How is Russia trying to build a domestic GenAI order under sanctions, supply vulnerability, and limited external access? The picture that has emerged suggests that Russia is constructing an AI system clustered around a few national champions, prioritizing the Russian language, and deeply intertwined with the state.

Yandex occupies a central position in this landscape. Through Yandex Cloud, the company offers the YandexGPT model family and Yandex AI Studio. Yandex’s own documentation describes YandexGPT as a “native” model family, supporting around 20 languages but focusing especially on Russian-language use cases. In other words, a domestic model layer and a domestic service layer have taken shape. Public Yandex Cloud documents also show that its services use GPUs such as the NVIDIA V100, A100, and T4.

As a second major player, Sber’s GigaChat model is also offered to enterprise customers through domestic cloud infrastructure. This suggests that the GenAI market in Russia is evolving not into a free and fragmented competitive space, but into a more controlled structure shaped around a few large actors such as Yandex and Sber.

On the performance side, Yandex claims that its models outperform GPT-4.1 on some tasks. These claims matter, but they still need to be verified, since the measurements are internal to the company. Even if YandexGPT appears strong and competitive in Russian-focused enterprise tasks, there is not yet enough independent evidence to say that it is competing head-to-head with the strongest U.S. and Chinese models.

According to Reuters, executives at Sber have also stated that Russia is roughly six to nine months behind the United States and China in AI. That may be a long gap in AI time, but from Russia’s perspective it still means avoiding a complete fall behind in the race.

Rather than becoming a fully independent AI superpower, Russia appears to have succeeded in building a sovereign AI order that can continue functioning under sanctions, prioritizes Russian, and is carried by the state and large corporations. Of course, sustaining this structure under intense competition and ensuring reliable access remain difficult challenges. For that reason, the continuous expansion of GPU and data center infrastructure is critically important.