Strait of Hormuz carries more than oil, handling 20% of global data flows

Up to 20% of global internet and financial data traffic runs through subsea cables in the Strait of Hormuz, creating a geopolitical chokepoint for cloud, AI, and payment networks. Yardeni Research says heightened tensions and Iran’s comments on cable fees could force costly alternative routes, with new overland and regional links likely taking years to build. The risk is especially relevant for major tech firms and Gulf AI infrastructure projects that depend on at least seven major cables crossing the strait.

Analysis

This is less about immediate service interruption risk and more about a slow-burn repricing of network resiliency as a strategic asset. The first-order beneficiaries are not the hyperscalers themselves but the vendors that monetize redundancy: fiber builders, submarine cable installers, route-diversification contractors, and terrestrial backhaul providers. The second-order loser set is any AI/data-center operator with a Gulf growth thesis predicated on low-latency connectivity; if capacity needs to be duplicated across multiple paths, marginal ROI on new regional compute clusters falls and deployment timelines stretch by 6-18 months.

For AMZN, GOOGL, META, and MSFT, the issue is not revenue loss from a cable cut in a base case, but capex inflation and degraded service-level elasticity. These firms can absorb a one-off outage, but repeated headline risk raises the cost of doing business in a region they would otherwise use for latency arbitrage between Europe and Asia. That tends to favor larger, cash-rich incumbents over smaller cloud challengers, while simultaneously improving the bargaining power of local telecoms and sovereign-backed infrastructure partners that control alternative routes.

The most underappreciated angle is that AI infrastructure is becoming geographically constrained by physical chokepoints, not just power availability. That creates a medium-term tailwind for NVDA and AMD insofar as every new redundant cluster still needs accelerators, but the timing is lumpy: urgency can shift spend forward only after procurement committees formalize resilience budgets. Over months, this could also widen the valuation gap between AI-exposed software names and the picks-and-shovels layer, because resilience spend is easier to defer than compute spend when budgets tighten.

Consensus likely underestimates how quickly procurement behavior changes after the first credible incident. One well-publicized near-miss could trigger board-level mandates for dual-homing and alternative routing, which would be bullish for infrastructure spend but bearish for operating margins at the hyperscalers. The market is also probably over-discounting the optionality in non-Gulf routes: Oman, Saudi overland, and Red Sea adjacency become more valuable if strategic risk premium persists for 12+ months.

Key Decisions for Investors

• Long BIP / DLR / EQIX vs short a basket of AMZN, GOOGL, META, MSFT on a 3-6 month horizon: express the view that redundancy capex accrues to infrastructure landlords faster than to platform margins. Risk/reward improves if Gulf-risk headlines persist without an actual outage.
• Buy NVDA and AMD on dips, but pair against MSFT or META if you want cleaner exposure: a resilience-driven AI buildout still requires accelerators, yet the hyperscalers absorb the near-term capex drag. Best entry is after any headline selloff in semis, with 2-4x upside to any accelerated procurement cycle over 6-12 months.
• Initiate a small tactical long in a fiber/backhaul beneficiary or regional telecom proxy tied to Gulf route diversification, with a 1-3 month catalyst window around government or consortium announcements. Use a tight stop; if no concrete route investment emerges, the trade decays quickly.
• For event risk, buy 1-2 month out-of-the-money puts on a regional cloud/cloud-infrastructure proxy if cable disruption headlines intensify. This is a convex hedge: low carry, but meaningful payoff if a genuine outage or sabotage event forces immediate rerouting and SLA concerns.