The shortage of NVIDIA video cards may slow down Yandex's projects to develop drones and artificial intelligence technologies.
In the spring of 2022, Yandex was supposed to receive a large batch of NVIDIA devices, but after the start of the special operation, the delivery fell through. Since that time, the import of these devices to Russia is possible only through parallel imports, but it is problematic to import chips in the volumes necessary for the company, sources told Vedomosti.
The problem was discussed in mid-January with relevant departments and experts at various levels. It is extremely serious for the company. Yandex has invested a lot of resources in the development of AI, but cannot deploy the infrastructure for it without video cards.
The lack of NVIDIA GPUs is a serious problem for the development of artificial intelligence in Russia, says Vladimir Arlazarov, CEO of Smart Engines:
Sanctions have been imposed on high-performance equipment for training neural networks, in addition, some companies, for example, Sberbank, are under US sanctions, while Russia does not have its own comparable hardware and capacities for their production, unlike China.
Within the framework of the current paradigm of the development of neural networks, an extensive approach is used in Russia to achieve new levels of quality and functionality, says Vladimir Arlazarov. Such an approach, according to him, provides for a constant complication of models and, accordingly, an increase in equipment requirements:
In addition, thoughtless and poorly justified use of neural networks for solving any problems is typical, even if they can be perfectly solved by other approaches. Therefore, for companies that use this approach - and among them Yandex, and Sberbank, and VK - it will be difficult to maintain the pace of development.
According to Vladimir Arlazarov, the solution could be the transfer of development centers to neutral territory, gray import purchases, or a complete change in the AI development paradigm, namely, the transition from an extensive approach to an intensive one.
Most likely the problem will be solved by a combination of the first and second ways. The choice of an extensive path for the development of AI will require a shift in emphasis to fundamental science and the Russian Academy of Sciences as the main actor, although this path is the most beneficial for Russia, it is now unlikely,” concludes Vladimir Arlazarov.