Deutsche Bank strategists have highlighted that Chinese artificial intelligence (AI) models are rapidly closing the gap with US counterparts in terms of capability, while being priced similarly to mid-tier US systems such as Anthropic Sonnet. These Chinese models are increasingly released as open-weight systems, enabling developers and enterprises to download, modify, and run them locally, in contrast to the US approach of closed, proprietary models delivered via APIs [1].
The strategists note that the recent release of these Chinese AI models has already exerted downward pressure on AI and semiconductor stocks. Specifically, the Philadelphia Semiconductor Index fell by 9.97% last week, including a 1.63% drop on Friday, marking its largest weekly decline since the week of the Liberation Day tariff announcements last year [1]. This market reaction reflects investor concerns about the sustainability of current capital expenditure (capex) trajectories in the US AI sector, as similar performance may now be achievable at lower costs due to the competitive pricing and open nature of Chinese models [1].
At a macro level, Deutsche Bank suggests that the proliferation of open-weight Chinese AI models could accelerate global AI adoption, making it faster, wider, and cheaper, which would be positive for productivity. However, this development also raises the risk of a capex overcycle in the US if returns on AI infrastructure are pressured, and could intensify geopolitical fragmentation as competing technology stacks emerge [1].
CONCLUSION
The rapid advancement and open-weight release strategy of Chinese AI models are reshaping the competitive landscape, pressuring US AI and semiconductor stocks and raising questions about the sustainability of current investment levels. While this could boost global AI adoption and productivity, it also introduces risks of overinvestment and increased geopolitical fragmentation in the technology sector.