AWS Says Some Data in Bahrain and UAE Cloud Sites Cannot Be Recovered After War Damage

Amazon Web Services has told customers that some data and computing resources in its Middle East cloud infrastructure cannot be recovered after war-related damage, an unusually blunt disclosure from the world’s biggest cloud provider.

In a public AWS Health Dashboard update dated around Sept. 15 and reported by Reuters, DatacenterDynamics, The National and The Register, the company said it could not restore access to resources and data hosted only in its Bahrain cloud region, known as Middle East (Bahrain). In the United Arab Emirates, AWS said it could not restore access to resources and data hosted only in one availability zone — a cluster of data centers designed to back up one another — in its Middle East (UAE) region. AWS said most customers were able to resume operations in other AWS regions by restoring backups or copying accessible data, and that support teams would continue helping migrations.

For Bahrain, the AWS Health Dashboard said, “After a thorough assessment, we have determined that we are unable to restore access to the resources and data hosted exclusively in this Region.” For the UAE, it said, “After a thorough assessment, we have determined that we are unable to restore access to the resources and data hosted exclusively in the mec1-az2 Availability Zone.” AWS said the damage “spanned multiple Availability Zones and exceeded what our regional and multi-AZ services are designed to withstand,” according to the dashboard wording reproduced by DatacenterDynamics, The National and other outlets.

The damage traces back to March and April 2026, when drone strikes affected AWS-linked facilities in the UAE and Bahrain during the broader U.S.-Iran conflict. The Sept. 15 disclosure came roughly six months later. AWS said publicly that the facilities were damaged “as a result of the conflict in the Middle East,” but did not identify an attacker in the dashboard language. Iranian state and IRGC-linked media have claimed responsibility for multiple regional strikes, but individual attributions have not been uniformly independently verified.

The announcement matters because cloud regions are supposed to be built for failure. AWS regions are made up of multiple availability zones, separate facilities meant to keep applications running if one site goes down. AWS operates three availability zones in Bahrain and three in the UAE. In plain terms, customers that stored data only in the affected Bahrain region or only in the affected UAE zone may have permanently lost access to it. Customers that had backups or replicated data in other regions were far more likely to recover.

AWS also said it had exhausted its recovery options for data and resources that had not been moved before the region became unavailable, according to reports quoting the dashboard. Several news outlets described the statement as a rare public acknowledgment by a hyperscale cloud provider that region-exclusive resources could not be restored, and some said it appeared to be the first such disclosure by a major cloud company. AWS has not publicly said how many customers were affected or how much data was lost.

The company said it is replacing the affected infrastructure and will provide more updates in the coming months. For Bahrain, reports citing the dashboard said another update is expected in early 2027. Until then, the clearest takeaway for customers is also the starkest one: If critical systems existed only in the damaged locations, AWS now says it cannot bring them back.