New AI club to bestow nuclear-like power, says Russian tech boss

Written by: Elena Fabrichnaya and Gleb Bryanski

MOSCOW (Reuters) –Artificial intelligence A senior Russian artificial intelligence executive said this century will give them an edge, saying they will have the same level of impact as nuclear technology.

Alexander Vedyakhin, the first deputy CEO of Sberbank, which has transformed from a major credit institution into an artificial intelligence-focused technology holding, said in a statement to Reuters that it was a success for Russia to be among the seven countries with self-developed artificial intelligence.

“AI is like a nuclear project. A new ‘nuclear club’ is emerging globally, where you either have your own national large language model (LLM) or you don’t,” Vedyakhin said in an interview at Russia’s annual AI Journey event.

He said Russia should have at least two or three original AI models, rather than “retrained foreign models”, for use in sensitive areas such as online public services, healthcare and education.

“It is impossible to upload confidential information to a foreign model. This is strictly prohibited. Doing this will lead to very unpleasant consequences,” Vedyakhin said.

President ‌Vladimir Putin said last week that home-grown AI models are vital to preserving Russia’s sovereignty. Sberbank and technology firm Yandex are leading efforts to catch up with their U.S. and Chinese rivals.

Vedyakhin said Russia will have difficulty competing with leaders in the IT field, especially due to Western sanctions restricting access to technology, and the gap will likely grow.

AI CLUB MEMBERSHIP CLOSED

The United States and China are six to nine months ahead of the rest of the club, including Russia, and membership is effectively closed, Vedyakhin said.

“Every day counts in this race, but those who haven’t started are falling more than a day behind the leaders with each passing day. For those who decide to join now, it will be extremely costly, almost impossible,” he said.

“We appreciate what Chinese and American companies are doing. We understand that they have a strong advantage with plenty of money, expertise and computing power,” Vedyakhin said.

Vedyakhin said Sberbank’s GigaChat 2 MAX LLM is comparable to ChatGPT 4.0, while the new GigaChat Ultra Preview is on par with ChatGPT 5.0.

Sberbank is preparing to compete with next-generation models and plans to open source some of its latest models, including for commercial use.

RUSSIA Is Immune to ‘AI Bubble Risk’

Russia will rely on programmers and mathematicians to reduce costs and accelerate machine learning, Vedyakhin said, adding: “What we cannot achieve with numbers, we achieve with skill.”

Still, he says that developing artificial intelligence requires huge investment, estimating Russia’s energy sector at 40 trillion rubles ($506 billion) for production and 5 trillion for networks over the next 16 years.

He said a leap forward in ⁠LLM memory and the emergence of an ⁠AI architecture that does not rely on generative pre-trained transformers (GPTs) could mark the next breakthrough, similar to what China’s DeepSeek has made in 2024.

Vedyakhin warned that energy consumption levels make the return on AI investments “either remote or not visible at all” and warned against “overheated hype” on infrastructure spending.

“Given the rapid pace of technological development, we believe that excessive investments in AI infrastructure may indeed not pay off,” he said, adding that Russia is immune to the “AI bubble” because its investment is not excessive.

(Reporting by Gleb Bryanski; editing by Guy Faulconbridge and Alexander Smith)