The OKX exchange has banned employees in Hong Kong from using the Claude chatbot

8/20/2026, 01:30 PMЕвгения Слив

The corporate account of the OKX crypto exchange on the Claude chatbot by Anthropic has been temporarily blocked due to a violation of the terms of use. As a result, the exchange has banned its employees in Hong Kong, as well as those who are traveling in China, from using this service, Bloomberg reports. The publication did not disclose the exact date of the block, but it is known that it happened in early August 2026. Following the appropriate response from the crypto exchange, access to the account was restored.

In response to the publication’s inquiry, company representatives reported that they had recently become aware of a possible partial non‑compliance of access to Claude for some OKX employees working in Hong Kong with Anthropic’s geographical access policy. The exchange intends to restrict access to Claude for employees in Hong Kong where necessary and will redirect their requests to other suitable artificial intelligence models. Anthropic indeed prohibits the use of its models in certain regions, and Hong Kong along with China are among them.

OKX CEO Star Xu commented on the situation, noting that the exchange spends between six and eight million dollars a month on several artificial intelligence providers. According to him, the company also has its own platform, Oli, which serves as a single access point to various models. Regarding the ban on using Claude in Hong Kong, Xu replied that the company is currently strengthening its internal access control mechanisms for corporate language models to ensure that Claude is used only by users in jurisdictions permitted under Anthropic’s policy.

OKX is not the only company that has faced similar demands from Anthropic. In late April 2026, Goldman Sachs was also forced to restrict access to the model for its employees in that geographic area. Earlier, Anthropic representatives stated that the United States could lose its leadership in the field of artificial intelligence as early as 2028. Such situations show that regional restrictions imposed by AI service developers continue to affect corporate users around the world.

Popular news