The government is actively working on a scenario for the mandatory use of domestic AI models in key sectors: from security and public administration to education and electoral processes. This is not about a total ban on foreign technologies, but about creating economic incentives that will make choosing Russian software more advantageous for businesses. In particular, in the education sector, the innovations will only affect school and preschool levels, while companies that continue working with foreign solutions risk losing access to tax benefits.
The essence of the initiative and pressure mechanisms
Deputy Prime Minister Dmitry Grigorenko outlined the contours of the new policy at the WEF-2026 platform: there are industries where priority for domestic developments should be established by default. The Cabinet is currently refining this approach, seeking a balance between the interests of the state and business. According to my information, the main tool will not be a ban, but the withdrawal of state support. Organizations that choose foreign AI products will simply be unable to claim subsidies and preferences when implementing technologies. This effectively creates a powerful economic incentive for businesses to switch to the Russian stack now in order to maintain access to government contracts.
At the same time, until 2032, the possibility of using foreign AI remains even in priority industries. Such a lengthy transition period gives companies time to restructure without direct administrative shock. The authorities are betting not on the stick, but on the carrot: in parallel, a package of support measures for developers of sovereign models is being prepared. The logic is simple: priority industries generate demand, and tax incentives reduce the cost of creating the technologies themselves.
Tax relief and departmental ambitions
Behind the scenes, a whole set of fiscal incentives is being discussed. This includes reducing the tax on the purchase of or remote access to AI solutions, subsidizing the developer in the amount of the product price for small businesses. Coefficients for AI expenses, preferential VAT on services, and exemption of specialized data centers from property tax are also being developed. Grigorenko's office confirms: work on by-laws is ongoing, and the final parameters have not yet been fixed.
The ambitions of the departments are impressive. The Ministry of Construction, for example, plans to transfer half of the processes in construction and housing and communal services to AI by 2030. The Ministry of Finance already uses algorithms in preparing the federal budget, and Minister Siluanov calls AI a new point of economic growth. The example of the "Opturan" platform is also telling, which was the first in the world to optimize public transport schedules using AI, reducing inter-route downtime in Moscow by 34%. These are not just experiments, but working cases that set the direction for the entire country.
However, there are also objective difficulties. The share of products with real agentic functionality in the Russian registry is estimated at only 5%, compared to 39% globally. This is a serious gap that needs to be overcome. According to the AIANA baseline forecast, the market volume will grow from 316.1 billion rubles to 830 billion by 2030, and under a favorable scenario, to 1.12 trillion. The fight for a share in priority industries is becoming a matter of survival and growth for developers.
My view: this course is a pragmatic response to sanctions pressure and an attempt to create a self-sufficient technological framework. But success will depend not on bans, but on whether domestic models can truly compete with foreign ones in terms of quality. Tax incentives are a good start, but without investment in fundamental research and personnel, the 34% gap in agentic functionality may persist.