JPMorgan Blocks Anthropic's Claude AI for Hong Kong Staff

JPMorgan Chase has removed Anthropic's Claude AI models from the list of approved tools available to its Hong Kong-based employees, according to a Financial Times report citing three people familiar with the decision.

The bank's move stems from concerns about the wording in Anthropic's usage terms within its licensing agreement with JPMorgan. The Claude models were pulled from an internal dropdown menu that Hong Kong staff use to access approved large language models for their work.

Wall Street pattern emerges

JPMorgan's decision follows a similar action by Goldman Sachs, which removed Claude from its approved tools list for Hong Kong bankers in April. The parallel moves by two major Wall Street institutions signal growing caution around AI deployment in regions caught between U.S. and Chinese regulatory spheres.

Neither JPMorgan nor Anthropic responded to requests for comment, and Reuters could not independently verify the Financial Times report.

Why it matters

These restrictions reveal how U.S.-China technology tensions are forcing global financial institutions to navigate complex compliance landscapes. Hong Kong occupies a unique position — while AI models from U.S. companies remain blocked in mainland China, the territory has served as a market where these tools can operate, albeit with usage limits. Banks must now weigh the risks of licensing language that could expose them to regulatory scrutiny or inadvertently violate emerging export controls.

Broader export control pressures

The banking restrictions come as the U.S. government intensifies oversight of AI technology exports. Earlier this week, Commerce Secretary Howard Lutnick ordered Anthropic to suspend worldwide exports of its Mythos and Fable AI models to all destinations and foreign nationals. The directive cited concerns that these models could be accessed by military intelligence users in China, Russia, and other countries deemed security risks.

President Donald Trump said Wednesday that negotiations with Anthropic regarding these restrictions are "going fine," though he provided no additional details.

Hong Kong's shifting AI landscape

While U.S. AI models remain unavailable in mainland China, Hong Kong has maintained relatively open access to these technologies. American companies typically impose their own usage restrictions in the territory. The recent decisions by JPMorgan and Goldman Sachs suggest that even this limited access window may be narrowing as geopolitical pressures mount.

The restrictions affect how employees in one of Asia's major financial centers can leverage AI tools for research, analysis, and client services — capabilities that competitors in less regulated markets may continue to access.

The details were first reported by the Financial Times.