Yandex Shuts 3rd Data Center in Four Days After Drone Attack
Yandex suspended its Vladimir data center after a drone attack, the third in four days. With two zones offline, the cloud operates in emergency mode, exposing the limits of three-zone redundancy.

Yandex completely suspended operations at its data center in the Vladimir region, about 180 kilometers from Moscow, on Sunday (November 11). In an alert published on Telegram, the company stated that the site’s infrastructure was damaged by a drone attack and that no one was injured. This is the third installation of Russia's largest technology company targeted in four days, following Sasovo in the Ryazan region early on October 8, and Kaluga on Friday (October 9).
The governor of Vladimir, Alexander Avdeyev, reported that the attack caused a fire and power outages that temporarily affected around 5,000 residents. Kiev has not officially claimed responsibility for the attacks, and Yandex did not name the perpetrator in its statements.
A Cloud in Emergency Mode
The impact on corporate clients is the most relevant data for those operating infrastructure. More than 80 Yandex Cloud services went offline, according to the company's status panel and reports from following outlets, including the YandexGPT API, managed databases, storage, and networking tools.
The sequence explains the extent of the damage. The attack on Sasovo knocked out the availability zone ru-central1-b, with the power outage registered on the panel at 10:31 PM GMT on October 7. The attack on Vladimir affected ru-central1-a. With two zones unavailable, Yandex declared that the remaining resource configuration is unstable and advised clients to trigger their disaster recovery plans. The company has not disclosed a timeline for restoring operations.
Sasovo carries additional weight. In 2021, Yandex itself reported that this complex housed two of its three supercomputers based on Nvidia A100 chips, used to train YandexGPT. After the attack, according to Reuters, the company declined to say whether the machines were affected and stated it did not know if the site’s equipment could be recovered.
What the Three-Zone Architecture Did Not Anticipate
The standard design of public cloud is based on one premise: availability zones are physically separated so that a local failure, such as a fire or power outage, does not take down the entire region. Three zones are typically sufficient because the probability of two failing simultaneously due to independent causes is low.
Coordinated attacks break this independence. Those who choose the targets can hit zone after zone, and the redundancy calculated for accidents no longer holds. This was the challenge Yandex faced over four days, and it is not the first time in 2026.
On March 1, AWS reported that the mec1-az2 zone in its UAE region was hit by "objects" that caused sparks and fire, leading firefighters to cut power to the building and generators. The following day, the company confirmed that two installations in the ME-CENTRAL-1 region had been directly hit by drones and that one installation in Bahrain suffered damage from a nearby attack amid Iranian retaliation against the United States and Israel. AWS’s recommendation to clients was the same as Yandex’s now: migrate workloads to other regions.
The Cost for Cloud Customers
Both incidents affect very different markets. Yandex Cloud primarily serves Russian companies and agencies, isolated from American hyperscalers since 2022 and, therefore, without a domestic alternative of the same size to migrate within hours. In the Gulf, the issue was different: clients who maintained data only in one zone discovered that recovery depended on a building that had been affected.
For a CIO outside these regions, the practical lesson lies in the risk matrix. Most continuity plans treat availability zone as the unit of failure and region as a last resort. The attacks in March and October show that, in countries exposed to armed conflict, the entire region becomes the unit of failure, and recovery depends on replication in another geography, with data and permissions provisioned prior to the event.
This comes at a cost. Replicating across regions doubles part of the storage bill, adds data transfer fees, and, in regulated sectors, runs into data residency rules that prohibit taking information out of the country. The dilemma sharpens where the law requires location and the geography is in a risk zone, which is precisely the situation Russia faces today.
There is also a change in the nature of the target. Until now, attacks on infrastructure in modern warfare aimed at energy, refineries, and telecommunications. Three data centers of the same company in four days, one linked to the training of AI models, indicate that computing capacity has entered the list of strategic targets, alongside the power grid that supplies it. For war risk insurers and those deciding where to build the next GPU campus, this is the variable that changed this week.
Sources
- themoscowtimes.comhttps://www.themoscowtimes.com/2026/10/11/russias-yandex-says-data-center-in-vladimir-shut-down-after-drone-attack-a93942
- aljazeera.comhttps://www.aljazeera.com/news/2026/10/11/russias-yandex-says-data-centre-in-vladimir-shut-down-after-drone-attack
- investing.comhttps://www.investing.com/news/company-news/russian-tech-giant-yandex-halts-vladimir-data-center-after-fresh-drone-attack-4942599
- kyivpost.comhttps://www.kyivpost.com/post/86708
- status.yandex.cloudhttps://status.yandex.cloud/en/incidents/2092
- theregister.comhttps://www.theregister.com/2026/03/02/amazon_outages_middle_east/