Why are Hong Kong Tech Stocks Lagging in the AI Boom?

Since 2026, global technology stocks have generally risen, with the Nasdaq Index soaring by 13%, the A-shares Science and Technology Composite Index surging by 28%, and the Growth Enterprise Market Index rising by nearly 20%. However, the Hang Seng Technology Index, which gathers Chinese internet giants such as Tencent, Alibaba, and Meituan, has dropped by over 13% against the trend, with trading volumes continuously shrinking.

This level of divergence in market trends is extremely rare in history, and it is not driven by a single emotional factor but has deeper structural reasons behind it.

Three Liquidity Pressures Hit Simultaneously

The Hang Seng Index is a typical offshore institutional market, with foreign capital holding a proportion of 47% in the Hang Seng Technology Constituents, plus mainland institutions in southbound funds, the combined holding ratio exceeds 80%. This means that the Hang Seng Index is much more sensitive to the global dollar cycle than A-shares.

Historical data confirms the validity of this logic. From July to September 2024, as U.S. inflation cooled and expectations for interest rate cuts rose, the US Dollar Index fell from 105 to 100, and the Hang Seng Technology surged by 34% within three months; from the beginning of 2025 to April, as the dollar trends weakened, Hong Kong stocks in technology once again experienced a trend-following rise. However, this effective logic encountered a reversal in 2026.

On May 22, Kevin Warsh took over as the Chairman of the Federal Reserve, and monetary policy shifted to a complex pattern of parallel balance sheet reduction and interest rate cuts. In April, the U.S. CPI was recorded at 3.8% year-on-year, and the PPI was as high as 6% year-on-year, with market expectations for the probability of interest rate hikes within the year climbing to nearly 70%. The dollar ended its decline and entered a period of fluctuation, foreign capital flowed back to the U.S., and the valuation of Hong Kong stocks continued to be under pressure.

At the same time, southbound funds are also quietly "changing their minds." Between 2024 and 2025, southbound funds净流入 Hong Kong stocks averaged over 15 billion Hong Kong dollars per month, but after entering 2026, the monthly net inflow has repeatedly broken through 10 billion Hong Kong dollars, and even several times net outflow occurred. The reason is direct and realistic: A-shares in semiconductors, optical modules, servers, and other AI-related targets have continued to set new highs, creating a strong siphon effect, and capital naturally migrates to markets with higher returns.

What makes things worse for Hong Kong stocks is the continued expansion at the IPO end. The queue of enterprises waiting to be listed on the Stock Exchange of Hong Kong has approached 500, and since 2026, the cumulative fundraising of the main board of Hong Kong stocks through IPO has exceeded 15 billion Hong Kong dollars. The inflow of incremental funds has slowed, and the stock funds are continuously diverted by new shares, so it is almost inevitable that the secondary market will be under pressure.

Betting on the "Wrong" Position in the AI Industry Chain

Apart from liquidity, the constituent structure of the Hang Seng Technology Index may be a deeper crux of the problem.

Comparing the Hang Seng Technology with the Science and Technology Composite Index, their positions in the AI industry chain are completely different. The top ten weighted stocks of the Science and Technology Composite Index are dominated by semiconductor and hardware companies such as Cambricon, Hygon, and SMIC. Semiconductors and hardware equipment have a combined weight of over 60%, mainly covering upstream infrastructure links such as optical modules, computing chips, and liquid cooling. The Hang Seng Technology Index is completely different, with Tencent, Alibaba, Meituan, Baidu, and JD.com forming the main body of the weight, with consumer retail plus software services accounting for nearly 50%, and semiconductors accounting for only 12.3%, mainly concentrated in the middle and downstream application levels such as AI large models, terminal ecosystems, and consumer internet.

The current pricing logic of the AI market just prefers the upstream. Since the start of this round of AI wave in 2022, the market has gone through stages of expected speculation, model competition, and hardware computing power resonance, and is currently in the "performance realization" window period. Capital prefers the upstream infrastructure track with clear profit expectations and verifiable financial data - this is the logic behind NVIDIA and AMD's continued outperformance of Microsoft and Alphabet, and also why A-shares optical