According to a CITIC Securities opinion cited by foreign media, the global AI boom this year resembles a "bottleneck trade" driven by infrastructure investment, rather than a technology bubble simply relying on high valuations. The article argues that this trend is closer to the bull market of 2006-2007, driven by heavy assets and investment; therefore, rising interest rates may not immediately mark a turning point.
AI market trends are more like investment cycles.
The article notes that the strongest performing sectors this year are mainly concentrated in hardware segments such as storage, optical communication, and semiconductor equipment, while platform companies and some AI application companies have been relatively weak. This suggests that the market's main focus is more on supply shortages and capital expenditure expansion, rather than simply betting on long-term prospects.
CITIC Securities also compared the current market trend to the dot-com bubble of 2000, noting that the forward valuation of the Nasdaq 100 was much higher than it is now, while AI supply chain companies in many parts of the Asia-Pacific region have seen significant gains, but their valuations have not spiraled out of control. Their assessment is that this round of gains is more driven by profitability and investment.
The impact of interest rate hikes depends on demand and spending.
The article argues that simply revising interest rate expectations upwards may not be enough to directly compress the valuations of "AI-related cyclical stocks." The real variables that will change the market are whether demand for AI devices slows, whether commercialization assumptions are weakened, and whether the growth rate of capital expenditures by large technology companies declines.
The article cites historical experience, noting that after the Federal Reserve began raising interest rates in 1999, the Nasdaq peaked relatively quickly; however, the more investment-driven bull market from 2004 to 2007 did not end immediately after the first rate hike. Based on this, CITIC Securities believes that if AI infrastructure investment continues to expand, the impact of rising interest rates on related sectors may be limited.
The pressure on A-share non-AI sectors was more pronounced.
The article states that while a clear divergence is emerging between AI and non-AI sectors globally, the divergence is more pronounced in the A-share market. Its calculations show that the valuation gap between AI and non-AI sectors in the A-share market is significantly higher than in the US, Japan, and South Korea, indicating a higher degree of valuation expansion in the local market and a weaker non-AI sector.
CITIC Securities believes that since May, A-share non-AI cyclical sectors have underperformed overseas markets, with two prominent pressures: first, the US dollar has strengthened again; and second, broad-based ETFs have continued to see net redemptions, which have suppressed traditional cyclical and financial sectors that lack new narrative catalysts.
- From June 15th to 18th, the four CSI 300 ETFs experienced a total net redemption of approximately 42.2 billion yuan.
- During the same period, the US dollar index strengthened, and the 10-year US Treasury yield fell.
Pay attention to the self-recovery signals of non-AI sectors.
The article argues that for non-AI sectors to recover from their weakness, the key lies not in waiting for a correction in AI stocks, but in improvements in their own fundamentals and funding conditions. For example, a drop in oil prices leading to lower inflation expectations, or a synchronized recovery in global non-AI industrial activity, could become new catalysts.
CITIC Securities maintains its "AI + Energy & Chemicals" investment strategy at the end of the article, remaining optimistic about some AI hardware chains, energy and chemicals, and some non-ferrous metals, while believing that the undervalued securities sector may see a recovery in the second half of the year.
