Amazon Faces Months of Repairs After Drone Strikes Cripple Middle East Data Centers

Damaged data center being repaired after drone strikes

Amazon Web Services says recovery from drone strikes that hit its data centers in the United Arab Emirates and Bahrain will be measured in months, leaving customers in the region facing prolonged disruption and prompting a broader rethink of investments in Middle East infrastructure. The attacks, part of a wider bout of regional hostilities, knocked core compute racks offline, triggered flooding when fire suppression systems activated, and damaged cooling systems—problems that together complicate a simple, rapid restoration.

What happened

On April 30, AWS posted an update revealing that its ME-CENTRAL-1 (UAE) and ME-SOUTH-1 (Bahrain) regions “suffered damage as a result of the conflict in the Middle East” and are currently unable to support customer applications. The company said billing operations for the affected regions have been suspended while teams work to restore normal operations, a process AWS expects could take several months. The outage follows Iranian drone strikes in the region earlier this year that targeted multiple data center facilities.

Extent of the damage

Internal documents obtained by other outlets described damage to at least one data center that put 14 EC2 server racks offline and impacted another five racks. EC2 is AWS’s core virtual server service, meaning those losses directly affected customers that rely on scalable compute instances. Damage wasn’t limited to servers: fire suppression activation caused flooding and water damage at one facility, and mechanical failures in cooling systems further compounded the recovery challenge. Taken together, these failures require hardware replacement, extensive cleanup, and cooling and power systems repairs—hence the months-long timeline.

Customer impact and responses

AWS initially waived usage-related charges for March 2026 at an estimated cost of roughly $150 million and later advised customers to migrate resources to other cloud regions and restore services from remote backups where possible. Some customers managed rapid recoveries by moving workloads to unaffected regions; for example, Dubai-based super app Careem reportedly performed an overnight migration to get services back online. But for many organizations with complex deployments, interdependent services, or data residency requirements, migration and recovery will be slow and costly.

Industry and regional fallout

The strikes and AWS’s prolonged recovery timetable have ripple effects beyond individual outages. Pure Data Centre Group, a London-based developer, announced a pause on Middle East data center investments until the conflict eases, reflecting growing investor and operator caution. The damage also highlights the vulnerability of centralized infrastructure to regional geopolitical risks, and it may accelerate multi-region redundancy strategies, regulatory reassessments of cloud residency rules, and investments in resilient edge or hybrid-cloud architectures.

Wider context and implications

The conflict that led to the strikes began on February 28, with U.S. and Israeli attacks on Iran followed by retaliatory Iranian strikes across the region. The situation has since settled into an uneasy ceasefire punctuated by naval blockades around the Strait of Hormuz and a growing economic and energy shock. For global cloud customers and infrastructure planners, the AWS outages serve as a reminder that geopolitical instability can translate quickly into long-lasting service interruptions. Expect more thorough contingency planning, a renewed emphasis on cross-region backups, and potentially higher insurance and operational costs for cloud-reliant businesses operating in or near conflict zones.

What to watch next

Recovery progress from the damaged AWS regions will be the immediate metric to track—whether AWS can accelerate repairs and when normal billing and service levels resume. Regulators and customers may press for clearer disclosures about resiliency and risk mitigation, and data center developers will weigh whether to recommit to the region. Finally, enterprises that experienced disruption will likely review cloud architectures, testing failover procedures and considering multi-cloud approaches to reduce single-vendor, single-region exposure.