- Статьи
- Economy
- Skilful mind: a third of IT specialists will switch to domestic neural networks by the end of the year
Skilful mind: a third of IT specialists will switch to domestic neural networks by the end of the year
By the beginning of 2026, a third of IT specialists in Russia will switch to domestic AI tools that are used to create digital products, software solutions and software. The reason is the increased leakage of data from foreign resources. Already 17% of developers use Russian analogues in their work, another 11% are considering or testing them. The attractiveness of such tools is due, among other things, to complete Russification, the absence of the need to use a VPN, and the availability of local support. Experts expect the trend to strengthen as trust in local technologies continues to grow.
What is the reason for the transition to local platforms?
To create digital products, software solutions and software, domestic IT developers began to actively switch to Russian solutions - by the beginning of 2026, their number will grow to almost 30%, Yandex B2B Tech told Izvestia. The main reasons are the growing number of data leaks from foreign services, the increasing difficulty of accessing them without a VPN and the need for local support, as well as security requirements within companies.
According to a joint study by Yandex B2B Tech and ITMO University, 17% of IT specialists are already using Russian development platforms with embedded AI, 11% are considering testing or a full transition within a year. Another 10% are ready to choose between domestic and international solutions, evaluating the effectiveness in specific projects.
More than 600 respondents were interviewed, including practicing IT specialists, teachers of technical universities and students studying programming. In the segment of IT specialists, backend developers (programming specialists involved in the development of a functional part of a software application or website), department heads, system administrators, analysts, support engineers, and others were interviewed.
Among the key advantages of Russian platforms, developers highlighted full support for the Russian language (42%), reduced dependence on foreign providers (40%), increased data security (27%), and compliance with local requirements for secure development (24%), the study says. An equally important factor is the lack of need for a VPN.
Thus, against the background of limitations of foreign services and growing trust in local technologies, experts predict an increased trend towards the real implementation of Russian AI tools in development. In the coming years, local AI platforms can become the basis for improving the efficiency of development in companies of any scale, the study notes.
At Yandex, 62% of developers regularly use SourceCraft Code Assistant to automate everyday tasks, from auto—completion to test generation, said Dmitry Ivanov, head of the platform. Over 1.4 million lines of code are accepted and integrated through the system in a month.
The MWS AI press service noted that the company's developers use its own Kodify AI assistant to write computer code and other programming-related tasks. The company noted the appearance on the market of high-quality domestic products based on generative artificial intelligence, which are able to solve problems in Russian as well as, and sometimes better than, foreign ones.
Earlier, the world media reported that American specialists from Wiz Research, which specializes in cybersecurity, discovered a leak of confidential data from the Chinese startup in the field of artificial intelligence (AI) DeepSeek. More than 1 million lines of user information were left unprotected, including digital program keys and user query histories in the company's AI chats.
In addition, as Izvestia wrote, since the beginning of 2025 in Russia, the number of hacked accounts in neural networks has increased by 90% compared to the same period last year. This is due to the growing popularity of AI services and their active use. Due to restrictions on registration using Russian numbers, users are massively turning to buying or renting foreign numbers to access ChatGPT, Grok, DeepSeek and other AI assistants.
Severstal's press service told Izvestia that they have already started using their AI assistants for development, as well as for testing and code verification tasks.
— To reduce the risks from using open models, we have launched a platform for solutions based on the Da Vinci Genius. Several installed AI models are available to users: for text and image processing and code, they noted.
Why are developers in Russia abandoning foreign AI services?
IT specialists in Russia are increasingly paying attention to domestic neural networks, primarily because they handle many work tasks as well as their foreign counterparts and do not create additional difficulties, said Anton Khaimovsky, chief product owner of Content AI.
— Domestic solutions are available directly without bypassing, without the risk that the connection will be terminated or access will be blocked. This is especially important for businesses, since disabling an AI tool embedded in workflows can paralyze the work of a single unit or the entire company," he said, noting that this will not happen to Russian neural networks, they will not be disabled due to sanctions, political decisions or changes in licensing policy.
According to him, there is another important advantage — the availability of affordable technical support. You can quickly contact domestic developers, arrange improvements to meet the needs of the company, or get help with integration, for example, ask them to reduce or increase the degree of censorship of the neural network.
— It is important to understand: there are no guarantees that the data in Russian neural networks will be better protected. Leaks are possible everywhere. The problem is not where the neural network is located, but how it works, who controls it, and what security tools are used," explained Anton Khaimovsky.
Leaks from foreign services have become a signal for many companies to switch to local platforms, said Denis Nagaev, Technical Director of OSMI IT. And if earlier it was discussed as a long—term strategy, now it has become a tactical necessity to avoid reputational and legal risks tomorrow, the expert added.
In fact, this solution can be called proactive: neural networks remain a "black box" and how vendors use input data is still an unanswered question, he noted. Vasily Grebennikov, Executive Director of WPP.DIGITAL (which develops digital products for businesses), believes that domestic models can be used to control data, but this requires more costs for their support and training.
Переведено сервисом «Яндекс Переводчик»
