JPMorgan Chase has stopped its Hong Kong staff from using Anthropic’s artificial intelligence models, according to the Financial Times. The bank removed Anthropic’s Claude from its list of internally approved large language models available to employees in the region, leaving them without access to one of the tools that their colleagues elsewhere can still use.
See also: Anthropic provides access to Project Glasswing to 150 additional companies
The reason for this decision was not performance, but bureaucracy. The wording of Anthropic’s terms of use in its licensing agreement with JPMorgan prompted the bank to delist Claude in Hong Kong specifically. This is a contractual and jurisdictional decision, limiting the restriction in one of the most geopolitically sensitive financial hubs in the world.
JPMorgan is not the first major bank to take this action. Goldman Sachs removed Claude from the list of approved tools available to its Hong Kong-based bankers in April, making JPMorgan’s move the second by a Wall Street institution in as many weeks.
The backdrop to this situation is the steadily hardening US stance on advanced artificial intelligence and China. Washington has become increasingly concerned about where the most capable American models end up and who might use them, and that concern has hardened into policy. US Commerce Secretary Howard Lutnickordered Anthropic to suspend exports of its most advanced models, including foreign ones, citing the risk that they could be diverted to military or espionage users in China, Russia and other countries of concern.
See also: JPMorgan hires Nomura's head of AI strategy
For banks, Hong Kong presents a complicated scenario. It is a Chinese territory and a global financial center, and any tool whose licensing terms create ambiguity about cross-border data or access becomes a compliance issue rather than a productivity one. Removing Claude from the approved list there is a cautious response, one that institutions choose when the cost of getting it wrong is regulatory rather than commercial.
There’s an irony in the timing for Anthropic, whose ties to Wall Street have deepened even as its access has narrowed. The company has lined up Morgan Stanley and Goldman Sachs to lead its IPO, with JPMorgan also reportedly working on the deal and has formed partnerships across the financial industry. The same banks that sell Anthropic in the public markets are restricting its use to certain offices.
The contradiction goes deeper than the IPO. Anthropic has spent the last year embedding Claude across the financial sector, from a $1.5 billion venture capital venture with Wall Street partners to deployments within banks that are testing the model for accounting and compliance.
The Hong Kong restriction does not negate any of this; it isolates a single jurisdiction from an otherwise expanding relationship. However, it marks the first visible example where geopolitics, rather than capability or cost, sets the limit on where Claude can be used.
See also: JPMorgan urges redefinition of SaaS security
For Anthropic, the timing is awkward in another sense. The company is approaching investors in the public markets at a reported valuation in the hundreds of billions, a figure based in part on the breadth of its business. Every bank that pulls out Claude in a sensitive market is a minor blow to this story, and a reminder that the market that a supplier of advanced artificial intelligence can address is now shaped as much by export policy as by the quality of the product.
