Chinese developers are increasingly relying on proxy-based relay platforms to access leading American AI models unavailable locally. These services circumvent official restrictions imposed by providers such as Anthropic and Google, supporting advanced tasks like coding and image generation amid a widening performance gap with domestic alternatives.
- Relay platforms route AI access through overseas proxy servers to bypass Chinese restrictions.
- Demand stems from higher accuracy and capabilities of foreign AI models versus domestic options.
- US AI companies have recently enforced stricter KYC and regional access controls.
What happened
In China, a shadow market for API relay services has emerged, enabling developers to access prominent AI models such as Anthropic’s Claude and Google’s Gemini despite official unavailability in the country. These relay platforms route user requests through proxy servers based outside mainland China, allowing local developers to circumvent restrictions and use these US models for various tasks including coding, debugging, and image generation.
Popular Chinese online marketplaces like Taobao and Xianyu host numerous sellers offering subscriptions to these relay services with claims of native-level access, high context window capacity, and seamless integration with popular development tools. Although providers like Anthropic have implemented stricter controls such as phone and identity verification, these relay services remain a key resource for developers facing limitations with domestic AI offerings.
Why it matters
The ongoing demand for relay APIs highlights a significant performance gap between Western AI models and local Chinese alternatives. Users report that models like Claude are more reliable and accurate, with fewer hallucinations or irrelevant outputs, improving complex engineering workflows. This demand drives a grey market despite increasing challenges from official restrictions and enhanced enforcement mechanisms imposed by AI providers.
At the same time, the relay infrastructure concept is gaining broader recognition globally, with new entrants such as Justin Sun and a US-based crypto company launching platforms offering access to multiple international AI models through unified APIs. However, the tightening of account verification and restrictions by US AI firms indicates growing friction in balancing open access with compliance and security concerns.
What to watch next
Key developments to monitor include how stricter know-your-customer (KYC) and identity verification measures impact the viability and scale of relay API services in China. Enforcement by companies like Anthropic has already led to more account bans and potential losses for relay resellers, suggesting heightened risks in this grey market.
Additionally, pricing strategies and market competition will be important indicators. Relay providers offer API usage at reduced rates compared to official pricing, attracting users but also raising questions about sustainability and regulatory responses. Meanwhile, growing interest in relay solutions internationally could drive innovation or prompt coordinated policy responses affecting cross-border AI model accessibility.