Anthropic Cuts Off Hong Kong Users Who Reached Claude by VPN

Anthropic, the American company behind the Claude AI assistant, has started cutting off users in Hong Kong who were sneaking in through VPNs. A VPN is a tool that makes your computer look like it's somewhere else — in this case, Singapore or the US. On Thursday, a wave of complaints appeared on X and Reddit from people whose accounts had worked fine for months and then were suddenly disabled. One user said they'd used Claude through NordVPN for six months before getting an email from Anthropic's safeguards team about "suspicious signals."

Claude isn't officially sold in Hong Kong, so tech-savvy users had been working around the rules: routing their connections through other countries and paying with foreign bank cards. That workaround is now collapsing. Anthropic appears to have gotten better at spotting when a connection is really coming from a restricted place, and it's acting on what it finds — permanently, and without much warning.

Why should you care if you don't live in Hong Kong? Because this is a preview of how AI companies are starting to treat accounts in general. Access is tied to where you live, what card you pay with, and whether your behavior looks "normal" to an automated system. If you travel a lot, work abroad, or use a VPN for privacy, you could trip the same wire. And when an account is closed for a policy violation, you often lose your saved chats and work along with it.

The practical takeaway is simple: don't treat any AI assistant as your permanent filing cabinet. Export important conversations, keep your own copies, and don't assume an account you've paid for will always be there. AI tools feel like utilities, but right now they're closer to subscriptions that can be cancelled from the other side.