Corporate Firewall: Goldman Sachs Blocks Access to Anthropic in Hong Kong

The Asian banking sector is going on the deep defensive. On April 29, 2026, it was reported that Goldman Sachs disabled access to Anthropic’s AI models (the Claude family) for its employees in Hong Kong.

The official reasoning cites data security issues and the specifics of corporate contracts. In reality, this is a direct consequence of recent statements by banking associations in Singapore and Japan, which have begun to view frontier models as potential cyber threats. Furthermore, amid US bans on transferring AI technologies to China (including Hong Kong), global banks prefer to build rigid internal firewalls (ring-fencing) to avoid cross-sanctions and prevent leaks of proprietary financial analytics.

Source: Reuters / Business Times