The global race for artificial intelligence dominance is entering a new phase, where the primary competitive advantage is rapidly shifting from superior performance to dramatically lower costs. According to a recent Jefferies report, China's latest AI models are poised to reshape the enterprise AI landscape, presenting a significant challenge to established Western leaders.
At the forefront of this disruption is GLM-5.2, developed by Hong Kong-listed Z.ai (formerly Zhipu AI). Jefferies strategist Christopher Wood describes GLM-5.2 as another "DeepSeek moment," signaling a new wave of cost-effective AI. Industry feedback suggests that GLM-5.2 delivers performance close to Anthropic's leading AI systems, but at approximately one-quarter of the cost per token. This pricing advantage has profound implications for the industry, particularly as businesses increasingly prioritize return on investment, security, and affordability.
While companies like Anthropic and OpenAI have built their enterprise businesses on premium AI offerings, the emergence of powerful, lower-cost alternatives could dramatically alter market economics. Data from AI platform OpenRouter highlights this shift: Chinese AI models processed 21.37 trillion tokens in a single week in June, a sharp increase from 4.37 trillion in late April. During the same period, leading US AI models processed 5.76 trillion tokens, underscoring the rapid adoption of more affordable Chinese options.
This trend reinforces the gradual commoditization of large language models. As AI technology matures, factors such as pricing, deployment flexibility, and data privacy are becoming more critical competitive differentiators than raw model performance alone. Cheaper token costs may also encourage enterprises to deploy AI differently, potentially moving towards running smaller models on their own servers to enhance data protection and reduce dependence on public cloud providers.
Ironically, this shift towards cheaper AI models may ultimately benefit semiconductor companies. Citing the Jevons Paradox, Jefferies argues that lower AI costs will encourage broader deployment of AI across more applications, thereby driving higher overall demand for computing power, AI servers, and memory chips. Consequently, the brokerage remains bullish on AI hardware despite intensifying competition among model developers.
However, Jefferies cautions that the biggest long-term risk to the AI investment cycle is not the rise of Chinese models, but rather whether investors will eventually question if companies like OpenAI and Anthropic can generate sufficient returns on the massive sums being invested in AI infrastructure. Until that point, robust AI capital expenditure is expected to continue. In this evolving landscape, GLM-5.2's true disruptive power lies not just in its intelligence, but in its ability to challenge the industry's long-held premium pricing structures.