AI & Data Centers

The impact of Hormuz disruptions on the global AI and data center sectors is largely indirect but could be among the most consequential effects of the conflict. The AI sector is central to global growth at a moment when few other engines of expansion are firing. Hyperscalers worldwide have announced combined AI spending of roughly USD 725 billion for 2026. This is the largest technology investment cycle in modern history, with U.S. firms accounting for the bulk of committed capital but Asian and European cloud providers accelerating their own buildouts in parallel. That global buildout, and the market confidence underpinning it, has provided a crucial counterweight to economic uncertainty stemming from the Strait disruptions. 

But the AI sector’s physical foundations, from chips and transformers to energy systems to the materials used to build data centers, are deeply embedded in Asian supply chains that the Hormuz crisis is stressing. Semiconductor fabs in Taiwan and South Korea, transformer manufacturers in Korea and Japan, and copper refiners in China are all shared inputs for data center construction globally. The risks to the buildout are in these structural, and largely invisible, nodes in the supply shock instead of in shortages of critical components. With U.S. technology equities at the gravitational center of global capital markets, any sustained pressure on the AI investment thesis would reverberate through pension funds, sovereign wealth portfolios, and financial markets worldwide.

When the Strait of Hormuz closed to commercial shipping, the immediate alarm centered on helium, a critical component in advanced semiconductor manufacturing. Prior to the conflict, Qatar supplied over 30 percent of the global helium market; Iranian strikes on Qatar’s Ras Laffan gas fields took most of this production offline. South Korea and Taiwan, which manufacture most of the advanced and memory chips used to power AI systems, sourced 55 percent and 69 percent of their helium respectively from the Gulf. Helium spot prices surged 40 to 100 percent almost immediately. Any further or prolonged squeeze on helium supplies could create real constraints in the semiconductor ecosystem and at a minimum, force chipmakers to redirect supply away from legacy-node production.

Nonetheless, the helium scare appears to have been overstated in the near term. The global semiconductor industry operates with long-term supply contracts, inventory buffers, and increasingly sophisticated recycling systems. The margins on advanced AI chips are large enough for chipmakers to absorb significant spot-price premiums and pass elevated costs through to customers. In a market where demand for leading-edge logic and high-bandwidth memory chips already outpaces supply, a helium surcharge will not break the business model for major semiconductor manufacturers or chip designers. Even in the event of shortages, much of the semiconductor industry could outbid competitors for limited supplies, and the strategic and economic importance of the AI buildout means that governments would potentially step in to ensure alternative supplies — including from the United States, which remains the world’s largest helium producer. 

The picture is less straightforward down the supply chain. Smaller fabs clustered in Southeast Asia, particularly in Malaysia and Vietnam, tend to operate on thinner margins, carry fewer reserves, are more exposed to price shocks, and could not rely on the same level of diplomatic and geopolitical firepower in the event of a true shortage. The same logic applies to higher energy costs, which may not meaningfully impact leading-edge AI chip production but could squeeze lower-margin producers across the ecosystem. 

The more consequential threat is not a shortage or price shock for a single input or commodity, but the cumulative inflationary pressure that is already moving through every layer of the physical infrastructure that underpins the industry and could dampen global investment in the medium term.

The data center buildout requires extraordinary quantities of steel, aluminum, and copper. Aluminum prices are up 20 to 30 percent year on year, due in part to disruptions in Gulf supply. Copper is foundational to every cable, transformer, and grid upgrade, yet a significant share of its production depends on sulfuric acid derived from sulfur that transits the Strait and is now increasing in price. Diesel — the fuel of freight — adds a further layer of cost escalation to every input shipped to a construction site. 

Individually, each of these increases is manageable. Together, they put a compounding inflationary squeeze on an industry whose physical supply chains were already stretched to the breaking point. Bloomberg reported in April 2026 that roughly half of U.S. data centers planned for this year are now expected to be delayed or canceled — not for lack of capital, but because transformers and switchgear can take three to five years to deliver.

Asia sits at the center of this exposure in ways that go beyond semiconductors. South Korea supplies the U.S. transformer market. China is the world’s leading refined copper producer. Japan and South Korea produce construction equipment used in data center construction. Disruptions to their industrial supply chains accumulate across the physical inputs underpinning AI systems. As one of the binding constraints on AI capacity shifts to power infrastructure, disruptions to these supply chains can impact the pace and timeline of the AI buildout even if the cost impacts are minimal.

Pre-war expectations of interest rate cuts in 2026 have given way to the prospect of rate increases as central banks confront a supply-side inflation shock. As 10-year yields climb toward 4.5 percent, the financial model underpinning the AI buildout comes under meaningful strain. This matters because the buildout is increasingly debt-funded. The five largest technology companies will spend some 90 percent of their operating cash flow on capital expenditure this year, leaving little room for error. Private credit markets, which have bankrolled much of the data center expansion, are beginning to absorb concentration risk at scale.

Sustained elevated interest rates do not need to crash the AI investment boom to have an effect. Even a modest rise in borrowing costs could prompt investors to rethink hundreds of billions of dollars in planned AI infrastructure that has not yet broken ground. The AI buildout relies on cheap, abundant capital, and the compounding pressures from Hormuz disruptions put the business model at risk. AI spending has been one of the few forces cushioning markets from the broader economic shock of the crisis, but the sector absorbing that blow is itself vulnerable to the inflationary pressures the war is creating.

If Hormuz disruptions lead to sustained higher costs for fuel, electricity, and food, the contrast between trillion-dollar data center investment and strained household budgets could intensify the political backlash. In that scenario, a relatively permissive regulatory environment may shift in ways the industry is not prepared for. The Hormuz crisis did not create social tensions over data centers, but it has sharpened them. The AI buildout is unlikely to be derailed by a single input shortage, but it could become slower, more expensive, more geographically concentrated, and more politically contested.