On November 2, the volume of GEM hashrate ETF Huaxia surged 3.61%. As of midday trading, the turnover was 371 million yuan, ranking first among similar products in terms of liquidity. In terms of capital flow, net subscriptions have been obtained for 3 consecutive days, with the latest scale of 880 million yuan. Guolian Minsheng Securities said that the increase in domestic large-scale model calls is driving the continued expansion of demand for inference computing power, and model architecture optimization and special chip development are expected to further improve the efficiency of computing power usage. At the same time, supply and demand for HBM, server DRAM, and enterprise-grade SSDs are tight, and AI infrastructure construction is still in a booming stage. It is recommended to focus on domestic AI chips, servers, storage, optical communication and advanced packaging industry chains, as well as domestic computing power leaders with large-scale delivery capabilities. Guojin Securities believes that the DeepSeek open source inference framework accelerates the adaptation of domestic chips, that autonomous and controllable domestic computing power and domestic CSP inference side procurement jointly support server and IDC demand, that leading optical module repurchases stabilize market confidence, and that the demand for high-speed optical connectivity continues. GEM Computing Power ETF Huaxia closely tracks the GEM Computing Power Infrastructure Index. The index focuses on the AI computing power infrastructure industry chain, and is a convenient tool to lay out the core links of computing power hardware with one click.

On November 2, the volume of GEM hashrate ETF Huaxia surged 3.61%. As of midday trading, the turnover was 371 million yuan, ranking first among similar products in terms of liquidity. In terms of capital flow, net subscriptions have been obtained for 3 consecutive days, with the latest scale of 880 million yuan. Guolian Minsheng Securities said that the increase in domestic large-scale model calls is driving the continued expansion of demand for inference computing power, and model architecture optimization and special chip development are expected to further improve the efficiency of computing power usage. At the same time, supply and demand for HBM, server DRAM, and enterprise-grade SSDs are tight, and AI infrastructure construction is still in a booming stage. It is recommended to focus on domestic AI chips, servers, storage, optical communication and advanced packaging industry chains, as well as domestic computing power leaders with large-scale delivery capabilities. Guojin Securities believes that the DeepSeek open source inference framework accelerates the adaptation of domestic chips, that autonomous and controllable domestic computing power and domestic CSP inference side procurement jointly support server and IDC demand, that leading optical module repurchases stabilize market confidence, and that the demand for high-speed optical connectivity continues. GEM Computing Power ETF Huaxia closely tracks the GEM Computing Power Infrastructure Index. The index focuses on the AI computing power infrastructure industry chain, and is a convenient tool to lay out the core links of computing power hardware with one click.