cd /news/large-language-models/jpmorgan-cuts-off-anthropic-access-f… · home topics large-language-models article
[ARTICLE · art-32383] src=thenextweb.com ↗ pub= topic=large-language-models verified=true sentiment=↓ negative

JPMorgan cuts off Anthropic access for Hong Kong staff

JPMorgan Chase has blocked its Hong Kong staff from using Anthropic's Claude AI model due to licensing terms, following a similar move by Goldman Sachs. The restriction highlights how geopolitical tensions and US export policy are shaping enterprise AI access in sensitive markets like Hong Kong.

read3 min views1 publishedJun 18, 2026

*JPMorgan Chase has stopped its staff in Hong Kong from using Anthropic’s AI models, according to the Financial Times, a quiet decision with loud implications. *

The bank removed Anthropic’s Claude from the internal drop-down list of approved large language models available to employees in the territory, the FT reported, leaving them without access to one of the tools their colleagues elsewhere can still use.

The reason, per the report, was not performance but paperwork. The wording of Anthropic’s usage terms in its licensing agreement with JPMorgan prompted the bank to pull Claude from the approved list in Hong Kong specifically.

The detail matters: this is a contractual and jurisdictional decision, not a verdict on the model, and it isolates the restriction to one of the most geopolitically sensitive financial hubs in the world.

JPMorgan is not the first major bank to do this. Goldman Sachs removed Claude from the list of approved tools available to its Hong Kong-based bankers in April, the FT noted, which makes JPMorgan’s move the second from a Wall Street institution in a matter of weeks.

Two of the largest US banks reaching the same conclusion about the same vendor in the same city is the kind of pattern that tends to precede an industry norm.

The backdrop is the steadily tightening US position on advanced AI and China. Washington has grown increasingly anxious about where the most capable American models end up and who can use them, and that anxiety has hardened into policy.

US Commerce Secretary Howard Lutnick ordered Anthropic to suspend exports of its most advanced models, including for foreign nationals, citing the risk that they could be diverted to military or intelligence users in China, Russia, and other countries of concern.

For banks, Hong Kong sits awkwardly in that picture. It is a Chinese territory and a global financial centre at once, and any tool whose licensing terms create ambiguity about cross-border data or access becomes a compliance question rather than a productivity one. Removing Claude from the approved list there is the cautious answer, the kind institutions reach for when the cost of getting it wrong is regulatory rather than commercial.

There is an irony in the timing for Anthropic, whose ties to Wall Street have been deepening even as its access narrows. The company has lined up Morgan Stanley and Goldman Sachs to lead its IPO, with JPMorgan also reported to be working on the deal, and has built enterprise partnerships across the finance industry. The same banks selling Anthropic to public markets are restricting its use in certain offices.

The contradiction runs deeper than IPO underwriting. Anthropic has spent the past year embedding Claude across the financial sector, from a $1.5bn enterprise joint venture with Wall Street partners to deployments inside banks piloting the model for accounting and compliance work.

The Hong Kong restriction does not undo any of that; it carves a single jurisdiction out of an otherwise expanding relationship. But it marks the first visible place where geopolitics, rather than capability or cost, sets the boundary of where Claude can be used.

For Anthropic, the timing is awkward in another sense. The company is courting public-market investors at a reported valuation in the hundreds of billions, a number that rests partly on the breadth of its enterprise reach. Each bank that pulls Claude in a sensitive market is a small dent in that story, and a reminder that a frontier AI vendor’s addressable market is now shaped as much by export policy as by product quality. The licensing language that tripped the restriction is the kind of detail that becomes load-bearing in a fractured regulatory landscape.

Neither JPMorgan nor Anthropic has commented publicly on the specifics beyond the FT’s account, and it is not clear whether the restriction will extend to other jurisdictions or be resolved through revised licensing terms.

What is established is the fact of it: a second Wall Street bank has pulled Claude from its Hong Kong toolset, and the reason points less at the technology than at the widening fault line running through it.

Get the TNW newsletter #

Get the most important tech news in your inbox each week.

── more in #large-language-models 4 stories · sorted by recency
── more on @jpmorgan chase 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/jpmorgan-cuts-off-an…] indexed:0 read:3min 2026-06-18 ·