Anthropic released a listing prospectus on Monday after months of speculation, identifying more than a dozen risks connected to China. Most disclosures appear in the 93-page risk factors section of the document filed with the US Securities and Exchange Commission. Anthropic is preparing an initial public offering that could be the largest ever. Anthropic also outlined plans to invest more than US$500 billion in AI infrastructure over the next 20 years.
Open-weight models make their numerical parameters publicly available for users to download without charge and modify. China has become a global leader in open-weight models, helped by startups including DeepSeek and Z.ai. Anthropic and CEO Dario Amodei have repeatedly raised safety concerns about open-weight systems. However, the prospectus represents Anthropic's first formal acknowledgment that open-weight models directly threaten the company's business. Anthropic warned that open-weight or open-source alternatives could win customers for particular uses and put pressure on Anthropic's prices. Poe Zhao, a Beijing-based technology analyst who founded Hello China Tech, said competitors from China were pressuring Anthropic, especially among enterprise customers prioritizing lower costs over intelligence. Anthropic's flagship Claude Opus 5.5 charges around five times as much per token as Z.ai's leading GLM-5.3.
Anthropic identified Beijing-Taipei tensions as a supply-chain risk because advanced chip production is concentrated at Taiwan Semiconductor Manufacturing Company. Beijing considers Taiwan Chinese territory and maintains that reunification could involve force. Most countries, including the United States, do not recognize Taiwan as an independent country. However, Washington opposes a forcible takeover of Taiwan and has committed to supplying Taiwan with weapons. Amodei has been a prominent critic of China in Silicon Valley and has characterized China as a technologically advanced surveillance state. Anthropic supported US controls on exports of advanced chip-related technology through lobbying. However, the filing acknowledged that those restrictions could harm Anthropic's operations by changing the cost or availability of computing resources.
The prospectus discussed Anthropic's recent enforcement against companies from China allegedly training models with Claude outputs, a practice called distillation. Anthropic disclosed that accounts later determined to breach usage policies or commercial terms accounted for up to 7% of revenue during the year's first nine months. Anthropic warned that revenue from those accounts was likely to decline following stronger enforcement in August and September. The enforcement coincided with users reporting stricter access restrictions in mainland China and Hong Kong in recent weeks. Despite competition from China, Anthropic expressed confidence in growth, citing the company's dominance in segments requiring the highest intelligence.
The filing showed revenue of US$16.3 billion in the first half of the year, 14 times the amount recorded a year earlier. The financial disclosures put cash, cash equivalents and liquid investments at around US$132 billion on June 30. That total exceeded the holdings of hyperscalers including Amazon.com and Meta Platforms. Anthropic reported a six-month net loss of US$80.7 billion. Anthropic attributed US$72.4 billion of that loss to a nonrecurring accounting adjustment. First-half gross margins were about 85%, compared with 26% for Hong Kong-listed Z.ai, also called Zhipu AI, and 18% for Hong Kong-listed MiniMax. Zhao cautioned that the companies calculate these costs differently, while emphasizing the size of the margin difference.
Felix Wang, managing director and global tech sector head at Hedgeye Risk Management, said investors generally felt relieved that Anthropic's IPO appeared likely soon. Wang said many investors wanted the IPO to proceed immediately. However, Wang said Anthropic was unlikely to achieve the US$2 trillion valuation the company had reportedly sought. Wang attributed yesterday's market gains to relief over progress toward the IPO, while cautioning that completion was not guaranteed.
Anthropic's IPO Prospectus Highlights Multiple China-Related Risks
FLASH AIShare
Summary
Anthropic's IPO prospectus, filed with the SEC, reveals over a dozen China-related risks, including open-weight model competition and geopolitical tensions. Chinese startups like Z.ai are challenging its pricing and market position. The firm also cited potential revenue declines from stricter enforcement against Chinese firms using its models. Despite these concerns, Anthropic reported $16.3 billion in H1 revenue and $132 billion in cash and investments. Open interest in AI-related assets remains high, reflecting market uncertainty. Fear and greed index readings suggest mixed sentiment ahead of the IPO.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.