Huge underground market emerges in China for Claude AI access

While the US company Anthropic is trying to strictly limit Chinese users' access to its Claude AI bot, a large-scale underground ecosystem for bypassing these bans has emerged in the country. According to an investigation by Wired, the sale of ready-made accounts, fake identities, and specialized intermediary services for accessing the Claude system have become an integral part of the Chinese tech market. This is reported by Ixbt.com news reports.

Anthropic, founded by former OpenAI employees, is known for its extremely serious approach to safety policies. The company has closed official access for Chinese users, and it is not limited to geoblocking or IP address checks. They actively block suspicious profiles using VPN and proxy, and recently began requiring identity verification via government documents through the Persona service.

Methods of bypassing bans and "relay stations"

However, Chinese users have found effective ways to circumvent these barriers. The simplest scheme involves using foreign phone numbers and payment cards, while more complex methods include purchasing pre-verified accounts on Taobao, Xianyu, and Telegram channels. Cybercrime experts note a steady growth in demand for Claude Pro and Claude Max subscriptions on the black market.

In particular, services called "relay stations" have become popular. These platforms are located in countries where Claude officially operates, purchase corporate access to Anthropic's API, and then resell it to users within China. For the customer, it looks like a simple chat interface: the request is sent to a local site, from there it goes to the Claude system via an intermediate server, and the response is returned.

Among developers and tech company employees, the Claude model is more valued than local analogs like DeepSeek or Z.ai. According to Oxford University researcher Zilan Qian, Chinese specialists emphasize that national language models lag behind US developments by 6 to 9 months, especially in programming issues. Therefore, Claude Code and similar tools have become a vital necessity for complex IT projects.

Geopolitical competition and security issues

Against the backdrop of the technological cold war between the US and China, Anthropic CEO Dario Amodei calls access to advanced models by Chinese companies a threat to national security. In his view, Chinese companies may use Claude's responses to train their own models (distillation method). Recently, Anthropic even openly accused the Chinese giant Alibaba of such practices.

Interestingly, the imposed restrictions are only prompting the emergence of new workarounds. After the requirement for document-based identity verification was introduced, the trade in pre-verified "digital identities" flourished. This situation is also distorting global statistics. For example, Singapore, with a population of 6 million, is among the world leaders in Claude system usage, while experts believe a large part of this traffic actually comes from China.