Goldman Sachs Group Inc. has pulled the plug on its Hong Kong branch’s access to Anthropic’s Claude, a generative‑AI tool that streamlines software development. Staff in the Asian hub can no longer rely on the agent that had been helping engineers draft code faster. The decision follows a brief internal review to enhance internal controls.

Claude, launched by San Francisco‑based Anthropic, gained traction in financial services for its rapid code generation and risk‑analysis capabilities. Goldman’s move signals a cautious stance toward deploying third‑party AI in sensitive research environments. The firm’s decision could ripple across its global teams, prompting a review of vendor agreements and compliance frameworks to ensure data security.

The suspension comes amid heightened scrutiny over AI reliability and data privacy. Analysts note that removing Claude may slow product development cycles, but it also reduces exposure to potential algorithmic bias and compliance violations. Investors will watch whether the firm’s broader AI strategy shifts toward in‑house solutions or alternative partners to maintain competitive edge globally.

For the Hong Kong office, the immediate impact is a halt on using Claude for coding tasks, potentially delaying project timelines. The broader market watches as banks weigh the trade‑off between cutting‑edge AI productivity and regulatory defensibility. This case underscores the growing importance of vendor vetting in tech‑heavy finance to protect client data and compliance standards.