Nvidia Enters the OpenClaw Battlefield! Tianrong Sci-Tech Chip Design ETF (589070) Valuation Already Significantly Lower Than Most Historical Periods, Pullback Presents Configuration Opportunity

Everyday Economics Editor: Xiao Ruidong

On March 12, the two markets bottomed out and rebounded, with chip design concepts declining. Regarding related ETFs, the Sci-Tech Innovation Chip Design ETF Tianhong (589070) closed down 1.64%, with a trading volume of 48.5608 million yuan and a turnover rate of 8.33%. Among its constituent stocks, Shengke Communication-U and Baiwei Storage fell more than 5%, with other stocks such as Anlu Technology, Xindong Lianke, and Zhenlei Technology also declining.

The Sci-Tech Innovation Chip Design ETF Tianhong (589070) has experienced a net capital inflow of 8.0455 million yuan over the past 30 trading days. As of March 11, 2026, the latest size of this fund was 589 million yuan, with a growth of 589 million yuan since the beginning of the year, ranking first among similar funds.

The Sci-Tech Innovation Chip Design ETF Tianhong (589070) achieves full coverage of three major sub-sectors by focusing on chip design companies listed on the STAR Market, enabling it to capture explosive growth in individual sectors while diversifying to reduce stock-specific risks. Currently, the semiconductor industry is in a recovery cycle, coupled with policy and demand benefits, further highlighting the investment value of the Sci-Tech Innovation Chip Design ETF Tianhong (589070).

The current PE-TTM of the Sci-Tech Innovation Chip Design Index is 188 times, at a historically very low level, only higher than 0.78% of the time. Compared with data from the past three years, the current valuation is significantly lower than most historical periods, demonstrating high cost-effectiveness for allocation.

Last night, NVIDIA launched a new generation open-source model, Nemotron 3 Super, specially designed for large-scale AI agents. It has 120 billion parameters, 12 billion active parameters, a 1 million token context, with inference speed tripled and throughput increased fivefold.

Guosheng Securities believes that NVIDIA’s GTC conference will unveil disruptive new chips and build a three-layer computing architecture for inference workloads. The shift from a one-size-fits-all approach to refined design indicates that chip design is evolving toward customization for specific workloads and heterogeneous integration.