The Machine Gun
On November 27, 2020, Mohsen Fakhrizadeh was driving on a highway east of Tehran when a Nissan pickup truck parked on the shoulder opened fire. There was no one in the truck. The weapon — a 7.62mm machine gun mounted in the cargo bed, operated by remote satellite link — fired thirteen rounds in under a minute.1 Fakhrizadeh’s wife, sitting beside him, was unharmed. He was dead before his bodyguards could return fire.
Fakhrizadeh was the most protected scientist in the country. He traveled with armed escorts. His schedule was classified. His home address was a state secret. The operation that killed him required penetrating Iran’s internal security apparatus at every level: identifying his route, positioning the weapon, establishing a satellite communication link, and firing with enough precision to kill the passenger and spare the driver.
The assassination was attributed to Israel’s Mossad. Iran’s response was public fury and private acceleration. Within months, enrichment climbed past 20 percent and then to 60 — a technical hair’s breadth from weapons-grade. The man who had been leading weapons research was dead. The program he oversaw was more advanced than ever.
What Fakhrizadeh had been building — and what this article examines alongside the campaign to stop it — would later be documented in extraordinary detail. The Iranian nuclear archive seized by Mossad in 2018 revealed he was the architect of Project Amad: a plan to build five nuclear warheads, each with a 10-kiloton yield, for integration with Shahab-3 ballistic missiles. Project Midan identified five underground test sites. The program included warhead designs, uranium metallurgy, and neutron initiator development.2 But before the sabotage campaign, there was a longer, larger concealment campaign — and the regime’s shadow came first.
The Archive
On January 31, 2018, fewer than 24 Mossad agents entered an industrial warehouse in the Shorabad district of southern Tehran. Working within a 6.5-hour window before the morning security shift, they breached iron doors, cut through safes with torches burning at 2,000°C, and extracted roughly half a ton of material from 6 of 32 safes: approximately 50,000 pages and 183 CDs documenting Iran’s nuclear weapons program.3
The archive documented Project Amad in granular detail: warhead designs for an implosion-type device, uranium metallurgy experiments, plans for a previously unknown underground tunnel facility for uranium metal weapon components (“Project 110”), high-explosive testing protocols, work on uranium deuteride neutron initiators, and the five test sites of Project Midan. A Harvard Belfer Center team that reviewed the documents found evidence that the weapons decision involved then-President Khatami, Security Council Secretary Hassan Rouhani, and Defense Minister Shamkhani, with Supreme Leader Khamenei reportedly informed.4
The program had not simply vanished in 2003 when it was formally “halted.” Archive documents from August–September 2003 meetings showed internal discussions “dominated by carrying out official orders to downsize and not build nuclear weapons but preserve the assets and many of the activities.”5 Key personnel continued working under successive organizational covers: first SADAT (2004–2008), then Malek Ashtar University, and finally SPND — the Organization of Defensive Innovation and Research — from approximately 2011 until Fakhrizadeh’s assassination. As David Albright of the Institute for Science and International Security concluded: “Iran’s decision to halt the Amad Plan merely served as a tactical retreat, not an abandonment of its nuclear weapons ambitions or activities.”
The IAEA used archive leads to identify previously unknown sites — and found physical corroboration. Uranium particles of anthropogenic origin appeared at Turquzabad, Varamin, and Marivan, exactly where the archive indicated nuclear material had been present. Iran denied everything. Iran’s UN mission called the archive “laughably absurd.” The particles said otherwise.
The Concealment
The archive answered a question that the shadow war often obscures: which shadow came first?
The standard narrative places the West’s covert campaign at the center — Stuxnet, assassinations, sabotage — as though Iran’s nuclear program was a transparent civilian effort attacked by rogue intelligence services. The technique that makes this framing work is what might be called “lessons taught by the West” — converting the regime’s own deliberate choices into consequences of Western actions, as though building secret nuclear facilities were a lesson Iran had no choice but to learn.
The actual record reverses the sequence. Iran’s concealment campaign preceded the sabotage campaign by nearly two decades. Here is what it looked like.
The Concealment Record:
Facility Secret Duration How Discovered Iran’s Response Natanz enrichment ~17 years (1985–2002) MEK press conference (Mossad intelligence) Admitted after exposure Kalaye Electric ~5 years (1997–2002) IAEA follow-up after Natanz 3 story changes; walls remodeled before access; 70% enriched particles found Fordow / Al Ghadir ~3 years (2006–2009) Western intelligence, revealed at G-20 Claimed legal notification compliance; IAEA disagreed Lavizan-Shian ~3 years (2000–2003) NCRI exposure, May 2003 Buildings razed, roads removed, 140 metric tons topsoil stripped; replaced with city park Parchin ~15 years of access denial (2004–2015) Western intelligence, IAEA 2004 request 6+ access denials; satellite imagery showed water flowing from building, demolitions, new paving Turquzabad warehouse ~9 years (2009–2018) Mossad archive leads Site sanitized after Netanyahu revelation; IAEA still found uranium particles Marivan test site Unknown Mossad archive leads IAEA found evidence of “full-scale hemispherical implosion systems” — nuclear weapon geometry
Not a single major facility was discovered by the IAEA through its own verification activities. Every revelation came from external intelligence — an opposition group’s press conference, three governments’ intelligence services, or a foreign intelligence raid. The IAEA’s system caught what it was designed to catch: diversion of declared material at declared facilities. It missed everything the regime chose not to declare.6
The concealment was not passive. At Lavizan-Shian, the regime stripped 140 metric tons of topsoil to eliminate environmental evidence and replaced the site with a park. At Parchin, satellite imagery documented systematic sanitization over three years: water flowing from the building of concern, two support buildings demolished, heavy machinery tracks, new roofing, asphalt paving, and pink polyethylene sheeting consistent with decontamination. When inspectors finally accessed Parchin in September 2015 — under a controversial arrangement where Iranian technicians, not IAEA inspectors, collected some samples — they still found chemically man-made particles of uranium.7
Build. Deny. Sanitize. Redefine compliance. This was the regime’s pattern for twenty years before anyone fired a cyberweapon or attached a magnetic bomb to a scientist’s car.
The Worm
The most sophisticated act of sabotage in the history of warfare was discovered by accident — and it targeted a facility that had been concealed for seventeen years.
In June 2010, a Belarusian cybersecurity firm called VirusBlokAda identified a worm spreading through Windows systems via infected USB drives. Analysis revealed something unprecedented: the malware was not designed to steal data or ransom networks. It was designed to destroy physical equipment.
Stuxnet — the name given by researchers — targeted the Siemens Step 7 software running programmable logic controllers at Natanz. Specifically, it targeted the controllers managing spin speeds of IR-1 centrifuges. The worm’s payload was elegant and vicious: it caused centrifuge rotors to fluctuate between speeds that would shatter them, while simultaneously feeding false “normal” readings to control room monitors. Iranian engineers watched their screens show stable operations while centrifuges were tearing themselves apart in the basement.8
The operation — codenamed “Olympic Games” and conducted jointly by the United States and Israel — was the first known instance of physical destruction achieved through cyber warfare. Estimates suggest Stuxnet destroyed approximately 1,000 centrifuges at Natanz. The worm’s design revealed an intelligence capability as alarming as its destructive power: someone had obtained the exact specifications of Iran’s centrifuge control systems in an air-gapped facility — meaning the initial infection required a physical agent inside the plant.
The strategic impact was temporary: Iran replaced the centrifuges and hardened its systems. The psychological impact was permanent: Iran knew that its most sensitive facilities had been penetrated at the software level by an adversary capable of physical destruction without firing a shot.
Context matters. Stuxnet targeted a facility that Iran had built in secret, never declared to the IAEA, and operated for years in violation of its NPT safeguards obligations. The cyberweapon set a dangerous precedent for state-sponsored cyber-physical attacks that extends far beyond Iran — AND it targeted a facility concealed for seventeen years in violation of international law. Both facts belong in the same sentence.
The Scientists
The assassinations followed a pattern. A motorcycle pulls alongside a car in Tehran traffic. A hand reaches out and attaches a small magnetic device to the driver’s door. The motorcycle accelerates away. Seconds later, the device explodes.
The Assassinations:
Date Scientist Role Method Jan 2010 Masoud Ali-Mohammadi Quantum physicist, Tehran University Remote-controlled motorcycle bomb outside his home Nov 2010 Majid Shahriari Key enrichment figure Magnetic limpet bomb attached to car door in Tehran traffic Jan 2012 Mostafa Ahmadi Roshan Deputy director of commerce, Natanz Magnetic car bomb Nov 2020 Mohsen Fakhrizadeh Head of AMAD program Satellite-controlled machine gun June 2025 Fereydoun Abbasi-Davani Former AEOI head, laser physicist Airstrike (Operation Rising Lion) June 2025 Mohammad Mehdi Tehranchi Physicist Airstrike (Operation Rising Lion)
Each assassination is both morally fraught and operationally specific. These were extrajudicial killings — conducted without legal authorization, congressional oversight, or judicial review, accountable to no electorate. If the legitimate exercise of authority requires lawful process, the assassinations failed that test.
They were also targeted at a concealed weapons program. Fakhrizadeh was not a dissident physicist pursuing civilian research. He was the architect of a plan to build five nuclear warheads, running a program his government denied existed while the IAEA found uranium particles at sites the archive identified. The moral calculus is genuinely difficult — and the difficulty is the point. Acknowledging it is more honest than either the “heroic Mossad” narrative or the “murdered civilian scientists” narrative.
The 2025 operations marked a qualitative shift. During Operation Rising Lion, Israel’s twelve-day air campaign, at least eleven nuclear scientists were killed9 — no longer covert assassinations deniable by all parties, but open military targeting of technical personnel. The institutional memory of the program was deliberately attacked alongside its physical infrastructure.
The Detection Gap
The shadow war — both the concealment campaign and the sabotage campaign — exposes a structural vulnerability in nuclear verification that no agreement has resolved.
The IAEA’s system verifies declared material at declared facilities. It counts what states report and checks that nothing declared is missing. When both a Comprehensive Safeguards Agreement and Additional Protocol are in force, the system is strong — environmental sampling can detect uranium at sub-trillionth-of-a-gram levels. But a determined state can advance significant weapons capability through activities that never trigger safeguards.
Computer modeling of implosion dynamics. High-explosive testing with non-nuclear surrogates. Engineering work on payload integration and re-entry vehicles. Procurement of dual-use equipment — high-speed cameras, electronic switches, flash X-ray systems. Metallurgical research using materials that mimic uranium’s properties. Development of exploding bridgewire detonators. None of these involve nuclear material. None produce radioactive signatures.
The JCPOA’s Section T nominally prohibited such activities. But as former IAEA Deputy Director General Olli Heinonen noted, quarterly reports provided no evidence of meaningful first-hand verification. The IAEA had not visited any military site since Implementation Day. The US intelligence community assessed that Iran would need approximately one year to complete weaponization steps that do not involve fissile material production — a timeline that existed independently of any enrichment constraint.10
The technique has a name: compliance theater — satisfying verification at declared sites to create the impression of transparency while advancing undeclared work where no inspector will look.
The PMD investigation — the “possible military dimensions” file — illustrated the pattern. Of twelve areas of concern identified in 2011, only two were even partially resolved by the 2015 Final Assessment. On ten issues, Iran either refused to discuss, denied everything, or offered explanations the IAEA could not verify. The file was politically closed to enable JCPOA implementation. The regime had complied where detection was certain. It had concealed where detection was not.
Imagine discovering that the country you’ve been negotiating with for a decade had a parallel weapons program hidden behind every concession — and that every major revelation came not from the inspectors in the room, but from a spy agency that broke into a warehouse.
The Honest Ledger
The shadow war is two stories that most accounts tell as one.
The first story is the sabotage campaign — Stuxnet, the assassinations, Operation Rising Lion. It was extrajudicial, legally unauthorized, and set precedents that extend beyond Iran. Stuxnet established that states could conduct cyber-physical attacks on sovereign infrastructure. The assassinations established that scientists could be killed on foreign soil without consequence. The 2025 strikes destroyed infrastructure but left knowledge intact and inspectors expelled. Each operation achieved its tactical objective. None resolved the underlying problem.
The second story — less often told — is the concealment campaign the regime ran first. Seventeen years of covert construction at Natanz. Fordow built inside a mountain from a weapons program blueprint. Lavizan-Shian’s buildings razed and replaced with a park. Parchin sanitized over three years while inspectors waited. Five nuclear warheads designed, documented, and preserved through successive organizational shells. Uranium particles at secret warehouses the IAEA never knew existed.
The sabotage campaign was morally costly. The concealment campaign was the context in which those costs were incurred. Both belong in the same ledger.
The regime’s preferred narrative casts these as lessons taught by the West — as though building Fordow inside a mountain were the inevitable response to Natanz being exposed, as though preserving a weapons archive were the natural consequence of sanctions. But the concealment preceded the sabotage. The covert facilities came first. The weapons research came first. The regime’s shadow was older, larger, and more systematic than the one cast by Stuxnet and magnetic bombs.
The concealment was also a betrayal of Iran’s own people. The regime claimed to represent Iranian sovereignty while building a secret weapons program that exposed 93 million civilians to the risk of military strikes, international isolation, and economic devastation — without their knowledge or consent. The Iranian people’s aspiration for national dignity was hijacked by a regime that needed confrontation with the West more than it needed the trust of its own citizens.
The regime has never voluntarily disclosed a significant nuclear activity. Not once. Every facility was found by someone else — an opposition group, a foreign intelligence service, a Mossad raid. Whatever comes next must contend with that fact.
This article is part of The Nuclear Pursuit. For the American genesis of the nuclear program, see Atoms for the Shah. For the deal that was supposed to constrain the program, see Heroic Flexibility.
Footnotes
-
United States Institute of Peace, “Assassinations of Iran Nuclear Scientists,” Iran Primer, December 2020 ↩
-
Mossad nuclear archive (2018); Institute for Science and International Security, AMAD warhead design analysis; Harvard Belfer Center review, 2019 ↩
-
Institute for Science and International Security, archive seizure documentation; Harvard Belfer Center authentication review ↩
-
Harvard Belfer Center, “The Iran Nuclear Archive,” expert review team including Matthew Bunn and William Tobey, January 2019 ↩
-
Institute for Science and International Security, “Iran’s Perilous Pursuit of Nuclear Weapons,” archive meeting minutes analysis ↩
-
IAEA Board of Governors reports, 2003-2025; Arms Control Association, “IAEA Investigations of Iran’s Nuclear Activities” ↩
-
Institute for Science and International Security, Parchin satellite imagery analysis series, 2012-2015; Iran Watch, Parchin inspection arrangements ↩
-
David Albright and Andrea Stricker, Institute for Science and International Security, “Iran’s Gas Centrifuge Program: Taking Stock” ↩
-
Times of Israel, “IDF Names 9 Iranian Nuclear Scientists Killed in Simultaneous Opening Strikes,” June 2025 ↩
-
Congressional Research Service, Iran’s nuclear weaponization timeline assessment; Institute for Science and International Security, Section T verification analysis ↩