AI-themed ETFs moved higher in the morning session, with E Fund Asia Semiconductor ETF (03486) rising 4.56% to HK$13.77 and E Fund Artificial Intelligence ETF (03489) gaining 3.12% to HK$10.57 as of the time of writing.
The E Fund Asia Semiconductor ETF focuses on memory, foundry, and advanced packaging, while the E Fund Artificial Intelligence ETF centers on core computing power plus mature applications, covering computing foundations such as Nvidia and Broadcom alongside application ecosystems including Microsoft.
On the news front, AI giant Anthropic released two new models on November 9 ahead of its IPO: Claude Fable 5.2 for the general public and Mythos 5.2 for cybersecurity and scientific research. The new models support long-cycle collaboration among 1,000 agents, such as completing a verification of Fermat's Last Theorem in 30 hours or developing an operating system from scratch in 8 days. Standard API pricing was cut by 20%, to US$8 for input and US$40 for output, and a faster Fast mode was also introduced. This launch marks the company's first new frontier model since CEO Dario Amodei called in September for slowing the pace of frontier AI development.
As of November 9, the three major U.S. stock indexes closed higher collectively, and the Philadelphia Semiconductor Index surged 6.85%. Notably, recent iterations of frontier models from Claude, OpenAI, and others have strengthened market optimism about long-term capital expenditure and revenue growth for AI supply chain companies such as Microsoft, Google, Nvidia, and Broadcom.
Some institutions believe AI computing power is undergoing accelerated restructuring, with custom chips becoming an important part of AI computing expansion. The ramp-up of ASICs is reshaping the computing landscape, and their large-scale deployment is gradually becoming an important choice for leading model developers and cloud providers seeking to optimize inference costs, energy consumption, and system efficiency. As customers including Google, Anthropic, OpenAI, and Meta advance multi-generation XPU projects, the AI chip market is forming a landscape in which GPUs and custom ASICs develop in parallel.
Comments