There is a specific kind of public-money failure that does not show up as a boondoggle, a fraud, or a cost overrun on a building, because the thing that was lost was not money or steel. It was knowledge. In the 1980s the United States manufactured a classified material for its nuclear weapons, then shut the production line down at the end of the Cold War. Two decades later, when it needed to make the material again, it discovered that it no longer knew how. The people who had made it were gone, the process had not been written down in usable detail, and the government had to spend years and roughly 92 million dollars reverse-engineering a substance it had invented itself. The material is code-named FOGBANK, and its story is the clearest illustration you will find that institutional memory is a national asset, and that letting it evaporate has a price.
What FOGBANK is, and what cannot be said
Start with the honest boundary of this piece, because most of the interesting detail about FOGBANK is classified and will stay that way. FOGBANK is the code name for a material used in American thermonuclear warheads. Publicly, the only sanctioned description is that it is an interstage material, the substance that sits between the fission primary and the fusion secondary stages of a two-stage weapon, and outside experts believe it is an aerogel-like substance (The War Zone). Its exact chemical composition and its precise function remain secret, and the National Nuclear Security Administration has confirmed as much. So this article does not tell you what FOGBANK is made of, because that is not public, and anyone who claims to know the recipe is guessing. What is public, and thoroughly documented by the government's own auditors, is what happened when the country tried to make it a second time.
The material is discussed publicly almost entirely in connection with the W76, the warhead carried on the Navy's Trident submarine-launched missiles, and its refurbishment program. FOGBANK was originally produced at the Y-12 National Security Complex in Oak Ridge, Tennessee, running from roughly the mid-1970s until about 1989. When Cold War warhead production wound down, the full-scale FOGBANK facility was deactivated and eventually decommissioned, leaving only a small pilot plant behind (public record summarized here). At the time, mothballing a production line for a material the country had plenty of looked like prudent housekeeping. It turned out to be the setup for an expensive lesson.
The knowledge that walked out the door
In the 2000s the W76 came due for a life-extension program, the process by which an aging warhead is refurbished to keep it safe and reliable for another few decades. The refurbished version, the W76-1, needed new FOGBANK. And when the National Nuclear Security Administration went to make it, it found that the capability had effectively been lost.
The Government Accountability Office documented what "lost" meant in concrete terms. Few detailed process records had been kept from the 1980s production run, and nearly all of the staff with hands-on manufacturing expertise had retired or moved on (GAO-09-385). This was not total amnesia, a pilot plant and some records survived, but the practical, at-scale know-how of how to actually produce the material to specification was gone. The agency had to re-establish the capability essentially from scratch and build a new production facility at Y-12 to do it. The auditors were blunt that this reflected poor management of institutional knowledge, an asset the government had let depreciate to zero without noticing.
The impurity that turned out to be the point
The most remarkable part of the story is what happened when they finally remade the material with modern methods, and it is a small parable about the danger of improving something you do not fully understand. The new, cleaner, more precisely controlled process produced FOGBANK that did not work correctly.
A root-cause investigation eventually traced the failure to an impurity. The original Cold War material had contained a contaminant that the purer modern process had inadvertently stripped out. That impurity, it turned out, subtly altered the material's structure in a way the weapon quietly relied on, and yet it had never been explicitly documented, specified, or controlled. The people who made FOGBANK in the 1980s had been producing an essential ingredient without knowing it was essential, because it rode along for free in their less-refined process. When a later generation cleaned up the process, they cleaned out the thing that made it work. Once its role was understood, the team reintroduced the impurity at a controlled concentration as a deliberate, monitored step (The War Zone). The chemical identity of that impurity is not public, and the point of the story does not depend on it: a critical property of a nuclear weapon component had been an undocumented accident, and rediscovering the accident took years.
What it cost, and what it delayed
The government's auditors put numbers on the failure. The Government Accountability Office attributed 69 million dollars in cost overruns on the W76 program specifically to the FOGBANK manufacturing problems (GAO-09-385). Public accounts commonly cite a larger total of roughly 92 million dollars to re-establish production, about 23 million spent exploring alternative materials that never panned out plus the roughly 69 million to reproduce the original, a total that is a secondary aggregation rather than a single GAO line (public record). Either way, relearning how to make a material the country had already invented cost tens of millions of dollars.
It also cost time, at an awkward place in the schedule. The Government Accountability Office found the FOGBANK trouble, together with an unrealistic timeline for the new production facility, delayed the first refurbished W76 by at least a year, slipping from September 2007 to September 2008 and creating logistical headaches for the Navy that carries the warhead (GAO-09-385). The facility itself started about a year late, the auditors found, because its construction schedule was unrealistic, there were disagreements over safety guidelines for handling the flammable solvent used in the process, and the W76 program manager did not have the authority to manage the construction. The broader W76-1 life-extension program was eventually completed in December 2018, having refurbished on the order of 800 warheads (public record). The material problem was solved. It was just solved late and at a premium, for want of records and retained expertise.
The ledger reading
FOGBANK is not a scandal, and it is not really a story about nuclear weapons, though the setting is dramatic. It is a story about the depreciation of knowledge as a hidden line on the public ledger. The government owned a manufacturing capability, a genuinely valuable and hard-won thing, and it let that capability lapse by mothballing a plant, keeping thin records, and losing the people, all of which looked like reasonable economy at the time. The bill for that economy came due twenty years later, denominated in tens of millions of dollars and a year of schedule, plus the strange months spent rediscovering that an undocumented impurity was the secret all along.
The lesson generalizes far beyond warheads. Any organization that stops doing a complex thing, and does not deliberately preserve how it did it, is quietly running the same risk: that the tacit knowledge in people's hands and habits, the part that never made it into the manual, will be gone when it is needed again, and that recovering it will cost far more than keeping it would have. FOGBANK is the version of that lesson with the highest stakes and the clearest paper trail, but it is not unique. It is what happens, everywhere, when the people who know how leave before anyone writes it down.
Related reading
- The MOX facility: a $7.6 billion hole in the ground: the nuclear enterprise's very different failure, a first-of-a-kind plant killed for cost.
- The Nuclear Waste Fund and Yucca Mountain: a repository paid for and never built, read through its two separate money pools.
- The working ledgers: the series that reads a public program by following what it actually cost.
Fact-check notes and sources
This piece describes only publicly documented facts. FOGBANK's chemical composition, its precise function, and the identity of the essential impurity are classified and are not asserted here.
- What FOGBANK publicly is (a classified interstage material in the W76 warhead, believed to be aerogel-like, with composition and function classified; originally produced at Y-12 in Oak Ridge from roughly the mid-1970s to about 1989, then deactivated except for a pilot plant): The War Zone and the public record. The W76 connection is documented; associations with other warheads are open-source speculation.
- The knowledge loss and the impurity (the W76-1 life-extension program's discovery that the manufacturing capability had lapsed, with few records kept and expert staff gone, requiring a new facility; and the root-cause finding that an impurity removed by a cleaner modern process had been essential and had to be reintroduced under control): GAO-09-385 and The War Zone.
- The cost and schedule (the 69 million dollars in cost overruns the GAO attributed to FOGBANK, the commonly cited roughly 92 million dollar total as a secondary aggregation, the at-least-one-year slip from September 2007 to September 2008, the facility's late start and its management findings, and the December 2018 completion of the broader W76-1 program covering on the order of 800 warheads): GAO-09-385 and the public record. The 69 million dollar figure is the GAO's; the 92 million dollar total is a secondary press aggregation in 2008 to 2009 dollars.
This post is informational and journalistic, describing publicly available records about a government program. It is not legal, financial, or policy advice, and it contains no classified information. Figures are drawn from Government Accountability Office reports and reputable reporting, with secondary aggregations and approximate dates labeled as such.
