Tech stocks continued their downward trend on Friday, extending a difficult week for the sector. While much of the initial sell-off was attributed to reports from the New York Times suggesting that OpenAI might delay its initial public offering following SpaceX’s lackluster post-IPO performance, analysts suggest a deeper structural shift may be at play. A significant emerging threat to U.S. AI dominance appears to be the rise of highly efficient Chinese artificial intelligence models.
The recent launch of GLM5.2 by Hong Kong-listed Z.ai (formerly Zhipu AI) has caught the attention of market strategists. According to Christopher Wood at Jefferies, the new model performs at a level nearly equal to Anthropic’s offerings but at a mere fraction of the cost. This significant price disparity could allow international competitors to capture substantial corporate market share from U.S. frontier models ahead of their anticipated IPOs. Wood characterized the current market sentiment as “another DeepSeek moment.”
Morgan Stanley traders noted that Z.ai’s new model demonstrates “very impressive coding capabilities.” However, they suggested that the trend of enterprises and hyperscalers migrating toward more affordable models represents a “recalibration in willingness to pay” rather than a decline in overall AI demand. The cost of high-end AI, measured in tokens, has become a critical concern for businesses that have recently faced budget exhaustion in their AI spending. This has fueled fears of a price war between major players like OpenAI and Anthropic, potentially compressing valuations prior to their public listings.
The emergence of high-performance, low-cost models poses a systemic risk to the current AI investment thesis. Jim Reid of Deutsche Bank noted in a recent report that certain models can perform the majority of everyday tasks at approximately 1.5% of the cost of leading-edge models like Anthropic’s Claude. Furthermore, Jefferies analysts indicated that Z.ai’s model likely provides privacy protections comparable to top-tier U.S. models. This capability could encourage corporations to migrate AI workloads from cloud service providers back to local servers, fundamentally altering the projected investment landscape for AI infrastructure.
As the demand mix shifts toward cost-effective solutions, concerns are mounting that the massive growth forecasts for AI infrastructure spending may be overly optimistic. Market reaction has been swift: Micron shares fell 7% in early trading, while AMD and Intel both dropped by more than 4%. Oracle also saw declines, marking a 19% loss over the past five days.
