Putin orders a state plan for AI as Russian business shuts down 90% of pilots

The Russian AI market is simultaneously accelerating and rigorously weeding out redundant experiments. Authorities are preparing a national plan for implementing neural networks, regions are scaling applied use cases, and major businesses increasingly acknowledge: only a small portion of pilots reach full launch.

State Plan and Rules

After AI Journey 2025, Vladimir Putin tasked the government with preparing by June 1, 2026 a national plan for implementing AI in the economy, state administration, and social sphere. An interdepartmental commission under the president should oversee coordination. In parallel, "Rosatom" and "Rosseti" received the task to approve data center infrastructure development plans for 2030–2036 taking into account future demand from AI services.

This is no longer talk about a trendy topic, but an attempt to lay out computing, energy, and responsibility between departments in advance. Another important layer is regulation. The Ministry of Digital Development published a draft law introducing definitions of "sovereign," "national," and "trusted" AI models.

For the first time at the legislative level, the very concept of artificial intelligence is fixed, and mandatory labeling is proposed for content created by neural networks. Citizens want to be given the right to challenge decisions made by state bodies using AI without court, and responsibility for unlawful results is to be distributed among the developer, operator, service owner, and user. The planned date for these norms to take effect is September 1, 2027.

Where AI Already Works

At the regional level, the story looks not theoretical, but quite applied. According to the digest, 55 regional governments have already implemented AI in their operations, with healthcare, state administration, and security emerging as the most active areas. The AI Development Center under the government reviewed around 400 solutions, more than 250 of which passed verification and made it to the "Digital Region" platform.

This is an important signal: the bulk of implementations in Russia is currently happening not through high-profile consumer products, but through concrete services where effects can be quickly seen and economics calculated. This is particularly noticeable in medicine, though public trust in such systems remains low. Only 1% of Russians are ready to fully trust AI with health matters, but in practice doctors already use neural networks as a second opinion.

Moscow AI-based services processed 8.6 million studies in 2025—almost three times more than Moscow's own patient flow. A separate indicator came from the IQDOC service: from February 6 to March 16, 2026, almost 3.

9 thousand medical workers asked the system about 25 thousand questions, and 87% of requests concerned diagnosis and treatment tactics.

• 55 regions already use AI in state authority operations
• More than 250 solutions passed verification for the "Digital Region" platform
• 8.6 million medical studies were processed by Moscow AI services in 2025
• About 25 thousand questions were asked to the AI system by doctors in one and a half months
• 87% of such requests were related to diagnosis and treatment

Pilots, Business and China

The most harsh part of the picture—corporate implementations. A study of 50 largest companies showed that only 7–10% of projects with LLMs, chatbots, and AI agents reached full industrial use. The rest remain in pilot, are reformatted, or closed entirely.

Moreover, 30–40% were halted precisely due to poor integration into real business processes. This is an important fork: the problem is often not in model quality, but in the company's failure to embed it into regulations, data, KPIs, and team routines. That said, talking about a reversal is premature.

97% of major Russian companies continue to work on AI implementation, half of industrial players are already ready to apply generative models in production, and 42% use such tools in automation. Where scaling did succeed, companies record output growth and productivity gains up to 20% and resource optimization up to 15%. Against this backdrop, a separate line is the political and technological bet on China: Putin tasked the government and Sberbank with developing cooperation with the PRC and other BRICS countries, and Sber directly points to DeepSeek as a model for cheaper and more pragmatic solutions.

"DeepSeek proved that even with curtailed capabilities, you can achieve quality comparable to

American standards."

What This Means

The Russian AI market is entering a selection phase where glossy presentations quickly give way to infrastructure, legal frameworks, and measurable benefits. The state is building rules and data centers, regions are accumulating working cases, and business is learning to close decorative pilots and keep only what embeds into processes. If the line toward local models and cooperation with China strengthens, the market will acquire a more pragmatic and less showcase-oriented growth trajectory.