TL;DR
Chinese individuals are successfully bypassing Anthropic’s geolocation restrictions through VPNs, account reselling, and proxy networks. Despite enforcement efforts, demand for Claude remains high, especially among programmers.
Chinese users continue to access Anthropic’s Claude AI despite official restrictions and bans, employing increasingly sophisticated workarounds. This persistent bypassing highlights the high demand for Claude’s capabilities among Chinese programmers and tech enthusiasts, and underscores the challenges faced by Anthropic’s enforcement efforts.
Anthropic has implemented measures to restrict access to Claude within China, including banning accounts suspected of being controlled from the country. Despite these measures, Chinese users have found ways to circumvent restrictions, such as using VPNs, proxy networks, and purchasing pre-verified accounts on e-commerce platforms like Taobao and Xianyu. A growing underground market has emerged, with intermediaries selling API tokens and accounts, allowing users to access Claude more reliably. Anthropic spokesperson Michael Aciman confirmed that the company employs detection systems, including identity verification, to prevent unauthorized access and disrupt proxy networks, but these efforts have not entirely stopped the underground activity. The demand for Claude remains high, particularly among programmers who prefer it over domestic models for coding tasks, even as Chinese companies develop their own large language models. Experts note that Chinese developers tend to favor Western AI tools for their superior performance, especially in coding and development tasks, despite geopolitical tensions.
Impact of Bypassing Restrictions on AI Development and Security
The ongoing circumvention of Anthropic’s restrictions reveals the high demand for advanced AI tools like Claude in China and illustrates the challenges of enforcing export controls. This situation has implications for national security, as Chinese companies may use access to frontier models for training and development, potentially undermining US efforts to limit technological proliferation. It also highlights the resilience of the Chinese tech community in accessing and deploying Western AI innovations despite regulatory barriers.
VPN for accessing restricted AI models
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Chinese Market and International AI Competition
Over the past year, Chinese developers and startups have sought ways to access Western AI models, with many turning to underground markets and technical workarounds. Despite restrictions, Chinese companies like DeepSeek and Z.ai have made significant progress developing open-source models, but these still lag behind proprietary models like Claude in performance. The demand for Claude persists, especially among programmers who value its capabilities for coding. The US government has expressed concern over Chinese access to frontier AI models, citing national security risks, and has taken steps to restrict exports. Nonetheless, the Chinese AI ecosystem continues to innovate and find ways to access advanced models, reflecting a broader global competition for AI dominance.
“We use a range of evolving detection systems, including identity verification, to enforce our policies against unauthorized access to Claude.”
— Michael Aciman, Anthropic spokesperson
pre-verified AI account for China
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Extent and Effectiveness of Enforcement Efforts
It remains unclear how effective Anthropic’s detection systems are at completely stopping underground access, and whether new workarounds will emerge to bypass these measures. The full scale of the underground market and its future growth is also uncertain, as is the potential for government intervention or policy changes to curb such activities.
Ubiquiti Networks Unifi AC Mesh 1167Mbit/s Power Over Ethernet (PoE) White
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Future Developments in AI Access and Policy Responses
Expect ongoing efforts by Anthropic and other companies to improve detection and enforcement measures. Meanwhile, Chinese developers are likely to continue refining their workarounds, possibly leading to new underground markets. Policymakers on both sides may escalate restrictions or develop new regulations to address the persistent demand and circumventing tactics.
Mastering Claude Code Token Usage Optimization: Reduce API Costs, Extend Context Windows, and Build More Efficient AI Coding Workflows
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How are Chinese users accessing Claude despite restrictions?
They use VPNs, proxy networks, and purchase pre-verified accounts on e-commerce platforms. Some also rely on underground markets selling API tokens and accounts.
Why is Claude so popular among Chinese programmers?
Because it offers superior performance for coding and development tasks compared to domestic models, making it highly valued despite restrictions.
Anthropic employs detection systems, including identity verification and proxy disruption techniques, but these have not fully stopped underground access.
Could restrictions on Western AI models impact China’s AI development?
Yes, restrictions may slow access to cutting-edge models, but China’s developers continue to innovate and develop domestic alternatives, maintaining competitive progress.
Source: WIRED
