Plenty of Russians reached for their phones this week to hail a car, pay for something, or find a route across town. Many got an outage instead.
Yandex’s ride-hailing, navigation, and payment services in Russia were disrupted after Ukrainian drones struck the company’s data centers. The attacks mark a significant expansion of Ukraine’s retaliatory campaign against Russian digital infrastructure.
Three sites, three strikes
The strikes unfolded over four days in October 2026. Drones hit a Yandex facility in Sasovo on October 8, one in Kaluga on October 9, and another in Vladimir on October 11.
Two of the data centers were shut down completely, while the third had key modules disabled.
Yandex operates five primary data centers in Russia. At least three of them were affected by the strikes.
Yandex Go, the company’s ride-hailing service, faced widespread outages. So did Yandex Pay and the company’s navigation tools.
Yandex reported that its advanced supercomputers, which are essential for training AI models, were housed at the targeted sites. The affected machines include systems named Lyapunov, Chervonenkis, and Galushkin.
Markets registered the hit. Yandex’s stock fell approximately 4% after the strikes.
Why Yandex, and why now
Yandex holds a 70% share of search engine usage in Russia, and its services extend into mobility, payments, mapping, and cloud computing.
The timing follows a pattern of escalation. Russian forces had stepped up attacks on Ukrainian digital networks starting in late September, and the Yandex strikes came in response to those assaults.
Ukrainian President Volodymyr Zelenskyy framed the drone campaign in exactly those terms, describing the strikes as a tit-for-tat response to Russian aggression.
What this means
For investors, the 4% drop may be the opening act rather than the full story. The strikes expose a concentration risk: when a large share of a company’s computing capacity sits in a handful of physical locations, those locations become single points of failure. Losing at least three of five primary sites shows how quickly that risk can turn real.
Training modern AI models requires large, dedicated compute clusters that cannot simply be swapped in overnight. Damage to the supercomputers housed at these sites could slow Yandex’s AI development and weigh on its cloud business.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.