Thunderstorm knocks out Microsoft Azure in West US

Microsoft ↗0.13% said lightning strikes overwhelmed backup generators at its West US 2 facilities, triggering outages starting at 4:27 UTC Friday.status
Dozens of Azure services were affected, including Azure OpenAI, SQL Database, and Virtual Machines; Copilot went down for thousands of users.status
By Friday afternoon, utility power and most services were restored, though storage-dependent workloads were still undergoing final recovery.status

Severe Thunderstorm Knocks Out Microsoft Azure Services in West US Region

Microsoft confirmed on Friday that a severe thunderstorm caused widespread utility power loss at its West US 2 datacenter facilities, triggering a multi-service outage that disrupted Azure cloud services and Microsoft Copilot for thousands of users.

Storm Overwhelms Backup Power Systems

The disruptions began at approximately 4:27 UTC on May 29, according to Microsoft’s Azure status page, after multiple lightning strikes affected several datacenter facilities in the West US 2 region. Backup generators activated as designed, but a subset of generator systems were unable to fully synchronize under the sudden facility load, while others shut down due to thermal protection mechanisms as cooling systems were also knocked offline by the broader power disruption.status

The concurrent failures across multiple availability zones compounded the impact. More than 2,600 users reported problems with Microsoft Copilot alone as of 8:58 a.m. PT, according to Downdetector. The Hindustan Times reported that thousands of users experienced slowness and connectivity failures throughout the day. A separate outage tracker confirmed the last detected Copilot outage occurred on Friday, May 29, lasting about one hour.gvwire

Dozens of Azure Services Affected

The list of affected services was extensive, including Azure Functions, Azure Kubernetes Service, Azure SQL Database, Virtual Machines, Storage Accounts, Azure Monitor, Application Insights, and Log Analytics, among others. The Azure OpenAI Service also experienced elevated error rates in multiple regions earlier in the day, according to Microsoft’s Foundry status page.status

By Friday afternoon, Microsoft said datacenter utility power had been fully restored and network infrastructure was fully operational. HVAC systems returned to normal operation and ambient temperatures stabilized. The company confirmed that numerous services — including Service Bus, App Service, Azure Cosmos DB, Azure Data Factory, Azure Databricks, and Azure Synapse — had been restored and were operating normally.status

Residual Recovery Underway

As of the most recent update at 18:21 UTC on Friday, Microsoft said residual impact was limited to storage-dependent workloads on two remaining storage stamps undergoing final recovery and data integrity checks, with full restoration estimated within three to four hours. The company advised customers with resources in paired or alternate regions to consider failing over traffic away from West US 2 until full recovery was confirmed, and recommended pausing new deployments to the region.status

The incident marks the second weather-related power disruption at Microsoft’s western U.S. data centers this year. In February, a power interruption at a West US facility disrupted Windows 11 updates and Microsoft Store functionality.x