AI infrastructure on the front line: Lessons for Asean from the Iran war

A plume of smoke rises after a reported Iranian strike on fuel tanks in Muharraq, Bahrain, on March 12.

Photo: AFP On March 1, after Israel and the United States initiated attacks against Iran, Amazon Web Services reported drone strikes against data centre facilities in the United Arab Emirates and Bahrain.

The strikes caused structural damage to the company’s infrastructure, impairing cloud services for those countries.

Iran warned that US tech companies with Israeli links, including Google, Microsoft, Palantir, Nvidia and Oracle, were on Tehran’s list of “legitimate targets” for countermeasures.

Strikes on data centres in West Asia matter far beyond their blast radius.

The implications are profound for Southeast Asia, a region aspiring to be a digital and artificial intelligence (AI) hub and which has sought to insulate itself from the scrappy blows of geopolitics.

As technological shifts substantially reshape the nature of conflict, the targeting of digital infrastructure – including privately owned data centres – was always only going to be a matter of time in an active conflict.

The convergence of AI-powered data analytics, cloud storage and military operations means that in highly digitised societies, the line between economic disruption and strategic jeopardy can be indiscernible.

The data centres and algorithms that underwrite banking, healthcare, education and public administration for millions of civilians could also double as military targets for destruction.

Unless companies are proudly announcing their participation in “kill chains” in the US and elsewhere, there is not always a clear separation between civilian and military uses of cloud storage or AI workflows.

US Secretary of Defence Pete Hegseth holds a briefing amid the US-Israeli conflict with Iran at the Pentagon in Washington on March 2.

Photo: Reuters In some cases, ambiguity affords companies a cloak of deniability through their terms of service.

Microsoft invoked this defence – until it was pressured into a U-turn – when accused by multiple sources, including its own employees, of allowing the Israeli military to use its cloud and AI technologies to decimate Gaza.

Corporations have also been exposed for masking their involvement in military operations through back-door arrangements to circumvent legal obligations.

However, even when companies want to establish boundaries, it may not be simple.

The recent row between Anthropic and the US Department of Defence over the

原文链接: 南华早报