The architect of stockpile stewardship

Victor Reis shares his experiences at the intersection of policy and science.

By Whitney Spivey | December 1, 2020

Victor Reis Opt
Victor Reis, pictured here in front of the Laboratory’s National Security Sciences Building, spoke at Los Alamos during the Lab’s 70th anniversary celebration in 2013. Los Alamos National Laboratory

The United States stopped testing nuclear weapons in 1992. In August 1993, Victor Reis became assistant secretary for Defense Programs in the Department of Energy (DOE) and was responsible for developing a new way for the United States to maintain its aging nuclear stockpile. Reis was instrumental in creating a science-based stockpile stewardship program, which uses computer simulations and nonnuclear experiments to evaluate the health and extend the lifetimes of America’s nuclear weapons. The program, which relies heavily on computing modeling and simulation, has spurred the growth of several generations of supercomputers.

Reis sat down with NSS to talk about the creation of stockpile stewardship, specifically how he was able to bridge the gap between scientists and policymakers.

Reis Speaking
Reis speaks about stockpile stewardship to NNSA employees in June 2019. Photo: NNSA

How do policymakers learn to trust the scientific community? How do they trust that scientists don’t have a hidden agenda or that scientific data isn’t skewed by funding sources? And how do scientists deal with that lack of trust by policymakers?

"The weapons labs became the U.S. computing companies’ biggest customers and subsequently drove the whole high-performance computing industry."- Victor Reis

A good deal of my career has been based on getting scientists and policymakers to trust each other, particularly during my service in each one of the “prune positions”—what diplomat and author John Trattner called the toughest, presidentially appointed science and technology jobs in Washington, D.C.

In August 1993, I was confirmed in my fourth “prune position”—assistant secretary of energy for Defense Programs. President Clinton had just announced that the United States would maintain a moratorium on nuclear testing while seeking a Comprehensive Nuclear-Test-Ban Treaty (CTBT). Because nuclear testing had been an essential and dominant element of the DOE weapons program, the DOE weapons labs (Los Alamos, Livermore, and Sandia) had traditionally opposed a comprehensive nuclear test-ban treaty. I found myself stuck in the middle: a member of what President Eisenhower in his famous farewell address called the dangerous “scientific-technological elite” and in a position of being able to affect national policy.

The important thing was to bring the national laboratories together to help them look toward a future in which they wouldn’t be designing new weapons but rather maintaining old ones. I had to find a sweet spot between science, technology, and policy so the president’s CTBT objective and the Department of Defense’s deterrence mission could work together.

The solution was a science-based stockpile stewardship program that would maintain the effectiveness and safety of the current nuclear weapons without testing. It was a challenging scientific problem, so the laboratories really couldn’t help themselves because that’s what scientists love to do.

To start, I asked each of the three nuclear weapons laboratory directors to send me a senior weapons designer and a senior scientist, for an offsite, “truth-be-told” meeting. John Immele and John Browne were the Los Alamos participants. The laboratories had been created to compete with one another—not for money but to design and build the next nuclear weapon. I told them, well, we’re not likely to build any more weapons, and we’re not likely to do any more testing. Our new strategy will be based upon new experiments and computer simulations that can validate the health of our aging weapons. At that time the necessary computing power was not available—not by a long shot—so the weapons programs had to strengthen their own computing groups, working with IBM and other major computing companies to develop a new class of high-performance computer. The weapons labs became the U.S. computing companies’ biggest customers and subsequently drove the whole high-performance computing industry.

Science-based stockpile stewardship put the laboratories in a powerful policy position because the responsibility of determining the health of the stockpile was theirs. If they ever believed that we had to go back to testing, they would have to say so. This is where scientific integrity comes in: you have to be able to say, no matter what the policy is, “This is what the science tells us.” You can’t be complacent. You can’t say, “We’ve done this already.” You need people saying, “Do we really understand?” You have to be able to criticize yourself all the time. I think the laboratories were originally skeptical about stockpile stewardship, but they were willing to give it their best shot if resources were available.

The Department of Defense, on the other hand, said “What do you mean you’re not going to test? That just doesn’t make any sense. We always test our airplanes before we fly them!” I went to the chairman of the Joint Chiefs of Staff, General John Shalikashvili, and suggested he look the laboratory directors in the eye and ask them directly, “Do you think you can maintain the stockpile without testing?” And he did that. In fact, he said that question should be asked every year as the weapons age and as we learn more. This became an official annual procedure; every year the labs assess the health of the stockpile.

This annual assessment process has been very important; it puts the laboratories’ scientific reputations on the line. The laboratories have to give their view, even if it goes against policy. And you have to ensure that the policymakers trust the scientists to give them the right technical answer, even if it’s not what they want to hear.

At the time, we had not really demonstrated that stockpile stewardship—maintaining effective and safe nuclear weapons using science and computing in the absence of nuclear testing—would work. Some thought we wouldn’t have enough data to be confident that our weapons are safe, secure, and effective. Some thought our confidence would erode over time. Maybe the continual questioning would cause potential adversaries to doubt our commitment to deterrence.

On the other hand, the policy risk must also be considered. If the United States resumes testing, then the Russians are going to test, and the Chinese are going to test. If that happens, would a return to testing improve our strategic posture? At what point do policy issues override the scientific uncertainty? The challenge is finding a balance.

Nearly 30 years later, science-based stockpile stewardship has actually evolved better than we expected. Many in the weapons community believe that the stockpile is more robust now than when we stopped testing—it’s more reliable, and additional safety and security features have been added over the years. The 2018 Nuclear Posture Review reported that continuing to effectively extend the lifetimes of our nuclear weapons means that there’s no need to resume testing.

Nonetheless, the system is configured so that if the laboratories say we have to return to testing, they will be trusted by the policymakers. Maintaining trust in the laboratories is—and will continue to be—the key to the future of the stockpile.


In 1992, John Trattner wrote The Prune Book: The 60 Toughest Science and Technology Jobs in Washington. Reis held four of those jobs throughout his career:

  • Assistant Director for National Security and Space in the Office of Science and Technology Policy
  • Director of Defense Advanced Research Projects Agency
  • Director of Defense Research and Engineering for the Department of Defense
  • Assistant Secretary for Defense Programs for the Department of Energy