Namecheap took all services dark for 15 hours after its Phoenix data center lost cooling

On August 13, 2026, a major storm knocked out cooling at the RadiusDC Phoenix data center, the facility hosting what Namecheap's own post-mortem calls "some essential Namecheap operations". Temperatures climbed toward critical levels, and the operator instructed Namecheap to take services offline entirely rather than risk equipment overheating and lasting damage. Namecheap — a 25-year-old registrar and hosting company — complied, and everything went dark at once: shared, VPS and dedicated hosting, EasyWP, DNS management, Private Email, billing, and the company's own website.

The post-mortem, written by CEO Hillan Klein the next day, is candid about the cascade. Customers could not manage DNS records or redirects. Many websites were unavailable, slow, or erroring; EasyWP sites kept throwing database connection errors even after the dashboard returned; Private Email could neither send nor receive. Crucially, the support helpdesk ran on the same failing infrastructure — for fifteen hours customers had no live chat, no email support, no way to reach a human. Recovery took temporary chillers and staged restarts, with full service about fifteen hours in.

The sting is in the facility's credentials. RadiusDC Phoenix is a Tier 3 data center "designed with a high level of redundancy", and Namecheap noted it had never experienced anything like this in its 25-year history. But a Tier rating certifies the building, not the tenant's architecture: Namecheap had concentrated its essential operations — and its customer-facing control plane — inside that single building, so one cooling failure in one Arizona storm disabled management interfaces, communication channels, and revenue systems simultaneously.

15 hours of downtime across hosting, EasyWP, DNS management, Private Email and billing; the helpdesk went down too, leaving customers nowhere to turn

Essential operations were concentrated in a single Phoenix facility, so the facility's failure became Namecheap's failure — no second site could keep the control plane alive.

The helpdesk ran on the same failing infrastructure, so customers with downed sites and stopped email also lost the ability to ask for help — an availability failure became a communication failure.

There was no continuity plan for the management plane: DNS management, billing, and email admin stayed dark for the whole incident even though DNS zone resolution itself stayed up.

Tier 3 redundancy at the facility level masked the tenant-level single point of failure; the post-mortem calls the event unprecedented in 25 years.

A facility's Tier 3 badge is not your redundancy: keep your control plane — DNS management, email, billing, support — dependent on one building and one cooling failure blacks out everything at once.

Cooling was restored with temporary chillers directed into Namecheap's area of the facility, and services came back progressively as temperatures stabilized — staging the company defended as the only safe order. The post-mortem promised to implement even more redundancy across its other datacenters in Europe and Asia. Namecheap also published a correction: early posts on X had blamed the wrong operator, PhoenixNAP, when the facility was run by RadiusDC. The fifteen-hour outage drew widespread customer anger, with the company publicly conceding "we failed you".