Iran threatens a "Persian Gulf Data Authority": AI data centers are now a declared target

By The Times of Israel · October 18, 2026.

The speaker of Iran's Parliament, Mohammad Bagher Ghalibaf, has announced on X that a "Persian Gulf Data Authority" is coming. According to The Times of Israel reporter Lazar Berman, the message suggests it will be part of an effort to target data centers in the region.

Ghalibaf warned "governments, funds and companies in the region" not to provide resources to any AI or data company that, in his words, aids invasions or the killing of children. He listed what must not be handed over: "financing, capital, land and energy." And he added: "Help them and you will face our enforcement."

The message comes days after Mohsen Rezaei, whom the outlet describes as Iran's top security official, threatened that undersea cables, data centers and other AI infrastructure could be targeted if the U.S. or Israel attacks Iran again. The article does not detail what powers the new "authority" would have or whether it exists beyond the announcement.

Our take: what matters is not the specific threat, which for now is rhetorical and undeveloped, but the framing. Iran is trying to turn the AI supply chain (money, land, electricity) into a front line of conflict. If a data center is considered part of a rival's war effort, it ceases to be a neutral commercial asset.

In general, the Gulf's large computing campuses have been sold as an economic and prestige bet. Statements like this introduce a political risk that investors and insurers will have to weigh, although nothing in the text indicates that attacks or withdrawal decisions have already occurred.

There is also an ambiguity worth watching. Talking about companies that "aid invasions" does not define what military use turns an AI provider into a target. That blurred line is precisely what makes the escalation dangerous: a civilian service and a military one share the same servers.

In the short term, this points to a rough transition: more spending on physical protection and redundancy, and more pressure to separate civilian AI from military AI. In the long term, our thesis remains that the abundance AI promises requires shared, stable infrastructure, and that is achieved with clear rules that protect civilian assets, not with bloc logic. The sooner there is agreement on what is off-limits, the less likely it is that the network meant to cure diseases ends up on the target list.

This is a brief liveblog item: details are lacking, and the prudent course is to treat the announcement as a political signal until it takes concrete shape.