it’s fair to wonder how much faith one should put in Russia’s ability to make the same weapon over and over without detrimental changes creeping in.

LOSING YOUR EDGE

Underlying the United States’ science-based stockpile stewardship approach was a commitment to invest in supercomputing. In 1995, the computing requirements were considered daunting. The Department of Energy estimated that it would need a computer capable of about 100 teraflops, an amount of processing power one official called “insane.” Today, the fastest computer in the world, at Oak Ridge National Laboratory in Tennessee, is about 10,000 times faster than that.

Of course, the United States no longer has the supercomputing lead it once did. For many years, the fastest supercomputer in the world has been at a nuclear weapons laboratory in China, not in the United States. But even with some of the world’s fastest supercomputers, there is one thing that China doesn’t have: the testing data for its computer codes. Without that data, which the United States has collected from years of conducting nuclear explosions, Chinese designers can less readily make changes to existing weapon designs that have already been tested. That might make it harder for China to make new, more miniaturized nuclear warheads.

The technical superiority of the American nuclear stockpile exists only because Russia and China quit testing and the United States invested heavily in science. If a second Trump administration resumed nuclear testing, Russia and China would surely follow suit—and because they have more to learn from each test, they would erode the United States’ advantage. Moreover, new or emerging nuclear weapons states—say, Iran or Saudi Arabia—would feel no constraints against carrying out explosion tests. The result would be that the United States’ nuclear-armed foes would be even more capable.

The naive belief that resuming nuclear explosions would lengthen the United States’ lead evokes a long tradition of shortsighted thinking about the bomb, with politicians and pundits unable to plan more than one move ahead. It’s a failing that dates to the very inception of the bomb in 1945.

Early on, advocates of using the bomb against Japan believed that the United States would enjoy a nuclear monopoly over the Soviet Union for decades. Moscow acquired its own bomb in less than four years. The same people then argued that developing a thermonuclear weapon, or hydrogen bomb, would restore the advantage. The Soviet Union obtained one less than two years after the United States, followed eventually by others, including China. At each step in the arms race, policymakers have succumbed to such wishful thinking. If the United States moves first to resume nuclear weapons testing, it will quickly learn how naive it has been once again.