JPMorgan has quietly cut off its Hong Kong employees from Anthropic’s Claude AI model, according to people familiar with the matter. Staff in the city can no longer select Claude from the bank’s approved list of large language models. The move comes months after Goldman Sachs imposed a similar restriction — and it signals that the fine print in AI licensing agreements is starting to bite on Wall Street.
The ban stems from the terms of service in Anthropic’s licensing agreement, which explicitly excludes the Greater China region, including Hong Kong. Goldman earlier interpreted those terms just as strictly. US AI companies have been blocking access to mainland China and Hong Kong in part to prevent “distillation” — the practice of using frequent API calls to train a competing model from a frontier system. ChatGPT and Claude are both inaccessible directly in Hong Kong.
Multinational corporations have historically sidestepped such geographic blocks by signing global contracts and hosting compute outside the region. But the moves by JPMorgan and Goldman suggest that these contractual barriers are now trickling down to individual employees. With the finance industry increasingly relying on frontier AI models for tasks like coding, barring access could hurt Hong Kong’s competitiveness as an international financial hub.