Iran's Ghalibaf announces a Persian Gulf Data Authority and warns AI firms' Gulf backers

What Ghalibaf said

Iran's parliament speaker, Mohammad Bagher Ghalibaf, has told governments, investment funds and companies in the Gulf to stop supplying artificial intelligence and data companies that he says are helping attacks on Iran. In an English-language post on X at 16:55 UTC on Sunday 18 October 2026 he wrote: "The Persian Gulf Data Authority is coming. To governments, funds, and firms in the region: you are kindly advised to give no resources to any AI or data company aiding invasions or killing children. No finance, equity, land and energy. Assist them and face our enforcement."

The post had been viewed more than 580,000 times by Monday morning. Iran's state-run Mehr News Agency reported it the same evening, translating the body's name as a "Persian Gulf Data Management Authority", and said Ghalibaf warned that anyone who helps such companies "will face our coercive action". Ghalibaf did not name any company, country or investment fund.

The four things he listed, finance, equity, land and energy, are what an AI data centre project needs from a host country: money, ownership partners, a site and a power supply. That makes the post less a statement about software than a warning aimed at the physical build-out of AI capacity around the Gulf.

What a Persian Gulf Data Authority could mean

Ghalibaf gave no details of how the authority would work. Iran International, a London-based outlet opposed to the Iranian government, noted that he did not explain what powers it would have or how Iran could enforce restrictions on companies operating in other countries.

Iran's state broadcaster Press TV offered one interpretation. It said the planned body would be modelled on the Persian Gulf Strait Authority, which Iran set up after the US-Israeli attacks earlier this year to control shipping through the Strait of Hormuz, and described the new authority as a way to impose checks and controls on the flow of data in the region. Press TV linked the announcement to Iran's demand that the United States lift what it calls a maritime blockade of the country.

Read that way, the authority would treat data the way Iran already says it treats tankers: as traffic passing close to its territory that it can regulate, tax or threaten. That is Press TV's reading, not a published plan. No legislation, decree or founding document for the authority had appeared by 19 October.

A week of threats against digital infrastructure

The post capped a week in which Iranian officials and state media repeatedly named digital infrastructure as a possible target. According to Iran International, Mohsen Rezaei, secretary of Iran's Supreme National Security Council, said in a Press TV interview that "many cables, data centers and artificial intelligence facilities, valued at $2 trillion, will be within range" if war with the United States and Israel resumed. On 17 October Mehr published an analysis headlined "Persian Gulf Internet, AI Infrastructures within Iran's reach", which said renewed aggression could lead to attacks on cables and data centres.

Iran International also reported that Hossein Shariatmadari, editor of the hardline Kayhan newspaper, endorsed Rezaei's warning and called for using the undersea fibre-optic cables that run through the Gulf and the Strait of Hormuz to pressure Washington. He claimed damage to those cables could disrupt more than 80 percent of commercial, banking, military and AI-related communications, a figure Iran International said has not been independently verified. The IRGC-affiliated Tasnim News Agency ran an article titled "Why Is Artificial Intelligence on the Target List?", arguing that attacks on AI centres could hurt US technological competition with China and unsettle American stock markets.

The same outlet listed earlier proposals from Tasnim, including licensing and annual renewal fees for foreign operators of undersea cables crossing the strait, and a demand that companies such as Meta, Amazon and Microsoft operate under Iranian law. Ghalibaf's post turns those scattered ideas into a named institution, even if it remains only a name for now.

The precedent from the spring war

The threats are not hypothetical. Iran International reported that Iranian drones damaged Amazon Web Services data centres in the United Arab Emirates and Bahrain in early March, during the war, and that Amazon said two UAE facilities were directly struck while a nearby strike damaged its infrastructure in Bahrain, causing power disruptions and cloud outages for customers across the region. Press TV, for its part, described "limited attacks" on US-run data centres during the 39-day conflict that it said ended in mid-April, and presented them as a response to strikes on banking data and AI infrastructure inside Iran.

On 31 March, according to Iran International, the Islamic Revolutionary Guard Corps published a list of 18 companies whose regional operations it said could be targeted. The list included Microsoft, Google, Meta, Apple, Oracle, Nvidia and the Abu Dhabi AI group G42, and the IRGC accused American technology and AI companies of assisting US military operations and told employees to leave their workplaces.

Data centres have become targets in other conflicts this year too. Drones shut down three Yandex data centres in Russia between 8 and 11 October, taking the company's AI-training supercomputers offline. Compute that once sat in quiet industrial parks is now treated by militaries as strategic infrastructure.

Who the warning is aimed at

Ghalibaf's phrase "AI or data company aiding invasions" is broad enough to cover most large US cloud and model providers, several of which work with the US military. He addressed the hosts and financiers of those companies rather than the companies themselves, which puts the pressure on Gulf governments and sovereign funds that have spent heavily to attract data centres and AI investment.

Mehr's 17 October analysis made the same point from Tehran's side, describing Saudi Arabia and the United Arab Emirates as having made extensive investments in cloud services and AI to diversify away from energy revenue. The UAE in particular hosts regional AI developers such as G42, whose Inception42 unit worked with MBZUAI on the Jais 3 Arabic models released last week.

The question of which AI companies count as military suppliers is also live in the United States, where Anthropic is still fighting the Pentagon's supply-chain risk designation of its Claude models in federal court. Ghalibaf's wording draws no such distinctions, and he offered no list or criteria.

What is still unknown

Much about the announcement is unclear. There is no published mandate, budget, membership or timeline for the Persian Gulf Data Authority, and no indication of whether it would be created by parliament, by the government or by the Supreme National Security Council. Iran has no legal jurisdiction over data centres in neighbouring states, so "enforcement" in Ghalibaf's sense appears to mean coercion rather than regulation, but he did not say what form it would take.

AI Lookout found no public response from Gulf governments or from the companies on the IRGC's March list in the reporting available on 19 October. Most of the reporting so far comes from Iranian state media, which presents the authority as a defensive measure, and from Iran International, which presents it as one of a series of threats. Both describe the same post; neither has seen a plan.

Why it matters for AI infrastructure

For AI companies, the practical issue is risk pricing. A large AI campus depends on a stable power supply, long-term land deals and fibre routes, and Gulf states have courted AI builders with exactly those inputs: energy, sites and capital. An explicit threat from the speaker of Iran's parliament against anyone who supplies those things adds political and physical risk that insurers, lenders and boards will have to weigh, whether or not the authority ever takes shape.

That comes as investors are already questioning the returns on AI infrastructure. Gartner warned last week that more than $3.5 trillion of AI infrastructure spending could be exposed to impairment by 2030. Geopolitical threats to specific regions add another variable to those calculations.

The next things to watch are whether Iran's parliament or government publishes a founding text for the authority, whether Gulf states or the US respond formally, and whether data centre or cable operators in the region change their security posture or disclose new risk factors. AI Lookout will update this article if the authority gains a legal basis or a concrete mandate.