Utility power failure stops cooling and overheats Newark data centre
July 2025Published 13 Oct 2026Responsible editor: Chief Editor
Get the next one by email
Free alerts on data centre incidents in the United States, and an email when EIR-2334 is updated. No account needed; unsubscribe in one click.
Get a free sample brief Three complete verified records and the incidents-per-GW table, by email within a minute.
By signing up you agree to the privacy policy.
What happened
On 27 July 2025, at about 08:30 UTC, a utility power failure at the Akamai data centre at 165 Halsey Street, Newark, New Jersey, caused the loss of HVAC cooling. The data centre building remained operational, but temperatures rose and Akamai hardware overheated, forcing core services to shut down. Linode Compute Instances, Object Storage, NodeBalancers, Linode Kubernetes Engine and block storage in the Newark (us-east) region were affected. Kubernetes Engine services in Dallas, Fremont, Sydney, Tokyo 2, Toronto and Washington DC were degraded because of dependencies on Newark infrastructure.
Mitigation included replacing failed network hardware, rerouting traffic away from Newark and migrating workloads to backup systems where possible. Services were restored in phases once temperatures stabilised, some beginning to recover on 28 July, with full recovery completed on 29 July 2025 at 16:22 UTC.
Contributing factors
The loss of utility power removed cooling at the facility, and elevated temperatures then damaged or overheated hardware. Cross-region dependencies on Newark infrastructure spread the impact to other regions. The source does not state whether backup generators or uninterruptible power supplies were available to the cooling plant. Restoration was staged to avoid hardware damage.
Lessons learned
- Confirm that cooling plant has assured backup power, and test generator start and transfer under loss of utility supply.
- Map cross-region service dependencies so that a single-site power fault does not degrade services elsewhere.
- Alert early on temperature trends so staff can shed load before hardware overheats.
- Plan staged restoration with thermal checks before returning load to recovered equipment.
More lessons on outages, backup supply and restoration, from every incident on the register.
Source
Published 2026-10-13. Written from public reporting. Descriptive, not investigative. See the data accuracy disclaimer.
