Russia's AI Ambitions Under Fire

Independent outlet T-invariant asked leading Russian supercomputing experts to assess how Ukrainian drone strikes on Yandex data centers have affected the company's artificial intelligence operations. Their conclusion: The damage could set Russia's AI development back years and push it behind Kazakhstan in the global supercomputing race.

All three Yandex supercomputers are reportedly offline indefinitely following a series of attacks. The first strike occurred on October 8 in Sasovo, Ryazan Oblast, home to Chervonenkis, Russia's most powerful publicly ranked supercomputer, and Lyapunov, its third most powerful. Ukrainian forces struck Yandex's largest data center in Kaluga the following night, then attacked a facility housing the Galushkin supercomputer in Vladimir on October 11.

Four specialists interviewed by T-invariant believe Yandex's computing infrastructure suffered extensive damage. The main vulnerability, they say, is the company's direct free-cooling system, which uses outside air to cool servers. Even if flames never reached the equipment, smoke and soot could have contaminated circuit boards. Carbon particles conduct electricity and can bridge tightly packed connections, causing widespread short circuits and motherboard failures.

Experts estimate that cleaning the systems could take six months. Replacing damaged components, particularly scarce and expensive NVIDIA chips or their equivalents, could extend recovery to one or two years.

The disruption threatens YandexGPT and Alisa AI just as demand is surging. According to Yandex B2B Tech figures cited by financial news outlet Finam, token consumption at Yandex AI Studio increased eighteenfold in six months. Yandex's service-status reports indicate that the platform is currently unavailable.

Experts argue that the facilities were designed for peacetime, not warfare. "Everything was built for normal civilian life. Nobody imagined drones hitting the roof," one supercomputing expert told T-invariant, noting that natural disasters are rare in the regions where the facilities operate.

A BBC source said Yandex regularly practiced shifting workloads between data centers after individual outages. However, simultaneous failures across several major sites were apparently not anticipated.

Russia's remaining supercomputing infrastructure may also be vulnerable. Sber's Christofari and Christofari Neo systems are housed in Skolkovo, near Moscow, where air defenses are stronger. MTS operates its GROM supercomputer in Lytkarino, while Yandex maintains additional facilities in Mytishchi and Ivanteyevka. All three Moscow Oblast cities face recurring Ukrainian drone attacks.

The consequences could soon become visible in the global Top 500 supercomputer rankings. Chervonenkis currently ranks 101st, narrowly ahead of Kazakhstan's Alem at No. 104. Another Kazakh system ranks 122nd, above every other Russian machine. T-invariant's experts expect the November rankings to place Kazakhstan ahead of Russia in top-end supercomputing performance.

Russia already trails the United States, China, and Japan by roughly two orders of magnitude. The contrast extends to Nebius, the Netherlands-based company led by former Yandex chief Arkady Volozh. Its two supercomputers, ISEG and ISEG2, have combined peak performance of approximately 425 petaflops. ISEG ranked 12th globally when first listed.

By comparison, Russia's five Top500 systems collectively deliver less than 100 petaflops of peak computing power, according to the Russian government's mathematical sciences development strategy through 2036. That means Volozh's single company commands more peak supercomputing power than all of Russia's Top500-listed systems combined.