On July 22, 2026, the United States and Saudi Arabia signed a civil nuclear cooperation agreement — a Section 123 deal, in the language of the Atomic Energy Act — that gives the Kingdom a documented pathway toward enriching its own nuclear fuel. It was signed as Washington was actively bombing Iran to stop that country from doing the same thing. Arms control researchers flagged the contradiction within days. Senator Chris Murphy called it a mechanism that would “set off a nuclear race in the region.”
None of that is speculation. It’s the news, from nine days ago. What’s less discussed is that the enrichment technology sitting at the center of this deal has a paper trail running back to a captured Austrian engineer and a Cold War prison camp — and that the trail never once stayed where it was supposed to.
A German ordnance lab, not a Soviet invention
The centrifuge did not begin in Soviet hands. It began in the autumn of 1941, at Anschütz & Co., a gyroscope manufacturer in Kiel, where physicists Paul Harteck and Wilhelm Groth, together with engineer Konrad Beyerle, began building a uranium centrifuge under a Heereswaffenamt contract issued by Kurt Diebner — the same rival administrator whose Army Ordnance channel competed with Heisenberg’s Kaiser Wilhelm Institute faction for control of Germany’s wartime uranium effort. By 1943 the Kiel group had reached 5% enrichment. Allied bombing destroyed the facility in July 1944; Beyerle relocated south and merged his work with Harteck’s group in Freiburg and Kandern. The war ended before the design could be industrialized. The expertise didn’t end with it.
The centrifuge nobody built on purpose
Gernot Zippe was an Austrian engineer taken into Soviet custody after the war and put to work alongside German physicist Max Steenbeck at the Sukhumi Physical-Technical Institute — continuing, under Soviet direction, the same line of centrifuge research the Kiel group had already established. The USSR did not originate the centrifuge. It captured the men who understood it and continued their work as part of the broader postwar program that pulled roughly a million German and Austrian specialists into Soviet institutes, the same pattern that pulled German rocket engineers into Operation Paperclip on the other side of the ledger.
He was not released by the regime that built the system that held him. Stalin died in March 1953. Zippe’s release came in 1956 — timing that tracks a wartime Allied understanding permitting the Soviet Union to retain German specialists for up to ten years, and that coincided with a deliberate 1955 diplomatic offensive by the post-Stalin leadership: the Austrian State Treaty that May, restoring Austria’s full sovereignty, and Konrad Adenauer’s September visit to Moscow, which secured diplomatic recognition of West Germany in exchange for releasing the last German POWs and detained specialists. The record shows this wasn’t carelessness — declassified debriefings describe a deliberate “cooling-off” practice, moving specialists into routine, non-sensitive work for a period before release specifically so their knowledge of current programs would already be stale by the time they crossed back West. Zippe’s notes were confiscated on the way out. The apparatus treated the paper as the danger.
It wasn’t. Zippe reconstructed the design from memory once he reached the University of Virginia. A system built entirely around extracting value from a person has no real mechanism for imagining that the person might still generate new value once you’ve decided he’s empty — and that gap, not any single guard’s mistake, is what actually let the design out.
Freed of Soviet secrecy, Zippe published and refined the centrifuge design that now bears his name. The Zippe-type centrifuge became the technical foundation for civilian uranium enrichment across the non-Soviet world.
The other channel
Zippe’s route wasn’t the only one. Wilhelm Groth never left Germany at all — he continued centrifuge research openly at the University of Bonn through the 1950s. In 1953 he met with Admiral Álvaro Alberto, president of Brazil’s National Research Council, alongside Otto Hahn and Paul Harteck. The result: an agreement to ship three working centrifuges, plus supporting equipment from fourteen German companies, to the University of São Paulo. The U.S. High Commissioner in Germany, James Conant, tried to block the shipment. It went through anyway, in 1956 — the same year Zippe walked free.
The research continues every week.
Hidden history. Forgotten symbols. New articles before they disappear into the archive. Delivered once a week. No noise.
Join the dispatch →Prefer Telegram? Daily updates here.
Two channels, one origin. One ran through a Soviet institute and a decade of captivity. The other ran in the open, under West Germany’s own flag, with Otto Hahn’s name on the arrangement. Both left German-Austrian hands and reached beyond the West’s intended perimeter within a decade of the war’s end.
It became, specifically, the foundation for Urenco — the Anglo-Dutch-German enrichment consortium established in 1970 to give Western Europe an enrichment capability independent of the United States.
The engineer who walked out with the blueprints
Abdul Qadeer Khan worked at a Urenco subcontractor in the Netherlands in the early 1970s, with access to classified centrifuge designs. In 1975 he returned to Pakistan carrying that design with him. It became the technical backbone of Pakistan’s nuclear weapons program — and, over the following two decades, of Khan’s own private proliferation network, which sold centrifuge technology onward to Libya, North Korea, and Iran.
A single line of engineering, originated in a German ordnance lab, carried through a decade of Soviet custody, refined in the West, industrialized by a European consortium, and walked out the door by one man with a security clearance and a plan. It reached three more countries after that. This is not contested history — Khan confessed to the network’s existence in 2004, and the technology transfers to Libya, North Korea, and Iran are documented in IAEA and national intelligence assessments.
Where the line runs today
Which brings the thread back to July 22. The 123 agreement with Saudi Arabia doesn’t hand over centrifuges directly — it establishes the legal and commercial framework under which American firms (Westinghouse, Bechtel, BWXT, Centrus among them) become the Kingdom’s nuclear partners of choice, with an enrichment pathway built into the structure rather than excluded from it. Congressional reviewers and arms-control researchers have already flagged that this arrangement sits uneasily next to the “No Nuclear Weapons for Saudi Arabia Act,” introduced in the same Congress, which would require Saudi Arabia to formally renounce enrichment before any such deal takes effect.
Whether this deal produces a Saudi enrichment capability, and whether that capability stays civilian, is not something this piece will predict. What’s documented is the pattern underneath it: enrichment knowledge has never, in eighty years, stayed inside the border it started in. It began in a German ordnance lab in Kiel. It survived Allied bombing and a decade of Soviet custody, carried out in one engineer’s memory. It left West Germany a second time in the open, under Otto Hahn’s signature, bound for Brazil. It moved from a European consortium to Pakistan in one engineer’s luggage. It moved from Pakistan outward to three more capitals. Every one of those moves was, at the time, described as an aberration — a leak, a betrayal, an exception to an otherwise well-contained system.
Every containment failure in this chain shares the same shape. A system convinced it has extracted everything of value from a man, a document, or a program stops watching — not because it has grown gentler, but because it has stopped imagining what’s left. The centrifuge got out of Sukhumi the same way it got out of Urenco a generation later: not through a hole in the wall, but through the blind spot of an apparatus built to measure what it had already taken, not what a mind or a market might still produce.
Sources: Congressional Research Service (IF10799), U.S. Department of Energy, Al Jazeera, The Washington Post, CNBC — July 2026 reporting on the U.S.–Saudi civil nuclear agreement; IAEA and public reporting on the A.Q. Khan network.
