Crypto exchange OKX has restricted employees in Hong Kong and staff traveling through China from using Anthropic’s Claude after its corporate account was temporarily suspended, even as the company spends between $6 million and $8 million every month across multiple artificial-intelligence models.
The restriction followed a brief suspension of OKX’s corporate Anthropic account in early August, Bloomberg reported. Access has since been restored, but OKX will prevent affected employees from routing requests through Claude and instead direct their work to alternative AI models.
OKX said it recently became aware that some employees’ Claude usage in Hong Kong may not have fully complied with Anthropic’s regional access policies. Anthropic does not permit Claude access in either Hong Kong or mainland China, according to the company.
The episode is significant because OKX operates globally, with major offices spanning Hong Kong, Singapore, Dubai, London, California, New York, Malta and Istanbul. Its dependence on multiple large language models also illustrates how rapidly generative AI has become embedded within major crypto companies.
OKX Is Spending Up to $96 Million Annually on AI Models
OKX’s disclosed monthly AI-model spending translates into an annualized run rate of approximately $72 million to $96 million if expenditure remains at current levels. Rather than depending exclusively on Claude, the exchange uses multiple large language models, giving it the ability to redirect Hong Kong employees toward alternative systems following Anthropic’s restriction. OKX has also been recruiting senior AI personnel, including advertising for a senior director or vice president of AI position covering Singapore and Hong Kong during 2026.
The scale of expenditure reflects a wider increase in enterprise spending on generative AI as companies deploy models for software development, customer operations, research and internal productivity.
For OKX, maintaining a diversified set of AI providers now carries an additional strategic benefit: geographic resilience. Regional restrictions imposed by one provider do not necessarily prevent employees from accessing competing models permitted in the same jurisdiction.
The exchange said it appreciated Anthropic’s engagement in resolving the corporate-account suspension and indicated that its response was designed to ensure compliance with model providers’ policies.
Hong Kong Restrictions Are Affecting Other Financial Firms
OKX is not alone in encountering Claude’s geographic limitations.
Goldman Sachs previously removed Claude from the AI tools available to Hong Kong employees after determining that its Anthropic agreement did not cover use in the territory. JPMorgan subsequently restricted Claude access for Hong Kong staff after reviewing similar licensing terms.
Those decisions illustrate an increasingly complicated operating environment for multinational financial and technology companies. U.S.-developed AI models are becoming critical enterprise infrastructure at the same time that Washington and Beijing are tightening controls around advanced artificial intelligence, semiconductors and computing capacity.
Hong Kong presents a particularly important test because it remains a major international financial and cryptocurrency hub while falling within geographic restrictions imposed by some U.S. AI companies.
For OKX, the immediate operational impact appears manageable because Claude represents only part of a broader portfolio of models. The larger implication is that AI-provider policies can now determine which software tools employees of a global crypto exchange can use depending on where they are physically located.
With OKX spending as much as $8 million monthly on AI services, model availability has therefore become more than an employee-productivity issue. It is increasingly an infrastructure, compliance and geopolitical consideration for the company’s global operations.