On November 12, 2026, the American company Anthropic, developer of the Claude family of neural networks, implemented an updated regional access policy for its services. The key change is that restrictions now apply to organizations from countries where Anthropic does not officially provide services — primarily Russia — regardless of where their employees are physically located. In essence, the company has shifted from a "user geography" model to a "jurisdiction and control of the organization" model, which deprives Russian entities of workarounds involving overseas workers.
What exactly has changed in the regional policy
Previously, Anthropic focused primarily on the geographic location of the individual user: accounts linked to unsupported regions were blocked. The new policy revision shifts the focus to the organization itself. Companies registered or headquartered in countries where Anthropic services are not provided are now banned. Restrictions also apply to entities that are directly or indirectly controlled by individuals or legal entities from such states. Thus, a Russian organization will not be able to use Claude even if access to the service is registered to an employee working abroad.
Perimeter of restrictions: from Claude.ai to cloud platforms
The new terms cover the entire spectrum of the company's products: the Claude.ai web interface, the Claude Code developer tool, the Cowork service, the API, as well as models available through cloud platforms and authorized resellers. This means that restrictions cannot be bypassed by moving to third-party cloud providers or services operating on an agency model — all access channels are subject to a unified regional policy.
Context: passport verification and waves of blocking
The tightening of the policy is occurring against the backdrop of ongoing control processes. According to industry reports, at the end of October, Claude began mass requests for identity verification via the Persona service — uploading a photo of a document and a live selfie; Russian passports, according to users, in the vast majority of cases led to account blocking, including paid Pro and Max subscriptions. Separately, there were reports of blocking profiles associated with Russian media resources, and in September, Anthropic published a report on cases of misuse of models, including attempts to extract code potentially related to the development of biological weapons. These episodes form the background against which the company is consistently narrowing access to its services.
Anthropic's position: clarification, not a new ban
The company emphasizes that this is not about introducing a ban from scratch, but about clarifying and expanding the application of previously existing regional restrictions. Russia is not officially on the list of supported countries, so formally, the new package of rules only makes existing restrictions stricter and closes the gap related to the use of services through employees located abroad. For Russian organizations, this means the actual loss of legal access to Claude and related tools regardless of the structure of their international presence.
