Anthropic, the developer of the advanced artificial intelligence model Claude, is facing a wave of criticism amid mass bans of user accounts. As of October 2026, account holders worldwide report a sharp tightening of security measures aimed at detecting and stopping the circumvention of regional restrictions using virtual private networks (VPNs). This situation has affected numerous users from countries and territories where official access to the service is initially restricted.
Large-Scale Outages and User Complaints
The first alarming signals came from users in Hong Kong, where the service does not officially operate. One account holder reported that they had successfully interacted with the neural network through a popular VPN service for about six months before receiving an official notification from the security team about the irrevocable cancellation of their account due to "suspicious signals." Similar cases have been recorded in other jurisdictions with comparable bans.
Situation in Russia and Expert Opinions
The issue has also affected Russian users, despite the service not being officially supported in the country. In particular, Roman Maresov, head of the e-commerce block at Yandex, encountered the unavailability and blocking of his Claude account. Experts note that Anthropic's security algorithms have learned to effectively detect attempts to mask IP addresses, threatening any workaround methods for domestic developers and specialists.
Contradictory Data
There are certain discrepancies in the assessment of the scale of events between the expert community and the corporation. An official and detailed press release from Anthropic regarding a change in internal VPN policy has not been published as of the current date. While independent analysts and affected users claim an unprecedented punitive campaign against geo-fencing workarounds, company representatives have so far limited themselves to standard responses regarding violations of the safety user agreement.
Consequences for the Industry
The tightening of control by Claude's creators demonstrates a general trend among leading artificial intelligence market players toward strict compliance with licensing and geographical restrictions. This could lead to a significant outflow of audiences from unauthorized regions and stimulate the search for alternative local or less IP-restrictive language models.
