On July 17, the Federation Council gave the green light to a framework law regulating the development, implementation, and circulation of artificial intelligence technologies. This regulatory act is not just another bureaucratic formality, but a fundamental step in shaping the national legal architecture for a rapidly developing industry. The bill was introduced in the State Duma on June 25, passed its first reading on July 7, and is now one step away from final approval.
Sovereign and National Models: Two Levels of Protection
The key innovation is the legislative codification of the concepts of "sovereign" and "national" AI models. The law covers large fundamental models (LFMs) — programs with a billion or more parameters capable of solving complex tasks at a human level. Clear and stringent requirements are established for them:
Sovereign model: The developer must be exclusively a Russian legal entity with full control at all stages. Data placement and reproduction are allowed only in domestic data processing centers (DPCs). The use of foreign components is completely excluded.
National model: The developer must be a Russian legal entity. The infrastructure must be domestic, but the use of foreign components under open licenses is permitted.
Both types of models are required to strictly comply with Russian legislation and the country's traditional spiritual and moral values. In return, developers gain access to unprecedented state support measures: financial, property, guarantee, and informational. Moreover, they will be granted the opportunity to legally train neural networks on data from federal and regional information systems. The specific procedure for accessing this data will be determined by the government after coordination with security agencies.
Content Labeling and the General Trend Toward Increased Control
A separate provision of the law obliges owners of popular internet resources with a daily audience of over 500,000 users to implement tools for labeling materials created by neural networks. The format of such labeling will be developed through an agreement between service developers and consumers.
These changes are part of a general trend toward strengthening state presence in the technology sector. Simultaneously, the government is actively integrating dozens of AI services into its own apparatus, while the Ministry of Digital Development is designing national platforms for managing communication networks. The Federal Antimonopoly Service has even proposed recognizing the circulation of AI technologies as a strategic activity — an initiative sharply opposed by the RATEK association, which warns of a risk of electronics shortages due to vague wording.
Against this backdrop, other controversial issues are brewing in the industry. The media community demands the protection of copyright when training neural networks. Deputies are discussing a possible restriction on access to social networks and chatbots for children under 14 in 2027-2028, which is planned to be synchronized with the launch of the school profile "Artificial Intelligence" starting September 1, 2026.
My analysis: The adoption of this law is a logical and timely step, but it only sets the general framework. The key intrigue lies in the by-laws, which will determine how stringent the requirements for "sovereign" models will be and how exactly the mechanism for accessing state data will work. For the crypto industry and DeFi, this could mean tighter control over AI services used in financial operations, but it also opens a window of opportunity for developing fully domestic, secure solutions.