Artificial intelligence will hand enormous geopolitical leverage — comparable to nuclear capabilities — to the countries that manage to get ahead now, according to Alexander Vedyakhin, the first deputy CEO of Sberbank.
He said dominance in AI will grant nations strategic superiority throughout the century, drawing a direct line between technological leadership and national power.
Speaking with Reuters at Russia’s annual AI Journey conference, Vedyakhin argued that it was no small feat that Russia sits among what he described as the “seven countries with home-grown AI.” Sberbank, which has rapidly expanded from a major state-backed lender into an AI-driven technology conglomerate, has become one of the key players in this push.
“AI is like a nuclear project. A new ‘nuclear club’ is emerging globally, where either you have your own national large language model (LLM) or you don’t,” he said. Moscow believes it belongs inside that club, but officials openly admit the challenges ahead.
Vedyakhin said Russia must have at least two or three entirely original AI models — not foreign models retrained with domestic data — to power platforms in online public services, healthcare, and education. He warned that using foreign systems in such sensitive areas would be unacceptable.
“It is impossible to upload confidential information into a foreign model. It is simply prohibited. Doing so would lead to very unpleasant consequences,” he said.
The Kremlin has been sharpening that message. President Vladimir Putin said last week that home-grown AI was essential for protecting Russian sovereignty, reinforcing Moscow’s view that AI is no longer just a commercial tool but a national-security asset. Domestic champions such as Sberbank and tech giant Yandex are leading efforts to catch up with U.S. and Chinese developers, though hurdles remain.
The Race Leaders — and a Closing Door
Vedyakhin admitted Russia lags the United States and China in raw computing power, talent pool depth, and access to high-end chips — the last being heavily constrained by Western sanctions. He estimated that the U.S. and China are ahead of all other AI nations, including Russia, by “six to nine months,” a gap that he said is widening.
“In this race, every day matters,” he said. “But those who haven’t started are falling behind the leaders by much more than a day with each passing day. For those who decide to join now, it will be extremely costly, almost impossible.”
He added that the AI club is effectively “closed,” given the capital and expertise now required to build competitive large language models.
Still, Moscow is trying to show momentum. Vedyakhin said Sberbank’s GigaChat 2 MAX LLM is comparable to OpenAI’s ChatGPT 4.0, and the company’s new GigaChat Ultra Preview model is on par with ChatGPT 5.0. To expand its footprint, Sberbank plans to make some newer models open source, including for commercial use.
The push comes as Russia looks for ways to offset its disadvantage in chip supply. Vedyakhin said the country would increasingly rely on domestic programmers and mathematicians to accelerate training and reduce costs.
“What we can’t achieve with sheer numbers, we achieve with skill,” he said.
The Cost of AI Power — and the Energy Question
Vedyakhin noted that the structural demands of AI development are immense. He estimated that Russia’s power sector alone needs 40 trillion roubles ($506 billion) for electricity generation and another 5 trillion roubles for grid upgrades over the next 16 years to meet anticipated AI-related computing needs.
The energy burden is already a global concern. He pointed out that a breakthrough could come from an LLM architecture that isn’t based on classic generative pre-trained transformer (GPT) designs, noting China’s DeepSeek in 2024 as an example of a step-change model structure.
But he warned that current AI infrastructure spending comes with serious economic uncertainty. Energy consumption is so high that returns on investment are “either very distant or not visible at all,” he said — a point that has been raised internationally as companies and governments pour hundreds of billions into data centers, chips, and electricity networks.
“There is overheated hype around infrastructure spending,” Vedyakhin told Reuters. He argued Russia is less exposed to “AI bubble risk” because its investment levels remain comparatively restrained.
However, the overall message was that the global race has already hardened into a contest where very few countries have the capacity to compete at the top. But that created the objective for Russia to build sovereign AI models, try to close the computing gap, and secure a place in what Vedyakhin described as the emerging “nuclear club of AI.”







