JPMorgan has released a research report indicating that China's AI ecosystem is far from a simple replication of overseas AI stocks. On one hand, Chinese optical module, PCB, and OSAT companies play a crucial role in the global supply chain, with leading firms deriving over 70% of their revenue from overseas. This positions them as low-cost, high-leverage instruments to capture the ongoing expansion of global AI computing expenditure, estimated at $800 billion.
On the other hand, a policy-driven domestic substitution cycle is creating an endogenous growth trajectory independent of global tech fluctuations. This cycle is expected to underpin a significant expansion in domestic AI computing capital expenditure from 2026 to 2029. Domestic inference token demand is projected to grow at a compound annual rate of 330% through 2030, with sales of AI servers powered by domestic GPUs forecast to rise by 41% in 2027 and 65% in 2028.
Furthermore, with few exceptions, forward 12-month price-to-sales valuations for Chinese AI companies across various segments of the AI value chain are now on par with those of global AI leaders. The rare combination of high growth, robust earnings visibility, and attractive valuations makes Chinese AI a significant source of differentiated alpha in global investment strategies. The report highlights JPMorgan's preferred stocks within the Chinese AI ecosystem.
The report notes that while AI adoption in China is still in earlier stages compared to global peers, a policy-led diffusion agenda and massive infrastructure development are creating a credible path for catching up. Global large language model (LLM) commercialization is rapidly expanding, with JPMorgan predicting the annual recurring revenue (ARR) for China's major LLMs could grow approximately four to sevenfold by 2026. Chinese LLM providers are competing with a value-oriented pricing strategy, aiding in the expansion of their global market share.
The bank identifies the core challenge as the sustainability of monetization. Rapid price reductions and low switching costs could compress profit margins. Without sustained state-of-the-art (SOTA) performance and workflow lock-in, valuation support may weaken. JPMorgan views pricing power as the most critical signal for valuation premium and, consequently, maintains an "Overweight" rating on KNOWLEDGE ATLAS (02513).
Comments