Amazon Web Services suffered a significant power outage in its US-EAST-1 region on May 7, impacting EC2 instances and EBS volumes after a thermal event triggered cooling system failures. The incident, first reported at 5:25 PM PDT, affected the use1-az4 availability zone and left customer resources inaccessible. AWS acknowledged that services depending on the affected EC2 instances and EBS volumes also experienced impairments.
The outage began when AWS detected problems in the availability zone, later confirming that hardware lost power during a thermal event. By 8:06 PM PDT, the company reported that progress in restoring normal temperatures was slower than anticipated, despite making incremental improvements to cooling systems. Customers experienced elevated error rates and latencies across multiple workflows as AWS worked to stabilize the situation.
AWS responded by shifting traffic away from the impacted availability zone and bringing additional cooling capacity online by 10:11 PM PDT (05:11 UTC May 8). The company reported recovering some affected racks and stated it was working to restore additional racks in a controlled manner. However, the recovery process remained gradual, with customers reporting that even the AWS status dashboard was timing out during the incident.
The outage affected customers across the region, with users reporting complete loss of access to their servers. AWS suggested that organizations shift workloads to other availability zones within US-EAST-1, though the company acknowledged that provisioning times were significantly longer than normal. The incident disrupted operations for an unknown number of customers relying on the affected infrastructure.
US-EAST-1 has a history of major outages, including significant incidents in 2021 and October 2025 that disrupted large portions of the internet. While AWS executives maintain the region is not inherently more fragile than others, they acknowledge it operates at larger scale than other regions, which places additional stress on services. Organizations using US-EAST-1 should review their disaster recovery plans and consider multi-region architectures to reduce exposure to single-region failures.
Source: https://www.theregister.com/off-prem/2026/05/08/aws-warns-of-ec2-impairment-as-power-loss-hits-notorious-us-east-1-region/5235509