How did AWS data centers go offline?

What happened and who was affected

Amazon Web Services reported that facilities in the Middle East experienced power and connectivity disruptions after being struck by objects during a period of regional military activity. Several AWS availability zones in the United Arab Emirates and Bahrain were taken offline or reported degraded service; Amazon said restoring normal operations would take time because strikes damaged infrastructure and complicated recovery efforts.

Customers across the region saw outages and service interruptions spanning web hosting, databases, and cloud services that rely on those availability zones. Multiple news outlets and Amazon statements described sparks and fires at impacted sites, and initial recovery work focused on assessing physical damage, rerouting traffic, and repairing power and network links.

Immediate consequences:

• Outages and degraded performance for regional customers relying on the affected availability zones.
• Increased operational strain as traffic was shifted to other zones and contingency plans were executed.
• Broader supply‑chain and continuity concerns for multinational services that depend on geographically distributed cloud infrastructure.

Why this matters

Cloud providers design resilience around redundant regions and failover, but direct physical attacks on data‑center infrastructure test those assumptions. Damage to site power and networking can extend downtime and complicate incident response, especially in conflict zones. For enterprises, the incident highlights the need to consider geopolitical risk in cloud architecture and to plan cross‑region redundancy and rapid failover strategies. For cloud operators, it raises questions about physical security, insurance, and how to maintain service continuity when infrastructure faces real‑world attacks.