Citi: True Tech Stocks Have Low Weighting in Hong Kong Market; Hang Seng Unlikely to Rally Significantly in H2
Alina Collins
Citi strategist Liu Xianda says frontier-technology stocks account for just 4% of Hong Kong's market — far below the US, Korea, and Taiwan — making the HSI's structural weakness the key reason a significant second-half rebound is unlikely.
How small is Hong Kong's 'real tech' weighting?
Frontier-technology stocks make up roughly 4% of Hong Kong's market. The US, Korea, and Taiwan are far higher.
What Hong Kong labels "tech" is mostly platform stocks — companies like Meituan and JD.com whose revenues come from food delivery and retail, not AI hardware.
This means → when global capital chased AI-infrastructure plays — memory chips, optical communications, PCBs — Hong Kong had almost nothing to catch that flow.
A-shares have about 15% in real tech. Still low, but nearly four times Hong Kong's share.
Why did Citi cut its year-end HSI target?
Citi set a year-end HSI target of 30,000 at the start of the year, then cut it to 29,600 before H1 ended.
Three reasons drove the cut: interest-rate expectations shifted more than Citi foresaw, regulatory intensity exceeded forecasts, and platform stocks stayed under pressure.
This means → with all three headwinds stacked, Liu described the HSI reclaiming 30,000 in H2 as "a castle in the air."
Can the platform-stock weighting come down?
The Hang Seng Index Company is unlikely to directly cut platform-stock weightings. But as the constituent list expands, newly added pharma, industrial, and resource names will gradually dilute them.
If platform weighting falls from the current ~32% to 25%–28%, it would spread out some volatility and regulatory risk.
In plain terms = the index won't kick platform stocks out — it will water down their share by adding new members.
Because platform stocks still weigh heavily, Citi expects no major shift in capital style in H2.
What stage is the AI trade in now?
Liu divides the AI boom into three stages: Stage 1 (2023–~2025) was concept-driven speculation; Stage 2 chased upstream AI-infrastructure suppliers — chips, compute hardware; Stage 3, starting in H2 this year, shifts focus to AI monetisation — who can turn the concept into real revenue.
This means → the market will no longer pay a high premium for vision alone. Actual order volumes and revenue growth become the new yardstick.
Upstream compute hardware demand will persist, but gains may lag the past two to three years, and price pullbacks are possible.
How should investors pick IPOs?
Liu recommends focusing near-term on AI-themed IPOs. New listings with AI and semiconductor elements deserve a premium.
Traditional metrics still work: cornerstone-investor line-ups and oversubscription multiples remain key quality signals.
This means → concentrating on these IPOs is expected to deliver returns above the 6%–9% annual average from holding traditional blue chips.
Content is for reference only, not financial advice.