Last month, U.S. President Donald Trump rekindled a decades-old debate about nuclear testing. “Because of other countries testing programs,” he wrote on social media, “I have instructed the Department of War to start testing our Nuclear Weapons on an equal basis.”

The United States has not tested a nuclear weapon since September 23, 1992. Energy Secretary Chris Wright quickly tried to walk back Trump’s post, explaining on Fox News that the president had meant testing nuclear delivery systems, not explosives. But Trump doubled down on his original meaning, telling60 Minutes that he ordered resumption of nuclear testing in response to alleged secret underground nuclear detonations by China and Russia. “They test way underground where people don’t know what’s happening with the test,” he said.

Intelligence officials in both the first Trump and Biden administrations questioned whether U.S. adversaries, especially China and Russia, are testing nuclear weapons clandestinely. Both the Heritage Foundation’s Project 2025 and Trump’s former national security adviser, Robert O’Brien, have called for resuming testing to ensure the United States’ aging nuclear arsenal is safe, effective, and reliable. The United States has roughly 3,700 nuclear weapons, of which about 1,770 are believed to be deployed militarily—but all these weapons were produced and tested before the United States conducted its last test in 1992.

I am deeply familiar with this debate. When I first arrived at the Los Alamos Scientific Laboratory in New Mexico as a summer graduate student in 1965, the United States had conducted more than 400 nuclear tests, about half in the atmosphere and the rest underground. Its nuclear arsenal consisted of some 30,000 weapons. I served as director of the laboratory from 1986 through 1997 during the collapse of the Soviet Union, which resulted in drastic reductions of Russian and American nuclear arsenals, a nuclear testing moratorium in 1992, and a comprehensive nuclear test ban treaty in 1996.

Now, some 33 years after the last American nuclear test, the aging nuclear arsenal presents increasing challenges to maintain its safety, effectiveness, and reliability. But a return to testing at this time would likely benefit U.S. adversaries more than it would the United States. Worse still, it might rekindle an even greater and broader arms race than in the first few decades of the Cold War.

TESTING SPREE

Nuclear testing was deemed essential in the 1950s as the Soviet Union and the United States raced toward hydrogen bombs and missile-deliverable nuclear weapons. By the late 1950s, atmospheric explosions had inflicted a heavy health and environmental toll on the world with radioactive fallout. The consequences were even more dramatic in the northern Marshall Islands, where the United States was conducting most of its tests. Global public opposition led to a testing moratorium from 1958 to 1961 by the three nuclear weapons states at that time: the Soviet Union, the United Kingdom, and the United States.

During the moratorium, scientists at the Los Alamos lab using computer models discovered potential major safety concerns. They found that some of the nuclear weapons were not “one-point safe.” That is, an accident rather than intentional arming might trigger a nuclear explosion. They devised an ingenious scheme of very low-yield tests to find out. President Dwight Eisenhower was informed because of the sensitivity of potential abrogation of the test moratorium, and he authorized what were called “hydronuclear experiments” to be conducted unannounced, underground at Los Alamos. The nuclear yields from these experiments were typically much less than that generated by one pound of TNT, not tons or kilotons.

Thirty-five such experiments were conducted at Los Alamos and a smaller number at the Nevada Test Site by the Lawrence Livermore Laboratory. The experiments ended up being critical to making safety adjustments to Los Alamos systems. In 1966, for instance, a B-52 bomber carrying nuclear weapons crashed over Palomares, Spain. Had those weapons not been modified because of the hydronuclear program, the crash would likely have resulted in a significant nuclear explosion instead of setting off only the chemical explosives, which caused plutonium dispersal along the beaches.

Nuclear tests came roaring back in September 1961, when the Soviets shocked the world and abrogated the first testing moratorium. The United States was quick to respond—with ten tests later that year and with a mind-boggling 98 tests in 1962, with nearly half in the atmosphere. As concerns about nuclear fallout intensified, however, and public pressure again mounted, Soviet Premier Nikita Khrushchev, U.S. President John F. Kennedy, and British Prime Minister Harold McMillan agreed to limit nuclear testing to underground explosions in the Limited Test Ban Treaty signed in 1963. The last atmospheric tests by the three were in 1962. France and China, which joined the nuclear club around that time, conducted their last atmospheric tests in 1974 and 1980, respectively.

TESTING GOES UNDERGROUND

Despite the Limited Test Ban Treaty, the Soviet Union and the United States were still in an intense arms race. The Soviets were playing catch-up, adding more than 10,000 weapons every ten years (finally topping out at 41,000 by the end of the 1980s). The United States’ arsenal had peaked at 31,000 in 1967 and was slowly decreasing, but at the same time, the military was substantially increasing missile accuracy and survivability. The Cold War lingered on and nuclear testing, now underground, was still essential to both. In the ten years following the new treaty, the United States conducted nearly 400 underground tests and the Soviet Union about 170. Public opposition was muted because there were essentially no visible signatures of testing and the detrimental effects were dramatically less.

Most nonnuclear weapons states, however, still called for a complete test ban. The Soviets and Americans could only agree on limiting the size of underground tests to less than 150 kilotons with the Threshold Test Ban Treaty signed in 1974. The 150-kiloton limit was seen as a balanced compromise between politically demonstrating some nuclear restraint while maintaining the ability to modernize nuclear arsenals. It was initially viewed as being verifiable, yet it took 16 years for the two countries to ratify the TTBT because each side suspected the other of cheating. In the meantime, the two countries conducted another roughly 300 nuclear tests each.

The wheels were finally set in motion to ratify the TTBT once Soviet leader Mikhail Gorbachev and U.S. President Ronald Reagan met in their historic October 1986 summit in Reykjavik. Although the two men fell short of their goal—total elimination of nuclear weapons—the summit led to weapons scientists being asked to develop adequate verification techniques so that each side could get the assurances it needed to ratify the treaty.

To provide this assurance, we hosted Soviet scientists at the Nevada Test Site for months. The Soviet scientists made onsite measurements during the detonation of one of our nuclear devices in August 1988. Nongovernmental U.S. scientists, meanwhile, measured seismic signals remotely to compare. The sequence was then reversed when our nuclear scientists made onsite measurements of a Soviet device detonated at Semipalatinsk Test Site one month later. These measurements—and adroit negotiations—led to strong support for the treaty from both governments. In the U.S. Senate, it passed by a remarkable 98 to 0 vote in 1990.

GETTING TO ZERO

With the demise of the Soviet Union in December 1991, advocates in the United States pressed for a comprehensive test ban. The bipartisan Hatfield-Exon-Mitchell amendment to the 1993 Energy and Water Development Appropriations Act allowed for a maximum of 15 nuclear tests to be conducted over the next four years before the United States would sign a Comprehensive Test Ban Treaty (CTBT) with other nuclear-armed countries. President George H. W. Bush signed the act reluctantly in October 1992; although he believed nuclear testing was important for national security, the amendment was cleverly attached to a spending bill that he greatly favored.

Upon assuming office just three months later, President Bill Clinton made a comprehensive nuclear test ban a top priority. There was a compelling logic to allow Los Alamos and other Department of Energy weapons laboratories to conduct an additional last few tests to prepare for a world in which no tests would be allowed, but Clinton chose to stop testing immediately to demonstrate his commitment to a comprehensive test ban.

During the next few years, Clinton administration officials led the laboratories and the military in intense discussions about the advisability of such a drastic change to the country’s nuclear stockpile preparedness. For example, in June 1995, the commander of the Strategic Command, who had responsibility for the United States’ nuclear forces, brought the lab leaders and officials from the Departments of Energy and Defense to Omaha for a “Stockpile Confidence Conference.” The labs explained what could be accomplished with testing limits of ten kilotons, one kiloton, and less.

Department of Energy Secretary Hazel O’Leary made it clear that kiloton yields were not in the cards. Based on the Los Alamos experience of the value of very low-test limits during the 1958–61 test moratorium, I made the case for the importance of retaining the right to do such hydronuclear experiments—that is, yield limits of pounds or less. Although the technical specialists differed on the value of such experiments, it did not really matter: for O’Leary and Clinton, there was a political imperative to get to zero.

Before Clinton made his final decision on the treaty, he instructed the chairman of the Joint Chiefs of Staff, General John Shalikashvili, to ask the weapons lab directors if they would support a zero-yield test ban treaty. I remember how, in July 1995, Shalikashvili looked me directly in the eyes in his Pentagon office and asked if he would have to tell Clinton that testing was needed to keep our nuclear weapons safe and reliable. I told him that the nuclear stockpile was safe and reliable—but he would need to ask that question again in the coming years.

As a result of that meeting, the three lab directors were required to certify the nuclear stockpile annually in writing to the secretaries of defense and energy, who would inform the president, together with a letter by the commander of U.S. Strategic Command. Clinton announced a zero-yield test ban on August 11, 1995. The laboratories were thus limited to subcritical experiments, which would not generate a sustained nuclear reaction. Hydronuclear experiments would not be allowed.

This represented a dramatic change for the laboratories and significant risk for the nation’s nuclear arsenal preparedness. The Clinton administration addressed concerns expressed by the laboratory directors and the military by issuing six safeguards that defined the conditions under which the United States would enter into the CTBT. At the top was “the conduct of a science-based stockpile stewardship program to ensure a high level of confidence in the safety and reliability of nuclear weapons in the active stockpile.” Victor Reis, Department of Energy assistant secretary for defense programs, led this effort, and the goal was for the laboratories to develop a better understanding of the materials used and the processes involved during the entire life cycle of a nuclear weapon—up to and including detonation—in the absence of nuclear testing.

Somewhat paradoxically, with nuclear testing we were able to produce a reliable stockpile without fully understanding the science. In other words, nuclear tests compensated for the lack of understanding the complex scientific fundamentals of nuclear explosions. Gaining such understanding, it was hoped, would allow the labs and the military to keep the nuclear stockpile safe, secure, and reliable without nuclear testing.

The safeguards also included maintenance of nuclear laboratories, nuclear test readiness, programs to improve treaty monitoring capabilities and operations, and commitments in the areas of intelligence gathering and analysis. The final safeguard specified the circumstances under which the president would be prepared, in consultation with Congress, to conduct the necessary testing if the safety or reliability of the U.S. nuclear deterrent could no longer be certified.

On September 24, 1996, the United States joined 70 other countries, including China, France, Russia, and the United Kingdom, in signing the CTBT. As Clinton stated at the time, “This treaty is the longest-sought, hardest-fought prize in the history of arms control. Its message is unmistakable: the era of nuclear explosive testing is over.”

GETTING THE SCIENCE RIGHT

I was not as enthusiastic as the president about the CTBT. I viewed moving from the testing to the no-test era as an enormous challenge that required steadfast government commitment and the ability to continue to attract the best and the brightest to U.S. laboratories. Moreover, by bypassing the final tests that had been laid out in 1992, 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 CTBT. Consequently, both were better prepared for a test ban than the United States was. I was encouraged, however, by Reis’s valiant effort to develop a well-funded Stockpile Stewardship Management Program (SSMP). It was a chance to get the science right to help alleviate the loss of nuclear testing.

The Clinton administration faced substantial hurdles to ratify the CTBT. When Senator Jon Kyl of Arizona, who was vehemently opposed to ratifying the CTBT, asked me and the other lab directors if the program would give us the same confidence as was achieved by nuclear testing, I replied: “I believe that the SSMP as currently configured and fully funded provides the best approach to keeping the confidence level in our nuclear stockpile as high as possible for the foreseeable future. . . . A robust nuclear testing program would undoubtedly increase our confidence. However, our long-term confidence in the stockpile would suffer if we substituted a program consisting of an occasional nuclear test for a robust stewardship program.”

In his reply to Kyl, Bruce Tater, the director of the Lawrence Livermore lab, also emphasized that a strong stockpile stewardship program is “necessary to underwrite confidence” in the U.S. nuclear weapons stockpile in an era without nuclear explosive testing.

That is where I believe we still stand today. Although the Senate failed by a large margin to ratify the CTBT in 1999, the United States has been committed to a voluntary testing moratorium and has not tested weapons at any yield. The stewardship program, meanwhile, has been funded generously. The laboratories have continued to attract outstanding scientists and engineers. And with new facilities, innovative nonnuclear experiments, and greatly improved analytical and computational tools, the United States has gained a better understanding of the materials and processes in nuclear weapons.

Nuclear testing provided a robust shortcut. With the stockpile stewardship program, the United States has greatly increased its knowledge, leading to an improved understanding but not necessarily improved confidence. Nevertheless, the directors of the laboratories have continued to certify the stockpile every year since I signed the first two letters for the Los Alamos lab in 1996 and 1997.

THE PLUTONIUM CHALLENGE

The most problematic concern in the stockpile today is the plutonium present in the cores, or pits, of the triggers of modern, two-stage thermonuclear weapons. Plutonium is used in nuclear weapons because of its nuclear characteristics, but its electronic structure makes it the most complex element in the periodic table and an engineering nightmare.

Although numerous materials in weapons present significant physics and engineering challenges, most can be addressed with nonnuclear tests. For plutonium, the United States has historically relied on nuclear testing to compensate for a lack of fundamental understanding. In the past three decades without testing, the United States has resorted to extending the lifetimes of existing pits or has tried to remanufacture them to previous specifications. Both have turned out to be monumental challenges.

Extending the life of existing pits with confidence is difficult because of how plutonium degrades.Plutonium ages not only from the outside in (like the oxidation and rusting of many metals) but also from the inside out. It radioactively decays by alpha-particle emission, and the result is that it transmutes into other elements such as americium, uranium, and neptunium, and it also generates helium during the decay. The helium atoms are stored in the plutonium metal structure, with some significant fraction forming microscopic bubbles. Moreover, the alpha decay of plutonium sends energetic uranium atoms crashing through the plutonium metal, greatly disrupting the regularity of plutonium’s crystal structure, changing its properties.

A return to nuclear testing might rekindle an arms race.

These challenges received little scientific attention during the testing days, and few aged nuclear devices were tested because the weapon systems were regularly replaced with new designs and, hence, new plutonium pits. The aging characteristics are now under intense study at the weapons labs, but without testing, the implosion characteristics are difficult to assess and the lifetimes are uncertain.

Until 1989, new plutonium pits were produced in the Rocky Flats nuclear facility near Denver, Colorado. But that facility shut down permanently because of alleged environmental violations, and every attempt to provide the nation with an adequate pit production facility since has run into political opposition or major time delays and cost overruns.Construction of a new pit production facility is now underway at the Savannah River nuclear site in South Carolina. The plutonium R & D facility at Los Alamos, meanwhile, was reconfigured to make the very few new pits that have been produced since 1989.

A fundamental challenge is that the United States does not know what level of pit manufacturing is necessary to maintain its nuclear deterrent because scientists do not know what the acceptable lifetimes are for existing plutonium pits that need to be replaced or if the U.S. government may choose to build different weapons to maintain its deterrent. In any case, the United States needs a fully functioning, resilient nuclear weapons production complex, especially for plutonium pit production in addition to a robust stewardship program.

TESTING TRADEOFF

Whereas resumption of full-scale nuclear testing would allow the United States to answer some pressing questions about the fitness of its stockpile, it would provide even greater benefit to China and Russia. When Trump posted on social media that China and Russia were already testing nuclear weapons, it is likely he was referring to hydronuclear experiments with ultralow yields, which are not detectable without an onsite presence. (If China or Russia were doing nuclear tests in the multikiloton range, it would almost certainly be detected by the elaborate international test-monitoring system established by the Comprehensive Test Ban Treaty Organization.)

Neither China nor Russia likely sees these experiments as violating the CTBT. Russia ratified the CTBT in 2000 but revoked ratification in November 2023. China has signed the treaty but is believed to be waiting for the United States to ratify it first. But like the United States, China and Russia claim compliance with the treaty in line with the international legal principle that once a country signs a treaty (even if it fails to ratify it) it is not supposed to defeat the “object and purpose” of the treaty.

Russia and China, however, likely interpret the Clinton administration’s insistence on a zero-yield treaty differently than the United States has. This is because zero-yield has no technical basis since plutonium fissions by itself. In fact, the treaty language accepted by the UN General Assembly on September 10, 1996, makes no mention of zero-yield. It only states, “Each State Party undertakes not to carry out any nuclear weapon test explosion or any other nuclear explosion.” Although the treaty negotiating records show that Russian and Chinese officials agreed to zero-yield, China was not interested in hydronuclear experiments at the time because it did not have the technical sophistication to benefit from them. (It certainly does now.) My Russian counterparts at the time, meanwhile, complained that zero-yield made no sense and is unverifiable. They did not consider hydronuclear experiments disallowed. After all, the United States and the Soviet Union justified hydronuclear experiments during the 1958–61 moratorium because they were not considered to be “nuclear explosions.”

The question of whether China and Russia are conducting these experiments while the United States is not will continue to be contentious until the zero-yield issue is fully resolved. Hydronuclear experiments should either be approved by the five established nuclear powers, making it possible for the United States to conduct them, or be specifically disallowed with adequate provisions for verification.

Resumption of full-scale nuclear testing, meanwhile, would allow the United States to answer some pressing questions about aging and remanufactured plutonium pits. But it would help China and Russia even more. Russian President Vladimir Putin, for instance, has been brandishing new, exotic weapons systems, such as a nuclear-powered, nuclear-armed cruise missile and the huge Poseidon nuclear-powered, nuclear-armed torpedo. Testing would likely give Russia more confidence in these systems.

China, however, would likely learn the most. Nuclear testing would augment what it learned from the few tests it was able to conduct in the 1990s on the weapon systems that make up its arsenal today. It would also help in fielding the greatly expanded nuclear arsenal that it is developing now. With the extensive activities observed at the Chinese and Russian nuclear test sites in recent years, it is also likely that both countries could resume full-scale nuclear testing much more quickly than could the United States. A resumption of full-scale testing by any of the three could also allow India, North Korea, and Pakistan to overcome political barriers to resume testing.

The United States has the richest test history. It has conducted 1,054 nuclear tests (24 of them with the United Kingdom). Russia has conducted 715; France, 210; the United Kingdom, 45; China, 45; India, 6; Pakistan, 6; North Korea, 6. What the United States does not have, however, is a nuclear weapons production complex, especially for plutonium pit production. That doesn’t require nuclear testing. It requires the national will to make it happen. The last 30 years give me little confidence the United States can do so.

My greatest concern about resuming full-scale nuclear testing is that it will fuel another dangerous arms race at a time when global tensions among the great powers are high. Engaging in another arms race is contrary to Trump’s comment that “it would be great if we could all denuclearize, because the power of nuclear weapons is crazy.”

Instead of suggesting an immediate return to nuclear testing, then, Trump should focus on returning to arms control measures to ensure strategic stability with Russia and with China. Hopefully, these measures would lead to a reduction in U.S. and Russian nuclear forces and reduce incentives for China to increase its arsenal. For nuclear testing, he should help erect the highest possible barriers for any country to test by leading an effort to ratify the CTBT. To settle the question of evasion of low-yield tests or hydronuclear experiments, the president and his counterparts in Beijing and Moscow would need to show the political will to agree on a verifiable low-yield limit. That will almost surely require onsite inspections, which were demonstrated to be possible in 1988.

The bottom line is that even though the United States could derive important benefits from resumed nuclear testing, it would lose more than it stands to gain.