Anouncement

OKX Bans Claude in Hong Kong but Continues Aggressive Internal AI Integration

HONG KONG — Major cryptocurrency exchange OKX has restricted its Hong Kong-based workforce and employees traveling through mainland China from accessing Anthropic’s artificial intelligence chatbot, Claude.

The move follows a temporary suspension of OKX’s corporate account by Anthropic in early August, highlighting how regional compliance policies and escalating US-China technology tensions are creating operational hurdles for global financial firms.

Geographic Restrictions Trigger Compliance Adjustments

According to reports, Anthropic briefly suspended OKX’s corporate account after detecting access from locations not supported under the AI company’s terms of service. Anthropic’s usage policies strictly prohibit access from unsupported territories, a list that explicitly includes both Hong Kong and mainland China.

Although OKX’s corporate account was subsequently restored, the exchange chose to proactively halt Claude usage for staff operating within or traveling through China. Employees in those regions have seen their AI-assisted tasks rerouted to alternative large language model (LLM) providers that comply with regional access guidelines.

AI Remains Central to OKX’s Internal Operations

The restriction comes at a time when OKX has been aggressively integrating artificial intelligence into its core business practices. AI proficiency is no longer optional at the exchange; it is formally integrated into employee performance evaluations.

To maintain its competitive edge, OKX spends between $6 million and $8 million per month on various major LLM services. Previously, staff were encouraged to leverage both Anthropic’s Claude and OpenAI’s ChatGPT for daily operations, coding, and workflow automation. While Claude is now restricted in Hong Kong, OKX continues to utilize other LLM frameworks to ensure workflow continuity.

A Growing Trend Among Global Tech and Finance Leaders

OKX is not the only industry titan forced to navigate strict geopolitical boundaries and AI licensing constraints:

  • Goldman Sachs: Restricted its Hong Kong employees—specifically software engineers—from using Claude, extending the limitation even to international staff visiting the region.
  • Alibaba: Implemented an internal ban prohibiting its engineers from utilizing “Claude Code,” classifying Anthropic’s coding assistant as high-risk software.
  • ByteDance & Ant Group: Reported attempts by Chinese tech firms to access foreign AI models through overseas subsidiaries have faced mounting friction due to tightened compliance protocols from AI vendors.

The Broader Impact on Global Crypto & AI Integration

As foreign AI developers double down on geo-blocking to align with regulatory requirements, international tech and crypto firms operating across Western and Asian markets face a balancing act. OKX’s shift demonstrates how enterprise AI strategies must continuously adapt to local regulatory environments, ensuring compliance without sacrificing internal efficiency.

Related Posts