Yandex Says It May Lose Equipment at Sasovo Data Center, Which Is Fully Offline

Yandex does not yet know whether it can recover the equipment at its data center in Sasovo, Ryazan Oblast. The site has been fully shut down since the attack, and the company is still assessing the damage to the infrastructure. Interfax reported this on 8 October 2026. There are no specific figures: no extent of the destruction, no number of affected racks, and no timeline for bringing the site back online. The wording "did not rule out the loss of equipment" suggests that even basic diagnostics are not finished.

What is known so far

There are few sources, and that is worth keeping in mind. Everything published boils down to a short statement from the company. The report does not disclose details of the attack or list the affected systems.

• Facility: Yandex's data center in Sasovo, Ryazan Oblast.
• Status: the site is fully shut down.
• Equipment: whether it can be recovered is unknown, and its loss has not been ruled out.
• Damage: the assessment of the damage to the infrastructure is ongoing.
• Source: Interfax, published on 8 October 2026.
• Not disclosed: the site's capacity, the number of affected servers, the restart timeline, and the impact on specific services.

A full shutdown is different from partial degradation. Had only one machine hall or one power line been hit, the company would most likely have described a limited outage. Here, the entire site is down.

Buildings and hardware are recovered differently

A data center has two layers of assets, and losing each costs a different amount. The first layer is the engineering infrastructure: power supply, cooling, fire suppression, and cable routes. Repairing it is expensive and slow, but the process is technologically straightforward, and contractors and equipment for such work are available on the market. The second layer is the computing hardware: servers, storage systems, and network switches. It cannot be repaired in the field if boards or disks are damaged, if the equipment was soaked with water during firefighting, or if it failed because of power surges.

That is why the line about possible equipment loss matters more than the rest. External restrictions make it harder for Russian companies to buy server hardware in the volumes they need, and accelerators for AI workloads are a separate problem. Replacing lost capacity "off the shelf" will not always be possible, and it will not be quick. The company has not said which workloads ran in Sasovo, so it is not yet possible to judge how severe the blow to its computing resources will be.

What to watch next

The coming weeks should clarify three things. The first is the outcome of the technical assessment: what can be repaired and what will be written off. The second is whether load has been redistributed to the company's other sites, if that was needed. The third is Yandex's public position on its services: whether it will report disruptions or say that redundancy worked without noticeable consequences for users. For now, there is no confirmation either of outages or of their absence.

What teams that depend on the cloud should do

For cloud customers and developers, this story is a reason to check their own architecture, not to panic.

• Find out in which zones and regions your critical services physically run and your data is stored.
• Check that backups sit at a different site, not in a neighboring zone of the same data center.
• Compare your real RTO and RPO targets with those written in the documentation. They often differ.
• Run failover drills: a plan that has never been executed counts as untested.
• Ask your provider for a written explanation of the status of your resources instead of relying on rumors.

Until official data appears, any forecasts about timelines and scale remain guesswork. Base decisions on confirmed statements from the company and specialized sources.