尽管Anthropic采取了严格的限制措施，试图阻止中国用户访问其开发的AI模型Claude，但在实践中这些防护机制经常失效。中国的初创公司、研究人员和技术爱好者通过使用虚拟专用网路（VPN）、国外电话号码及中介「转接站」等手段，开发出日益复杂的规避方法以获取Claude的使用权限。

这种对Claude的强烈需求催生了蓬勃发展的地下经济，用户在淘宝等电商平台和Telegram频道上购买已注册的帐号或API令牌。随著Anthropic引入身分验证（KYC）机制，黑市也迅速调整，开始贩售已通过身分验证的帐号，甚至提供绕过验证的解决方案。

中国的软体开发人员和学者表示，在编程和代码生成方面，他们显著偏好使用Claude等西方工具，而非DeepSeek等国内开发的模型。尽管Anthropic的管理层将中国对前沿模型的访问视为国家安全威胁，但技术人员仍持续寻求使用这些美国产品，即便这会面临潜在的资讯安全风险和诈骗威胁。

Although Anthropic has implemented strict restrictions to prevent users in China from accessing its AI model Claude, these safeguards have often failed in practice. Startups, researchers, and tech enthusiasts in China have developed increasingly sophisticated workarounds, such as using VPNs, foreign phone numbers, and intermediary "transfer stations" to gain access.

This strong demand for Claude has fueled a thriving underground economy, where users purchase registered accounts or API tokens on e-commerce platforms like Taobao and Telegram channels. As Anthropic introduced Identity Verification (KYC) mechanisms, the black market quickly adapted to sell pre-verified accounts and offer bypass solutions.

Chinese software developers and academics express a strong preference for Western tools like Claude over domestic models like DeepSeek for programming and code generation. Despite Anthropic's leadership viewing Chinese access to frontier models as a national security threat, technical users continue to seek these American products, even at the risk of security vulnerabilities and scams.