How does China's AI large language model industry compare to the U.S.
China's AI large language model sector is emerging as a formidable competitor to its American counterparts. Goldman Sachs recently released a detailed report highlighting this competitive positioning, demonstrating that Chinese AI firms are producing cutting-edge models at significantly lower prices. The report, authored by analyst Ronald Keung, provides a multi-dimensional analysis based on pricing, cost advantages, and financial health of the companies.
Identifying Leaders and Pricing Strategies
Within the competitive landscape of China's AI model industry, certain companies are establishing clear leadership. Zhipu and DeepSeek are recognized as the primary players in foundational text models, while ByteDance is leading in multimodal and video generation capabilities. The pricing strategy of these Chinese firms is noteworthy; high-end models are priced around $1 per million tokens, whereas their American counterparts can range from $4 to $8. This pricing disparity allows Chinese models to cost just 10-25% of what U.S. companies charge, with some models available as low as $0.06 to $0.20 per million tokens.
Technological Advantages
Chinese AI companies are successfully implementing more efficient architectures, often utilizing models that are merely 2-10% of the size of their U.S. equivalents. Techniques such as Mixture-of-Experts architectures allow these companies to activate only relevant sub-networks, conserving computational resources. Goldman Sachs has initiated coverage of Zhipu with a Neutral rating and estimated its valuation at $110 billion, while DeepSeek and ByteDance remain unlisted with positive outlooks for other firms like MiniMax and Kuaishou.
Projected Growth in the Sector
Goldman Sachs forecasts dramatic expansion in the Chinese AI model market, expecting revenues from APIs and subscriptions to soar from approximately 35 billion RMB in 2026 to an impressive 879 billion RMB by 2030. This 25-fold increase in daily token consumption reveals robust growth fueled by several factors. Open-weight models are fostering rapid developer adoption and creating an ecosystem that encourages retention. Additionally, enterprises are increasingly leveraging AI technologies, as they can access these solutions at a fraction of the costs associated with U.S. providers.
The landscape is shifting rapidly in favor of Chinese AI firms, and retail investors should monitor these developments closely to identify potential opportunities in this dynamic market.