Yandex’s Vladimir data center shuts after reported Ukrainian drone strike

Yandex suspended operations at its Vladimir data center after an October 11, 2026 drone strike, Ars Technica reported. It was the third Yandex facility hit in four days by attacks the outlet attributed to Ukraine. The affected sites house the company’s AI-training supercomputers, while reported outages have spread to banking, shopping and transport services.

Three facilities, different levels of damage

The Vladimir shutdown follows attacks on Sasovo, southeast of Moscow, and Kaluga, southwest of the capital. Yandex’s October 9 statement gives a more precise account of those first two incidents than descriptions suggesting every facility was knocked out entirely. The company confirmed a complete shutdown at Sasovo, but only a partial infrastructure outage at Kaluga.

Sasovo was seriously damaged during the night of October 7–8, Yandex said. Kaluga was attacked on the morning of October 9. No employees were hurt at either location, and emergency services and company specialists were working on site. Kaluga is one of Russia’s largest data centers, with a maximum capacity of 63 megawatts, according to Ars Technica.

Sasovo houses Chervonenkis and Lyapunov, two of the three supercomputers Yandex has used to train YandexGPT, its AI model. Both use Nvidia A100 graphics-processing chips. Their rated computing performance is 29.42 and 20.03 petaflops, respectively—a measure of how many calculations they can perform. Vladimir houses the third machine, Galushkin, rated at 20.64 petaflops.

Those locations put all three AI-training supercomputers inside facilities affected by the attack sequence. That does not establish that the machines themselves were destroyed. Ars Technica said the extent of damage to Sasovo’s supercomputers was not immediately clear; the Vladimir account describes suspended operations rather than a hardware assessment.

We are continuing to assess the condition of our infrastructure. For now, it is very difficult to say when the equipment at the data centers will be back up and running.

The Yandex Team, October 9, 2026

Cloud customers carry the disruption further

The consequences extend beyond developing AI models. Ars Technica reported interruptions or outages affecting payment apps, digital banking, email, e-commerce, video streaming, ride-hailing, food delivery and navigation. Yandex also operates Russia’s dominant search engine, accounting for about 70 percent of that market. Its infrastructure supports both its own products and businesses buying cloud services.

Damage also affected Yandex Cloud customers, including large grocery and fast-food chains and an appliance retailer, according to Moscow Times reporting cited by Ars Technica. That provides a route for disruption to reach companies whose customer-facing services do not carry the Yandex name.

Gizmodo’s October 11 coverage, citing the Kyiv Post, identifies several services reportedly affected that Sunday. These are dated outage reports, not confirmation that every service remains unavailable on October 12:

• Yandex’s Alice voice assistant, Yandex Music and Telemost, its video-meeting service, were reportedly disabled when the Kyiv Post published.
• Yandex Compute Cloud, which provides virtual machines, reportedly experienced extensive outages. Users also complained about banking and rideshare app failures.
• Apps belonging to retailers Magnit and Fix Price were reportedly interrupted, adding named examples of disruption beyond Yandex’s own consumer services.

Equipment replacement has no confirmed timetable

Yandex said its teams were working day and night to keep services running and help affected users, customers and partners. Its October 9 statement treated service continuity as the immediate priority while specialists assessed the infrastructure. The company said its technology helps hundreds of thousands of businesses take orders, serve customers and work with partners.

An unnamed IT expert interviewed by Russian newspaper Kommersant estimated that replacing damaged data-center equipment could take up to six months, Ars Technica reported. That is an outside estimate, not a Yandex recovery commitment. It concerns equipment replacement rather than a promise that each affected app will remain offline for that period.