Ukraine Strikes Expose Russia’s AI Hardware Dependence

Ukraine struck Russian data centres earlier this month, reportedly damaging three Yandex supercomputers used to train artificial intelligence models and putting pressure on computing infrastructure that depends on imported hardware. The extent of the damage remains unclear, and Yandex hasn’t commented. Experts quoted by independent Russian outlet T-Invariant said all three systems suffered serious damage and remain out of service for an indefinite period.

Yandex systems hit

The Sasovo data centre, home to the Chervonenkis and Lyapunov supercomputers, was hit on October 8. Satellite images showed a major fire and a burned section of the building, T-Invariant reported. The facility uses free cooling, which draws outside air through ventilation systems; smoke and combustion byproducts could therefore damage equipment beyond the immediate fire zone.

Yandex’s third supercomputer, Galushkin, sits in Vladimir. A Ukrainian attack on October 11 forced its data centre to stop operations entirely. According to infrastructure experts cited by T-Invariant, the attacks put all three systems used for machine-learning work out of service.

Two other Russian supercomputers, Christofari Neo and Christofari, were built for SberCloud and installed at Sberbank’s data centre in Skolkovo, outside Moscow. That centre hasn’t been attacked. Ukraine struck another Sber data centre in Balakovo, Saratov Oblast, on October 14.

Supercomputers handle quadrillions and quintillions of operations and can analyse large datasets. They also train large neural-network models. Yandex said it designed its systems specifically for machine-learning tasks and used them to train YandexGPT. Christofari and Christofari Neo helped develop Russian-language and generative models, including the GigaChat family and earlier projects.

Imported components and export restrictions

All five Russian supercomputers in the TOP500 run on American AMD or Intel processors and NVIDIA A100 or V100 accelerators. Chervonenkis and Galushkin used AMD EPYC processors and NVIDIA A100 graphics accelerators; Lyapunov used AMD EPYC and A100. Christofari Neo also used AMD EPYC and NVIDIA A100, while Christofari used Intel Xeon processors and NVIDIA Tesla V100 accelerators.

Russia developed its own infrastructure, software and application solutions, but its computing power depended exclusively on imported hardware components. Russia’s full-scale invasion of Ukraine in February 2022 marked a turning point in access to those components: many Western countries imposed broad export restrictions on semiconductors, equipment and other high-tech goods.

Those restrictions have made legal supplies of the previous-generation components used in Russian systems significantly more difficult. Repairing damaged machines requires compatible graphics processors, processors, networking equipment and memory. A modern replacement would require access to the next generation of accelerators. Russia doesn’t have an industrial counterpart to the NVIDIA A100 or the more advanced H100 and H200 accelerators.

Falling rankings, widening gap

In November 2021, seven Russian supercomputers appeared in the TOP500, the international list of the 500 most powerful computing systems. Chervonenkis ranked 19th, Galushkin 36th and Lyapunov 40th. Christofari Neo ranked 43rd and Christofari 72nd.

Chervonenkis still performs at about 21.53 petaflops, but ranked 101st in the June 2026 rating, the latest available. Russia now has five systems in the TOP500 and ranks 21st by number of systems. The June 2026 leader, LineShine, achieved 2.198 exaflops, more than 100 times Chervonenkis’s performance. Christofari Neo ranked 167th in that list, while Christofari also remained on it.

The Stanford AI Index 2026 said American companies created 50 of the most notable new AI models in 2025, compared with 30 from Chinese companies. Private AI investment in 2025 reached approximately $286 billion in the United States and $12.4 billion in China, figures that don’t account for a significant portion of Chinese government funding. Russia ranked 31st among 83 countries in the Global AI Index 2024. A Sberbank representative estimated that Russian generative models lagged those of the United States and China by approximately six to nine months. Significant damage to Russian supercomputers could make it harder for Russia to improve its position in the global AI race.