AI Iteration Risks Surface As Hong Kong Market Diverges; Low‑Valuation, High‑Dividend Legacy Stocks Attract Capital As Safe Havens

As fund managers confront the reality that one technology can supplant another, a pronounced shift toward risk aversion has emerged, prompting public funds to favor traditional sectors. Concerns over AI‑driven substitution have already precipitated dramatic declines in certain U.S. biotechnology names heavily held by QDII funds, with some falling nearly 60% within two trading days.

On February 24, that defensive posture became evident in Hong Kong, where low‑valuation, high‑dividend, small‑cap legacy stocks rallied and assumed the role of capital safe havens. These companies generally lack exposure to AI, while high‑growth segments that rely on technology narratives but have yet to deliver earnings faced correctionary pressure. The result was a marked increase in defensive stock selection among public funds.

The market exhibited sharp structural divergence. Stocks that had been overlooked by institutional investors and long sidelined now staged strong recoveries, while previously favored technology and biopharma names experienced deep pullbacks. Despite continued attention to large AI models over the holiday, post‑holiday sentiment turned distinctly defensive, and traditional sectors such as shipping, education, food and beverage, advertising and construction machinery led the advance, forming the primary buying force on the tape.

Illustrative moves included CIMC Group (02039.HK), a major holding of GF Fund, which jumped about 11.91% in a single session; Zhenjiu Li Du (06979.HK), held by Harvest Fund, which rose nearly 5%; Yum China (09987.HK), a notable position for manager Zhang Kun, which gained 2.55%; and Smoore International (06969.HK), held by Qianhai Kaiyuan Fund, which increased roughly 1.8%. These names share attributes of earnings visibility, relatively low valuations and higher dividend yields, aligning with the post‑holiday preference among public funds for steadier returns.

Small‑cap stocks that had been deeply oversold also attracted concentrated buying. Jiahao Education (01935.HK) surged about 10% on the day and has climbed more than 75% since early February, yet its market capitalization remains below HKD 2.5 billion. Andeli Juice (02218.HK) rose approximately 2.23%, bringing its year‑to‑date gain to about 19% and its one‑year advance to roughly 90%, while its free‑float market value is near HKD 1 billion. Advertising firm Duoxiangyun (06696.HK) rallied about 7% amid micro‑cap momentum, with February gains of 137% and a three‑month rise of 332%, though its market cap still sits under HKD 1.5 billion. The concentration of flows into these traditional, small‑cap names reflects both a hunt for high‑beta opportunities and a broader defensive stance.

By contrast, popular growth sectors suffered notable declines, particularly where rapid AI iteration raises the prospect of product obsolescence. In healthcare, advances in AI‑driven, noninvasive cancer screening—such as breakthroughs in CT imaging AI and capsule endoscopy for lesion detection—triggered severe disruption in the global blood‑based cancer screening space. U.S. company Grail (GRAL.US), once a heavy QDII holding for some mainland funds, plunged about 60% across February 20 and 23.

The shockwaves reached Hong Kong peers. On February 24, blood‑testing leader Mirxes‑B (02629.HK) plunged intraday by as much as 26% and closed down about 21.67% as investors rushed to price in substitution risk. Competition among AI large‑model providers remains intense and price pressure is mounting, heightening concerns about future profitability and market structure. On the same day, enterprise AI services provider Paradigm Intelligence (06682.HK) fell 10.35%, medical AI platform Ark Healthcare (06086.HK) declined 7.42%, Healthroad (02587.HK) dropped 8.79%, and AI marketing software firm MaiFushi (02556.HK) lost 8.60%.

Widely held technology names also corrected sharply. Kingdee International (00268.HK), Meitu (01357.HK), China Literature (00772.HK), Mobvista (01860.HK), Damai Entertainment (01060.HK), Kingsoft Cloud (03896.HK), SenseTime and XtalPi (02228.HK) each fell more than 5%, with constituents of the Hang Seng Tech Index (800700.HK) among the hardest hit. Since October, the Hang Seng Tech Index has been in a downtrend, diverging from the AI narrative; ETFs tracking the index have declined roughly 20% over the past four months. Market participants worry that rapid AI advances in content generation, knowledge retrieval and office automation could displace existing software and internet services, eroding user engagement and undermining the long‑term growth thesis for many incumbents.

This dynamic is mirrored internationally. On February 24, India’s Nifty IT Index experienced a sharp sell‑off, falling nearly 5% intraday and closing lower for the fifth consecutive session. Reports from U.S. investment institutions warn that widespread AI adoption could exert pressure on white‑collar employment and compress the software sector, posing risks to outsourcing revenues for major Indian vendors.

Against this backdrop of elevated AI‑related volatility, several public fund managers signaled a renewed emphasis on absolute returns and a tilt back toward traditional consumption. Cui Zhipeng of Huashang Fund expects food and beverage and other conventional consumer segments to reach a turning point in 2026 as fundamentals stabilize and valuations remain at multi‑year lows. He notes that China’s service consumption share still lags developed markets and that services could become a key driver of domestic demand and employment.

Huang Yisong, manager of Penghua Consumption Fund, advocates continued exposure to structurally advantaged consumer themes—youthful demographics, discretionary self‑care spending, domestic brand substitution and improved value propositions—focusing on food, jewelry, beauty and small appliances while adding gaming exposure. For core, defensive holdings he prioritizes industries with sound competitive structures and stable long‑term demand, applying bottom‑up stock selection and a buy‑and‑hold approach to pursue sustainable, steady returns.