How the Iran War Exposed the Physical Backbone of the AI Boom

The artificial intelligence (AI) boom is often framed as a story of capital, code, and compute. Alphabet, Amazon, Meta, and Microsoft forecast capital expenditures of about $650 billion in 2026. At the same time, AI-linked stocks are trading near all-time highs.

Beneath the software layer, however, sits a physical supply chain that depends on transformers, switchgear, batteries, grid capacity, and a less-discussed input: Helium. 

AI’s $650 Billion Buildout Has a Helium Problem

For context, Helium is a critical input for semiconductor manufacturing, supporting cooling, leak detection, and high-precision processes. Yet the supply is quite concentrated.

According to the US Geological Survey, Mineral Commodity Summaries 2026, global production in 2025 reached approximately 190 million cubic meters. The United States, Qatar, and Russia together account for roughly 84% of global helium output. 

Qatar’s output reached 63 million cubic meters that year, accounting for roughly a third of the global helium supply. Almost all of it comes out of Ras Laffan Industrial City, the world’s largest LNG complex.

That balance broke on March 18, when Iran struck Ras Laffan during the broader US-Iran war. QatarEnergy declared force majeure on some of its long-term liquefied natural gas (LNG) contracts. In a press release, it also warned repairs could take up to 5 years.

“QatarEnergy expects the damage to its Ras Laffan Industrial City caused by missile strikes, which occurred on Wednesday 18 March 2026, and in the early hours of Thursday 19 March 2026, to cost about $20 billion a year in lost revenue and to take up to five years to repair, impacting supply to markets in Europe and Asia,” the state-owned petroleum company wrote.

Follow us on X to get the latest news as it happens

Russia closed the second door in April. It imposed export controls on Helium through the end of 2027. Officials cited domestic supply needs, particularly for fibre-optic components used in military drones.

Electrical equipment bottlenecks are compounding the problem. Roughly half of the US data centers slated to come online this year are now facing delays or outright cancellations, with transformers, switchgear, and batteries in short supply. 

The same components are needed to expand the grid itself, which is also absorbing rising demand from electric vehicles and heat pumps. Domestic manufacturing capacity has not kept pace, leaving data center developers increasingly dependent on imports.

Thus, the AI boom is running into a physical chokepoint that capital alone cannot fix. Wall Street has priced in the execution, yet the inputs behind that execution are tightening one by one. 

Subscribe to our YouTube channel to watch leaders and journalists provide expert insights