JPMorgan Chase has blocked its Hong Kong staff from accessing Anthropic's Claude AI models, the Financial Times reported on Thursday. The move, which removes Claude from the bank's internal list of approved tools, signals how Wall Street firms are tightening controls on AI use in Asia as U.S.-China tensions escalate.
What prompted the restriction
The wording of Anthropic's usage terms in its licensing agreement with JPMorgan led the bank to pull the models from an internal drop-down list of approved large language models available to employees in Hong Kong, according to three people familiar with the matter. The bank declined to comment outside business hours. Anthropic did not respond to a request for comment.
Goldman Sachs made a similar move in April, removing Claude from the list of tools available to its Hong Kong-based bankers. Both decisions reflect a broader reassessment of how sensitive financial work intersects with AI tools hosted by U.S. companies in markets with complex geopolitical dynamics.
Wider regulatory pressure on U.S. AI firms
The restrictions come as the U.S. government intensifies scrutiny of AI exports. Earlier this week, Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei ordering the company to suspend exports of its Mythos and Fable AI models to all destinations and foreign nationals. Lutnick cited concerns the models could be used by military intelligence users in China, Russia, and other countries of concern.
President Donald Trump said on Wednesday that negotiations with Anthropic are "going fine." While U.S.-built AI models remain unavailable in mainland China, Hong Kong has historically served as a market where some models operate under usage limits set by the companies.
Why this matters for finance professionals
For compliance officers, risk managers, and trading desk leaders at global banks, the JPMorgan and Goldman moves are concrete examples of how geopolitical tensions are reshaping the tools available in key financial hubs. Banks have invested heavily in AI for Finance applications, but the pattern of withdrawals from the Hong Kong market makes clear that licensing terms and export controls will increasingly determine which models can run in which locations.
Finance teams that rely on third-party AI tools should expect more frequent reassessments of approved vendor lists in Asia-Pacific offices. The immediate task is to review current contracts for usage terms that could force a sudden removal, as happened with Claude. Relying on a single provider without a backup plan now carries added geographic risk.
Your membership also unlocks:
