Amazon Web Services has determined that it cannot restore access to resources and data stored exclusively within its Bahrain Region following physical infrastructure damage that compromised multiple Availability Zones simultaneously.
The outage started in March during regional conflict. According to Reuters, strikes hit two AWS facilities in the United Arab Emirates, and a nearby drone strike damaged AWS infrastructure in Bahrain itself. Banking operations experienced disruptions at that time.
AWS acknowledged that the extent of the damage surpassed what its regional design could handle. "The damage to our infrastructure spanned multiple Availability Zones and exceeded what our regional and multi-AZ services are designed to withstand," the company stated in an update.
The episode has exposed a fundamental constraint in cloud resilience architecture: while spreading workloads across Availability Zones within a Region can shield against isolated failures, it offers no protection against physical destruction severe enough to impact multiple zones at once.
AWS Directed Customers to Relocate Operations
When the initial Availability Zone suffered damage in March, AWS instructed customers to shift workloads to different Regions. The majority had completed this migration before a second Availability Zone failure in April rendered the Bahrain Region entirely inaccessible.
Since then, AWS has assisted remaining customers in restarting their operations in alternative Regions, drawing on saved backups or deploying alternative strategies to reduce losses from unreachable data.
Concurrently, AWS examined the damaged infrastructure and investigated potential paths for recovering data and resources that customers had not transferred before the Region went offline.
AWS has now concluded that these recovery options are no longer viable. The company stated it will continue supporting Bahrain-based customers going forward and plans to issue an additional status report in early 2027.
Understanding the Boundaries of Regional Resilience
The Bahrain situation underscores an important distinction for organisations building cloud resilience plans.
Availability Zones enable workload distribution within a single Region. In Bahrain, however, physical destruction extended across multiple zones, creating a scenario that AWS stated exceeded the design parameters of its regional and multi-AZ architecture.
For businesses, this raises the importance of clarity about which failure modes their infrastructure design actually protects against. Even when workloads span multiple Availability Zones, they remain vulnerable to disruptions affecting the broader Region itself.
The timeline in Bahrain also demonstrates the value of contingency planning that extends beyond regional boundaries. Organisations that relocated workloads following the initial March incident faced a different outcome than those still present when the second Availability Zone failed in April.
Sovereignty and Resilience Present Separate Challenges
The outage has raised broader concerns regarding the geographic placement of critical computing infrastructure and governance over it.
Mark Boost, chief executive of Civo, suggested the incident should prompt UK organisations to reassess their dependence on computing facilities outside their own borders. "Amazon has spent six months trying to bring its sites back and has now informed its customers that it cannot. That is not a criticism of Amazon," Boost stated. "It is simply what happens when critical infrastructure resides in foreign territories."
Boost contended that genuine sovereign AI requires a nation's computing capacity to be constructed and controlled within its own boundaries, rather than treating sovereignty solely as a matter of where data physically sits.
The Bahrain disruption does not prove that domestically controlled infrastructure would withstand physical damage better. Sovereignty and resilience address distinct concerns.
Infrastructure location and ownership shape regulatory authority and decision-making power. Resilience, by contrast, depends on how workloads, data, backups, and recovery processes are arranged across different systems.
AWS's experience in Bahrain demonstrates that even the largest cloud providers encounter failure thresholds. For organisations relying on cloud infrastructure, the practical question becomes identifying where those limits exist and whether their recovery strategies account for scenarios beyond them.
