What happened
Published by Brookings' Artificial Intelligence and Emerging Technology Initiative and Strobe Talbott Center on August 3, 2026 (the first in a planned series), this report examines the strategic tradeoffs of locating frontier AI data centers outside the United States. The authors argue overseas AI infrastructure faces direct, potentially irreversible physical-security risk — sustained low-cost drone swarms or attacks on supporting power infrastructure cannot be reliably defended against, unlike the political-pressure risks of dependency, which can at least be negotiated. The report frames this as a first step toward understanding domestic capacity constraints and the probability/severity of security risks to overseas compute, weighed against economic-competitiveness tradeoffs.
Why it matters
With billions in AI infrastructure investment flowing to the Gulf and other overseas locations, this gives boards and policymakers a framework for evaluating physical-security exposure of offshore AI compute commitments before signing long-term capacity deals.
Action needed
Boards overseeing overseas AI infrastructure investment or compute-supply agreements should request a physical-security risk assessment modeled on this framework before finalizing commitments.