Russia Sees AI as Nuclear-Level Power Shift

Russia runs no ‘AI bubble’ risk as its investment not excessive
Use of foreign AI models in sensitive sectors is risky
Global AI investment is ‘overheated hype’
Russia must invest $570 billion in power generation, grids

Artificial intelligence will grant nations influence comparable to nuclear technology, according to Alexander Vedyakhin (pictured), first deputy CEO of Sberbank. Speaking at Russia’s annual AI Journey event, he said countries that secure leadership now will hold superiority throughout the century. Russia counts itself among seven nations with home-grown AI models, a position Vedyakhin described as an achievement. He stressed that sensitive areas such as healthcare, education, and public services require original domestic models rather than retrained foreign ones.

Closed Membership in the AI Club

Vedyakhin likened the global AI race to the emergence of a “nuclear club,” where only nations with their own large language models (LLMs) can participate. He argued that the United States and China are ahead of Russia by six to nine months, and that joining the race late is nearly impossible. Countries that have not yet begun developing national AI systems face prohibitive costs and widening gaps. Western sanctions restricting access to advanced computing technology further complicate Russia’s efforts to catch up.

Sberbank is positioning its GigaChat 2 MAX model as comparable to ChatGPT 4.0, while its GigaChat Ultra Preview is said to rival ChatGPT 5.0. The company plans to release some of its models as open source, including for commercial use. Vedyakhin acknowledged the strong head start of U.S. and Chinese firms, citing their financial resources, expertise, and computing power. He emphasized that Russia must continue building multiple original models to maintain sovereignty.

Investment and Infrastructure Challenges

AI development requires vast resources, particularly in energy. Vedyakhin estimated Russia’s power sector would need 40 trillion roubles ($506 billion) for generation and 5 trillion for grids over the next 16 years. He noted that programmers and mathematicians will play a crucial role in reducing costs and accelerating machine learning. Skill, rather than sheer scale, will be Russia’s advantage in advancing AI.

Future breakthroughs may come from new architectures beyond generative pre-trained transformers, similar to China’s DeepSeek project in 2024. Vedyakhin cautioned that energy consumption makes returns on AI investment uncertain or distant. He warned against excessive infrastructure spending, describing it as vulnerable to hype. Russia, he argued, is immune to an “AI bubble” because its investments remain measured.

Sovereignty and Strategic Importance

President Vladimir Putin recently underscored the importance of home-grown AI models for preserving Russian sovereignty. Sberbank and Yandex are leading national efforts to compete with global leaders. Vedyakhin reiterated that uploading confidential information into foreign models is prohibited due to security risks. He said Russia must ensure its AI systems are trusted and independent to safeguard critical sectors.