Ukrainian Strikes Hit Russian AI Supercomputers

Ukrainian attacks on Russian data centres reportedly damaged three Yandex supercomputers used to train artificial intelligence models earlier this month. The extent of the damage remains unclear, and Yandex hasn’t commented. Still, independent Russian outlet T-Invariant quoted supercomputer infrastructure experts who said all three systems suffered serious damage and remain out of service for an indefinite period.

The Attacks

A Ukrainian attack hit the Yandex data centre in Sasovo on October 8. It houses the Chervonenkis and Lyapunov supercomputers. That was ten days ago. Yandex’s third supercomputer, Galushkin, is in Vladimir; a Ukrainian attack on October 11 shut down its data centre entirely. T-Invariant reported that satellite images of Sasovo showed a major fire and a burned section of the building.

The danger reaches beyond the immediate fire zone. These data centres use free cooling, which pulls outside air through ventilation systems. Fire, smoke and combustion byproducts could therefore damage equipment far from the flames, while Ukraine also struck another Sber data centre in Balakovo, Saratov Oblast, on October 14. The other two Russian supercomputers, Christofari Neo and Christofari, sit at Sberbank’s data centre in Skolkovo, outside Moscow, which hasn’t been attacked.

The Policy Gap

Supercomputers can carry out quadrillions and quintillions of operations and analyse large datasets. Generative artificial intelligence has made them essential for training large neural-network models, too. 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.

The picture has changed sharply since then. Russia’s full-scale invasion of Ukraine in February 2022 marked a turning point. Many Western countries imposed broad export restrictions on semiconductors, equipment and other high-tech components. Meanwhile, rival countries and major technology companies built more powerful systems with newer accelerators. Russian machines largely stayed the same and fell down the rankings. Chervonenkis still performs at about 21.53 petaflops, but its ranking dropped from 19th place in 2021 to 101st in June 2026, the latest rating available.

Imported hardware is the core problem. All five Russian TOP500 supercomputers use American AMD or Intel processors and NVIDIA A100 or V100 accelerators. Russia developed its own infrastructure and software, but imported hardware supplied all of its computing power. Since 2022, export restrictions have made even legal supplies of these previous-generation components significantly harder to obtain.

Competitive Stakes

Yandex said its supercomputers are designed specifically for machine-learning tasks, and the company used its own systems to train YandexGPT. Christofari and Christofari Neo helped develop Russian-language and generative models, including the GigaChat family and earlier projects. Yet the wider race is moving fast. The Stanford AI Index 2026 said American companies created 50 of the most notable new AI models in 2025, while Chinese companies created 30. Private AI investment that year reached approximately $286 billion in the United States and $12.4 billion in China, figures that don’t capture a significant portion of Chinese government funding.

Russia ranked 31st among 83 countries in the Global AI Index 2024. A Sberbank representative estimated at the time that Russian generative models lagged those of the United States and China by approximately six to nine months. That gap now looks harder to close. Repairing a damaged system requires compatible GPUs, processors, networking equipment and memory. Building a modern replacement requires access to the next generation of accelerators. Russia can operate large computing complexes and build competitive models, but it lacks an industrial counterpart to the NVIDIA A100, let alone the more advanced H100 and H200 accelerators.

What Comes Next

Yandex hasn’t said when, or whether, its three machines will return to service. Sberbank’s installations remain intact for now. The June 2026 leader on the TOP500, LineShine, achieved 2.198 exaflops—more than 100 times Chervonenkis’s performance. Whatever the final damage assessment shows, the gap between Russia’s AI infrastructure and that of its rivals is already wide.

Why This Matters: The strikes expose how deeply Russia’s AI capacity depends on imported chips, a vulnerability that years of sanctions have made harder to fix. For ordinary users of Russian-language AI tools, the consequences could mean slower development and a narrower range of models. For Europe, the episode is a reminder that technological competition now runs alongside the war itself. Strategic dependence on a few foreign suppliers, whether for chips or cloud infrastructure, leaves any government vulnerable. Europe’s own push for digital sovereignty and public investment in compute capacity looks more urgent when a major power’s infrastructure can be disrupted this quickly. The human costs of the war in Ukraine continue to unfold on the ground, and the technological fallout shows how far its effects reach.