Russia is investing in home-grown AI to secure a competitive edge, likening its strategic importance to nuclear technology. Alexander Vedyakhin, first deputy CEO of Sberbank, said the country must develop multiple original large language models (LLMs) for sensitive areas such as public services, healthcare, and education.
Sberbank and Yandex lead development, with the bank’s GigaChat 2 MAX model positioned against ChatGPT 4.0 and the new GigaChat Ultra Preview comparable to ChatGPT 5.0. Some models may be open-sourced for commercial use.
Vedyakhin acknowledged U.S. and China maintain a six- to nine-month lead, citing Western sanctions and limited access to computing resources as challenges. He emphasized reliance on skilled programmers and mathematicians to overcome these gaps, warning that excessive infrastructure investment may not yield immediate returns.
Despite the hurdles, Russia aims to remain among the “AI club” countries with indigenous LLMs, focusing on efficiency and technical skill rather than scale. Officials argue that domestic AI is vital for sovereignty and security, especially when uploading sensitive data to foreign models is prohibited.
