By Rana Agahi
When a country gives up or loses a nuclear weapons program, what actually disappears?
Unlike a warhead, laboratory or centrifuge, knowledge cannot simply be dismantled. Nuclear programs depend not only on machines and materials, but also on the tacit knowledge scientists gain through experience and cannot always pass on through instructions alone.
In June 2025, Israel and the United States launched a broad military campaign against Iran that included strikes on nuclear facilities and nuclear personnel. The International Atomic Energy Agency (IAEA) reported extensive damage across several nuclear sites, including Fordow, Natanz and Isfahan, and eventually withdrew its inspectors from Iran after the military attacks made it unsafe to continue in-field verification and monitoring. The attacks succeeded in damaging Iranian nuclear sites but did little to destroy the nuclear knowledge developed in Iran.
Israel also targeted the people behind the program. At least 14 Iranian nuclear scientists were reported killed during the attacks. Israel’s ambassador to France, Joshua Zarka, argued that eliminating senior scientists would significantly damage Iran’s ability to rebuild its nuclear capabilities. But nuclear experts also pointed out that Iran had other scientists capable of continuing the work, even if replacing more experienced researchers would take time.
Leaving aside the moral question of targeting scientists, killing individual experts is not the way to eliminate nuclear expertise. It may temporarily slow technical work, but it does not dismantle the broader scientific system that produced them or erase the knowledge they already passed on to others.
Leaving aside the moral question of targeting scientists, killing individual experts is not the way to eliminate nuclear expertise. It may temporarily slow technical work, but it does not dismantle the broader scientific system that produced them or erase the knowledge they already passed on to others.
Iran had already demonstrated this before the 2025 attacks. After the November 2020 assassination of nuclear scientist Mohsen Fakhrizadeh, Iran’s parliament approved legislation requiring the government to increase uranium enrichment to 20 percent, expand the use of advanced centrifuges and reduce some IAEA monitoring if sanctions relief did not occur. Rather than eliminating Iran’s nuclear knowledge, the assassination was followed by political decisions that expanded parts of the program.
If facilities can be destroyed and individual scientists removed without eliminating the knowledge behind a nuclear program, non-proliferation policy must address what happens to that surviving expertise.
The collapse of the Soviet Union provides one possible solution.
After the Soviet Union collapsed in 1991, an enormous weapons complex was left behind across newly independent and economically fragile states. Along with nuclear weapons, materials and facilities were thousands of scientists, engineers and technicians who had spent their careers working within the Soviet nuclear enterprise. U.S. policymakers worried that unemployment and economic instability could make some of those experts vulnerable to recruitment by governments or organizations seeking weapons of mass destruction.
The U.S. response was the Nunn–Lugar Cooperative Threat Reduction program and related non-proliferation efforts. These programs worked to secure and dismantle weapons and materials, but they also recognized that the people behind the former Soviet weapons complex presented a different kind of proliferation problem. Programs supported civilian scientific research and peaceful employment for former weapons scientists rather than treating physical dismantlement as the end of the process. The International Science and Technology Center, for example, was established in 1992 in part to provide former Soviet weapons scientists with opportunities for sustainable, peaceful employment.
In practice, this meant treating continuing employment as part of non-proliferation policy. The goal was not to make scientists forget what they knew. It was to reduce the possibility that economic desperation would create an incentive to sell that knowledge elsewhere by giving scientists another place to use their expertise.
The goal was not to make scientists forget what they knew. It was to reduce the possibility that economic desperation would create an incentive to sell that knowledge elsewhere by giving scientists another place to use their expertise.
Nunn–Lugar therefore approached the problem differently. Instead of asking how nuclear knowledge could be destroyed, it asked how the conditions surrounding the people who possessed that knowledge could be changed.
Abdul Qadeer Khan is, however, an important complication to that argument.
A.Q. Khan was a central part of Pakistan’s nuclear program, leading its uranium-enrichment effort using gas centrifuges and helping establish the laboratory at Kahuta. He later became the center of an international proliferation network that moved nuclear technology and expertise across borders. For this reason, Khan is considered an international criminal, except in his home country, where he is accounted something of a hero.
Khan and his associates supplied Iran and Libya with centrifuge components, designs and, in some cases, complete centrifuges. The United States also concluded that his network supplied centrifuge designs, equipment and technology to North Korea, while Libya received nuclear weapon design information. The Pakistani government denied any connections with Khan’s efforts, but as it benefitted from them, it is widely supposed that Khan benefitted at least from government awareness of what he was up to.
It is reasonable to ask whether stopping Khan earlier could have prevented at least some of the technology and expertise that later reached Iran, Libya and North Korea.
But Khan’s own network also shows why killing one scientist would not necessarily have solved the larger problem. By the time the network was operating internationally, Khan was supported by a much wider system of suppliers, manufacturers, financiers, shipping agents and procurement specialists, with government connections, spread across several countries. Even after Khan himself was removed from the operation, governments still had to identify and prosecute associates, shut down companies, strengthen export controls, track shipments and prevent the remaining pieces of the network from reconstituting themselves. Khan himself also benefited from knowledge and assistance that did not originate with him. A declassified U.S. government document noted that Pakistan’s nuclear effort drew on external suppliers and reported that China provided Khan’s program with important nuclear assistance in the early 1980s.
Killing Khan in 1983 might have prevented some later transfers, but it is likely that another would have taken his place. What is certain is that it would not automatically have eliminated the larger conditions that made proliferation possible, such as Pakistan’s nuclear institutions, other trained scientists and engineers, foreign suppliers, procurement networks, and the technical information Khan had already acquired and shared.
These cases show why non-proliferation cannot focus only on what can be physically destroyed. If bombing nuclear sites and killing scientists cannot eliminate the knowledge behind a nuclear program, then those actions cannot be the only steps taken. For Iran, part of the answer will still have to come through diplomacy, whether that means negotiating a new version of the Joint Comprehensive Plan of Action, better known as the Iran nuclear deal, or developing another agreement that places meaningful limits on the country’s nuclear program. But diplomacy also has to account for the people who already possess the technical knowledge needed to sustain or rebuild that program. This is where the Nunn–Lugar experience offers a useful lesson. Because reducing proliferation risk is not only about dismantling weapons and facilities, but also about giving scientists a reason and an opportunity to use their expertise elsewhere.
Non-proliferation cannot focus only on what can be physically destroyed.
The United States could build on that idea by using its relationships with Gulf states and the technical expertise of the Department of Energy to support a regional research center where Iranian and Gulf scientists work together on transparent, peaceful applications of nuclear science. Such a center would not eliminate proliferation risk, and it would not replace the need for a broader nuclear agreement. It could, however, create more transparency around how nuclear expertise is being used while giving scientists productive alternatives to weapons-related work. Bombing may damage a program in the short term, but without a political and technical alternative, the knowledge needed to rebuild it remains. If nuclear knowledge cannot simply be destroyed, non-proliferation policy should do more to keep that expertise visible, engaged and directed toward peaceful purposes.
