JPMorgan blocks Hong Kong staff from Anthropic models

JPMorgan Chase has reportedly removed Anthropic's models from the approved large-language-model menu for staff in Hong Kong, making it the second major Wall Street bank to restrict Claude access in the city this spring. Reuters attributed the report to the Financial Times and said the bank's move was prompted by wording in Anthropic's licensing agreement with JPMorgan. 1

The immediate issue appears contractual, not a model-performance problem. Reuters said JPMorgan removed Claude from an internal drop-down list of approved LLMs for employees in the Asian financial hub; JPMorgan and Anthropic did not respond to Reuters' requests for comment, and Reuters said it could not immediately verify the FT report. 1

This is not a broad retreat from enterprise AI, but it is a clean signal that Anthropic's enterprise distribution now has a jurisdictional fault line. Goldman Sachs removed access to Claude for its Hong Kong bankers in April; at that time, Reuters reported that other models such as Gemini and ChatGPT were still available on Goldman's internal platform, and that Anthropic's spokesperson told the FT that Claude models had never been officially "supported" in Hong Kong. 2

The JPMorgan report also lands during the separate Fable 5 and Mythos 5 dispute. Anthropic said on June 12 that a US government directive forced it to suspend access to those two models for all customers, while access to other Anthropic models would not be affected. 3 Reuters later reported that Commerce Secretary Howard Lutnick ordered the suspension because officials feared the models could be used by military intelligence users in China, Russia, or other countries of concern. 4

The distinction matters. JPMorgan's Hong Kong restriction is being reported as a licensing and approval-list decision; the Fable/Mythos shutdown is a government directive. Together, they point to the same operating problem for Anthropic: selling Claude into regulated financial institutions is no longer just a product-adoption question. It also depends on how banks interpret model access, geography, employee location, and US policy risk.

The next signal is whether JPMorgan's restriction stays limited to Hong Kong or is resolved through revised licensing terms. A broader pattern across banks would be more damaging than a single-office cutoff, especially while Anthropic is pitching Claude to regulated industries through partners such as TCS and DXC. For now, the verified fact is narrower: two large US banks have pulled Claude from Hong Kong employee toolsets, and the latest report remains unattributed beyond FT's unnamed sources and Reuters' follow-up.