CoreWeave incident
Facility power incident impacting services in US-EAST-15A
Ongoing since 09:32 PM UTC, affecting CoreWeave Infrastructure and CoreWeave Object Storage and 1 more.
Power has been restored and node recovery is underway. Nodes are being brought back online in phases, starting with air cooled systems; B200 (liquid cooled) nodes will follow once that first phase completes. We expect full recovery by Morning time due to the subset of racks which sustained a cooling issue during the event and require servicing before they can be powered back on. We'll share more updates as more nodes return to service.
Affected components
Update timeline
- identified Oct 22, 2026, 09:32 PM UTC
Hello, We are actively investigating a facility power event in US-EAST-15A following the loss of both utility power feeds at the site. Customers in this region may experience compute unavailability, job interruptions, , Inference disruptions and degraded or unavailable Object Storage. Our engineering and facilities teams are actively working to restore stable power and cooling. Impact may continue or widen until site conditions stabilize. If you are impacted or have questions, please contact [email protected]. We apologize for the disruption and thank you for your patience. Thank you, CoreWeave Support
- monitoring Oct 23, 2026, 12:05 AM UTC
The disruption has been traced to a tripped main feed from the local utility provider, who is actively troubleshooting on site. The facility is running on backup generator power, cooling has been restored, and room temperatures are returning to normal. When utility power gets restored, we will begin bringing systems back online in stages, prioritizing restoring customer capacity. A subset of racks sustained a cooling related hardware issue during the event and require servicing before they can be powered back on; we currently estimate these will be restored by Friday morning. We expect more information on the utility power timeline within the next hour and will post more updates
- monitoring Oct 23, 2026, 03:06 AM UTC
Power has been restored and node recovery is underway. Nodes are being brought back online in phases, starting with air cooled systems; B200 (liquid cooled) nodes will follow once that first phase completes. We expect full recovery by Morning time due to the subset of racks which sustained a cooling issue during the event and require servicing before they can be powered back on. We'll share more updates as more nodes return to service.
Your product runs on more than CoreWeave
Add the vendors you depend on and get one message when any of them reports trouble. Free for 3 vendors, no card.
Amazon Web Services