Anthropic Faces Challenges from Rising Chinese AI Competition Ahead of IPO
The increasing popularity of Chinese artificial intelligence models poses a significant risk for Anthropic as it prepares for its initial public offering (IPO), according to US research firm Hedgeye Risk Management.
Felix Wang, managing director and global tech sector head at Hedgeye, stated, “It’s all about Chinese competition when analysing Anthropic. Many of my clients are asking the same question.” This assessment comes amid renewed investor concerns regarding whether Anthropic’s substantial AI investments will yield adequate returns.
Recently reported financial figures have led to sell-offs of major AI chipmakers, including Nvidia and Intel. Anthropic informed investors at a private briefing last week that its third-quarter revenue rose nearly 50 percent from the previous quarter to US$16.8 billion, as reported by a source familiar with the matter.
Despite this quarter-on-quarter increase, the growth rate has slowed compared to the explosive expansion seen in the first half of the year, when revenue more than doubled in the second quarter as companies rushed to purchase subscriptions for its Claude Code coding agent tool.
Impact of Chinese Alternatives
This deceleration coincides with a notable rise in the global popularity of Chinese open-weight models. AI research group Exponential View reports that these models now account for approximately half of all AI tokens, which are the core units of AI usage worldwide.
While Chinese AI startups still lag behind their US counterparts in total revenue, they have experienced rapid growth over the past year. For example, Beijing-based Z.ai has seen its annual revenue run rate increase more than sixfold since March, driven by soaring demand for its GLM models.
Importantly, the Chinese AI industry has managed to grow sales while steadily increasing prices. Z.ai, also known as Zhipu, has raised prices for its model services at least three times this year, while Hangzhou-based startup DeepSeek more than doubled prices for its flagship V4 model in August.
Despite these price increases, Chinese models remain among the most affordable on the market and are only a few months behind leading US models in terms of capabilities, according to third-party benchmarking firm Artificial Analysis.
Wang noted, “Total demand for your tokens may be going up, but if the price is flat or falling, then you’re not making more money.” This situation indicates that Chinese open models are not only capturing market share from Anthropic but are also influencing AI pricing dynamics.
Challenges Ahead for Anthropic
Hedgeye’s internal models suggest that Anthropic has experienced a slowdown in the third quarter as growth in enterprise AI spending has cooled. Data from C-suite executives supports this finding, with Boston Consulting Group’s latest “Applied AI Index” indicating that nearly three-quarters of the world’s most sophisticated enterprise AI users have imposed usage limits or spending controls.
This trend implies that even as token demand continues to rise, downward pricing pressure is hindering Anthropic’s revenue potential, according to Wang. He emphasized that the competitive landscape is rapidly evolving, particularly with the resurgence of domestic rival OpenAI, which has seen increased demand for its coding tool Codex and new flagship model GPT-6 Astra.
Anthropic’s IPO is anticipated next month, although some investors are now discussing a reduced valuation of around US$1.5 trillion, down from the previously sought target of approximately US$2 trillion, as reported by Reuters.
Meanwhile, OpenAI has postponed its own IPO until next year. These challenges have led some analysts to predict a potential market downturn for the US AI industry, with Jefferies’ global head of equity strategy Chris Wood citing cheaper Chinese models as a key factor in his forecast of “massive capital destruction” in the US.
However, it remains uncertain whether Chinese models can maintain their competitive advantage against US counterparts. Anthropic has indicated that its crackdown on distillation has contributed to its recent revenue slowdown, with this campaign reportedly impacting top-line figures by 5 to 7 percent.
Wang noted that many analysts have underestimated the extent to which China’s AI progress has relied on distillation. While Chinese companies are expected to remain competitive, a successful crackdown by Anthropic could compel these firms to invest more in research and development, potentially affecting their profit margins.
Wang concluded, “The China dynamic is very fast-moving when it comes to analysing Anthropic and OpenAI.”
Source: SCMP Tech (Asia)