Although the latest policy became effective Thursday, its central restrictions are not entirely new. Anthropic previously introduced limitations on companies controlled by organizations in unsupported jurisdictions, and the November update provides more explicit guidance on how those requirements apply.
The policy establishes three conditions that can prevent individuals or businesses from accessing Claude. The first concerns physical location: anyone using Claude from an unsupported region falls outside the permitted access rules, even if they work for an organization headquartered in a supported country. Employees traveling to restricted locations are therefore subject to the same geographic limitations.
The second condition applies to businesses incorporated or headquartered in unsupported regions. Their employees and other users remain ineligible regardless of where they are physically working. The third extends the restrictions to companies with majority ownership or control linked to unsupported regions, including businesses that operate elsewhere but are directly or indirectly controlled by individuals or organizations in those locations.
These distinctions are particularly relevant to international businesses with complex ownership structures. A company incorporated in a supported country does not automatically qualify for Claude access if its controlling interests are based in an unsupported region. The restrictions also extend to employees and other users associated with those organizations, even when they work from countries where Claude would otherwise be available.
However, the policy does not prohibit access solely because a business has Chinese investors or employs Chinese nationals. Its language focuses on geographic location, corporate structure and control rather than citizenship or the existence of any minority investment. Control also operates as a separate condition from majority ownership, meaning a minority shareholding does not necessarily establish eligibility when another arrangement gives an entity in an unsupported region control over the business.
The requirements extend beyond Anthropic's own applications. Companies accessing Claude models through third-party cloud platforms or authorized resellers remain subject to the same regional eligibility rules, so using another provider does not create an exemption.
Anthropic first introduced its corporate ownership restriction in September 2025, when it announced that it would stop selling to businesses more than 50% owned, directly or indirectly, by organizations headquartered in jurisdictions where its products were prohibited. China was specifically identified in that announcement.
At the time, Anthropic cited concerns about laws in certain jurisdictions that can require companies to cooperate with intelligence agencies or disclose information. The company also pointed to potential military and intelligence applications of its technology, along with the risks of model distillation and the use of Claude to support competitors of American and allied technology companies.
Anthropic revisited its regional access rules on October 8 as part of a broader Usage Policy update. The company characterized much of that revision as an effort to clarify existing requirements and explain their enforcement rather than introduce an entirely new set of restrictions.
That distinction became relevant after reports circulated on Russian Telegram channels on November 11 suggesting that Russian companies and their overseas employees were facing newly imposed restrictions. Russian-language publication Meduza subsequently noted that the ownership-related limitations dated back to the previous year and pointed to an earlier version of Anthropic's supported regions documentation, dated May 12, 2026, which contained the same list of eligible countries and territories.
The November 12 effective date therefore marks the implementation of updated policy language rather than the introduction of every restriction it describes. For businesses already affected by the ownership rules, the latest version makes the conditions governing access more explicit.
Enforcement remains an important consideration for users and organizations. According to TokenPost, Anthropic began introducing identity verification measures in April that can require users to submit government identification and a selfie through Persona. Suspected access from restricted regions or unauthorized resale of Claude services may trigger additional verification, while accounts that fail those checks could face closure.
However, the available reporting does not establish that Anthropic has initiated a new round of account terminations specifically tied to the November update. The published rules also do not explain how the company evaluates every potential ownership arrangement or resolves disputes involving indirect corporate control. They establish eligibility requirements without detailing how individual organizations will be assessed.
Beyond regional access, Anthropic's October Usage Policy revision addressed several other areas of model use. The company introduced restrictions concerning deceptive campaigns and artificial activity, revised provisions involving elections, surveillance and weapons-related guidance, and updated its approach to higher-risk applications in healthcare, finance and legal services. The changes also included a prohibition on sustained abusive behavior directed at AI models.
Other AI providers are making changes to how their technology can be accessed through third parties. TokenPost reported that OpenAI announced on October 21 that platforms offering direct access to its GPT models would need advance authorization beginning November 20.
For businesses using Claude, Anthropic's regional policy makes eligibility dependent on more than where an employee opens the application. Corporate incorporation, headquarters and ownership or control can independently determine whether an organization is permitted to use the service, including when Claude is accessed through an outside provider.
The updated policy formalizes those requirements across Anthropic's product offerings, but it does not establish that every affected organization has already lost access or that a new enforcement campaign began on November 12.
This analysis is based on reporting from Tech Times.
Image courtesy of ITdaily.
This article was generated with AI assistance and reviewed for accuracy and quality.
