How China ‘risk factors’ loom over Anthropic’s hotly anticipated IPO filing

Anthropic, a Silicon Valley start-up focused on artificial intelligence, has faced numerous 'risk factors' related to China in its recent IPO filing. These risks, detailed in a 93-page section of the US Securities and Exchange Commission filing, highlight the potential commercial threats posed by China's rapidly growing open-weight AI model industry.

Open-weight models, whose parameters are publicly available for users to download and modify, are particularly proliferating in China, with companies like DeepSeek and Z.ai leading the charge. Anthropic CEO Dario Amodei has previously warned of the safety risks associated with open-weight systems, but the company's filing marks the first time it has acknowledged the direct commercial threat these models pose.

The prospectus also identifies geopolitical tensions between China and Taiwan as a supply-chain vulnerability, as Taiwan is a crucial hub for advanced chip manufacturing. Anthropic CEO's criticism of China as a "hi-tech surveillance state" has been echoed in the filing, alongside the company's recent crackdown on Chinese companies using its models to train their own systems.

Despite these challenges, Anthropic remains confident in its growth trajectory, boasting a 14-fold surge in revenue to $16.3 billion in the first half of the year and a substantial cash reserve of $132 billion. However, analysts expect investors to welcome the imminent IPO, although the anticipated valuation may not match the $2 trillion target initially sought.