The question the Red Sea poses is not who supplied the weapons. It is who passed on the knowledge of using them at sea, and through which channels that knowledge still travels.
On 24 July 2026 the Saudi refinery at Jizan caught fire. The Yanbu terminal, through which roughly 92 % of Saudi seaborne crude exports pass, was hit. Two tankers had been struck the day before, and Abdul Malik al-Houthi had declared a naval blockade of Saudi Arabia. Crude went above a hundred dollars.
Ten years ago, the people who carried out those strikes were a Zaydi movement from northern Yemen, born in Saada out of a local grievance and in insurgency since 2004. They had no maritime capability then. They had none in 2015 either, when they took Sanaa. What happened in between is the only thing that lets anyone anticipate what comes next.
Three layers, and only one that really counts
Iranian support to Ansar Allah was built in strata. The commonest mistake is to confuse them.
The first is ideological. Through the 2000s the movement claimed kinship with the Islamic revolution, borrowed its repertoire, sent cadres to study in Qom. That layer produces no military capability, and it is the one media coverage describes most readily.
The second is material. After 2014, a movement turned territorial authority needs weapons and parts. This is the period documented by the UN Panel of Experts on Yemen: repeated seizures in the Arabian Sea of missile components, drones and guidance systems out of Iran. That layer fills depots. It trains no one.
The third changes everything, and it is the least told. From 2019 and 2020, support becomes a transfer of naval knowledge. Mining a strait. Guiding a surface drone onto a hull that is moving. Working out a firing window against currents and traffic. Sustaining a campaign against oil terminals. None of that is improvised, and no crate of weapons carries it across. It takes instructors, a curriculum, months.
The pipeline
That teaching has a location. According to material reported by the British press and by research into the Iranian security apparatus, a maritime sciences and technology academy at Zibakenar, on the Caspian coast of Gilan province, has taken in Yemeni recruits since January 2020. The course runs six months, it is reserved for foreign trainees, it sits under Quds Force authority, and the trainees are housed away from other students. The first intake is said to have numbered around two hundred. Wider figures circulate, of the order of more than a thousand fighters trained in IRGC camps, some two hundred and fifty of them at a Quds Force garrison in Hamedan. This material comes from open and opposition sources, it has not been independently verified on the ground, and DORNA reports it with that reservation.
The chronology, for its part, leaves little doubt: Houthi naval capability changes scale in the two years that follow. And July 2026 shows the pipeline still working. On the 13th, a flight linked Tehran to Sanaa carrying, according to sources cited by international agencies, between ten and twenty-one IRGC personnel, commanders and advisers, along with missile and drone-related equipment. Six years after the first Caspian intake, the direction of travel has reversed. It is no longer trainees going to Iran, it is instructors going to Yemen.
Where the knowledge is captured, upstream
That leaves a question almost nobody asks: where does the knowledge those instructors pass on come from. Part of it comes from Iranian operational experience in the Gulf, accumulated since the 1980s. Part of it comes from university research, and this is where one Iranian text deserves close reading.
In February 2021 the Supreme Council for the Cultural Revolution, chaired by the Supreme Leader and acting on the recommendation of the Ministry of Defence, adopted a Comprehensive Act on Science and Technology in the Defence and Security Field. It requires every Iranian university to hand over to the IRGC and the regime's armed bodies the intellectual property and output of its research. It lists the priority fields: automated and unmanned equipment, aerospace propulsion, artificial intelligence, and three directly maritime areas, guidance and navigation, logistics and vehicles, infrastructure. It names international university departments among the capacities to be drawn on. Its signatories are the Ministry of Defence, the IRGC, the Ministry of Intelligence and the Office of the Supreme Leader. The Council that adopted it has been under European Union sanctions since March 2023 and British sanctions since July 2023.
In those precise fields, international academic cooperation exists. The World Maritime University in Malmö, created by the International Maritime Organization, a United Nations agency, signed a five-year memorandum in November 2016 with Iran's Ports and Maritime Organization and Sharif University. A memorandum between Chabahar Maritime University and the Swedish institution was reported in 2017 by Iran's own Ministry of Roads and Urban Development and by Iranian media. These come from official Iranian sources.
Let us be plain. DORNA does not claim, and has nothing on which to claim, that a European academic programme trained Houthi fighters. That would be false. The point lies elsewhere, and it is more uncomfortable: since 2021, no Iranian institution can promise a foreign partner that the fruit of a joint project will stay civilian, because its own country's law forbids it from promising that. A representative of the Supreme Leader sits in every university, and a security department answering to the Ministry of Intelligence oversees travel, conferences and exchanges. The European partner is therefore not dealing with an autonomous scientific community. It is dealing with one link, and the end of the chain is written in a public text.
The blind spot is procedural before it is moral. European controls were built around dual-use goods, meaning around an object you can point at. They were not designed for a state that captures an entire national scientific output upstream, by law and without distinction. A course in hydrodynamics is not a dual-use good. A thesis on hull behaviour in shallow water is not exportable in any regulatory sense. Under the Iranian framework of 2021, both end up in the same place.
What the Houthis are worth to Tehran
That leaves the question of why Tehran invested six years in this pipeline. The answer lies in a cost to benefit ratio few military arrangements ever reach.
A few hundred men trained, a handful of air rotations, advisers and components: that is the investment. The return is measured on world markets. On 23 July, two tankers hit and a blockade declared were enough to push crude back above a hundred dollars. The next day, the fire at Jizan and the strike on Yanbu touched the infrastructure through which roughly 92 % of Saudi seaborne crude exports pass. Not one Iranian missile was fired, no Iranian territory was exposed, and deniability remains formally available.
The effects on the West go well beyond the oil price. A credible threat over the Red Sea and Bab al-Mandab pushes freight around the Cape of Good Hope, lengthens rotations, raises war risk premiums and feeds into European supply chains with a lag of several weeks. It ties down Western naval assets and escort budgets for an open-ended period. It erodes Egyptian transit revenue at the moment Cairo can least afford it. And it puts Gulf capitals in front of a dilemma they had spent four years avoiding, by reopening a front the truce with Riyadh had closed.
Above all, the sequence produced an immediate political result. After thirteen consecutive nights of strikes on Iranian territory, Washington suspended operations on 25 and 26 July, at the precise moment the war swung out to the Red Sea. What twelve days of bombing Iran had not achieved, a change of theatre achieved in forty-eight hours. The United Nations warned on 24 July against wider escalation, without holding any leverage over the actor concerned.
This is the definition of a successful asymmetric strategy: outsource the risk, keep the benefit, and pay in Yemeni currency for an effect measured in dollars across the world economy. The price of that success is settled elsewhere. By the crews who burn. By Yemenis, whose country serves as an aircraft carrier and who will absorb the retaliation. And by Iranians, whose state spends scarce resources and rare expertise on a war that returns them nothing, while the rial collapses and the power is cut at home.
The paradigm, not the incident
One can treat the Red Sea as an episode and look for the technical measure that closes it. That would repeat the mistake of the past twenty years. What this sequence shows, once again, is that armed projection beyond its borders is not one policy among others for the Islamic Republic. It is how the system survives. A power that has given up grounding its legitimacy in the consent of its population grounds it instead in the capacity to do harm abroad, and that capacity has to be maintained, displayed, periodically used. Training foreign sailors on the Caspian follows the same logic as capturing the scientific output of its universities: putting every national resource at the service of an external balance of power.
From this follows a difficulty chancelleries know without naming it. As long as the paradigm holds, containing Tehran requires instruments that also reach, and sometimes reach first, Iranian society. Broad sanctions impoverish families before they dent an apparatus that has built an economy of circumvention. Strikes destroy civilian infrastructure whole cities depend on. An academic break hits researchers already under surveillance. Every available tool carries that flaw, and clarity means saying so rather than dressing it up.
On this, DORNA holds a simple and constant position. Recognise first the nature of the regime and the structural role the IRGC plays within it, rather than treating each crisis as a repairable slip. Then condition, systematically, everything that can be conditioned: scientific cooperation, technical exchange, access to markets and institutions, with verifiable criteria and clauses that actually bite. Control what stays open, in the knowledge that the regime has written into its own law that it will capture what it can. And break where control is impossible, without extending the break to those who are not the apparatus.
That requirement carries a corollary DORNA presses with equal consistency: protecting Iranian society, its researchers, its professionals, its bridges to the world. They are not the allies of the arrangement described here. They are the first to be watched by it and the first to be impoverished by it. A policy that hits them while aiming at the apparatus mistakes its target and, in time, forfeits the Iran that will make another relationship possible.
In the Red Sea, in July 2026, crews burned. The skills that made it possible to reach them were taught somewhere, to someone, by someone. As long as the regime that organises that teaching remains what it is, the question will never be how to close an incident, but how much longer the world agrees to absorb its cost.