(This article is authored by Wandian Research and published on Titanium Media with authorization.)
Article by Wan Dian Research, author: Feng Yan
From the 18C to the AI Index, the Hong Kong Stock Exchange has completed the chain of technology assets.
On October 23, the Hong Kong Exchanges and Clearing announced the official launch of the HKEX Artificial Intelligence Index and the HKEX Semiconductor Index. The Artificial Intelligence Index tracks 40 Hong Kong‑listed companies closely linked to the AI industry chain, spanning a five‑tier architecture that includes chip design, data centers, AI infrastructure, models and software, and real‑world applications. The Semiconductor Index tracks 30 of the largest Hong Kong‑listed semiconductor‑related companies, covering manufacturing, equipment, components, computing hardware, and storage devices.
The Hong Kong Stock Exchange also announced the signing of a licensing agreement with Southern Fund, under which the latter will launch an ETF in mainland China that tracks AI-related indices. Specifically, both major indices include leading companies in their respective sectors, such as SMIC, Lenovo Group, Xiaomi Group, Innoscience, Acacia Communications, Biren Technology, and Tensilica.
Relevant primary-market investors stated that the two major indices have established new pricing and allocation benchmarks for the Hong Kong‑listed AI and semiconductor sectors, which are expected to further steer capital toward leading companies in these areas, enhancing the sectors’ representativeness and liquidity. With the indices now officially launched and ETF authorizations granted, the Hong Kong‑listed AI and semiconductor sectors have gained new investment tools and a fresh avenue for capital inflows.
1 HKEX Launches Thematic Indices: AI and Semiconductors Gain New Pricing Benchmarks
The Hong Kong Stock Exchange has launched two major thematic indices, ostensibly to provide the market with additional benchmarks, but in reality, it is making strategic moves along the productization chain of technology assets.
Objectively speaking, for a long time, the authority over Hong Kong stock market technology-themed indices has been largely held by Hang Seng Indexes Company. According to a report by 21st Century Business Herald, as of the end of June 2026, the total assets under management of passive products tracking the Hang Seng Tech Index had reached US$40.4 billion (approximately RMB 272.5 billion), nearly 26 times the US$1.5 billion at the index’s launch in 2020.
Additionally, according to disclosures by Hang Seng Indexes Company, as of the first half of the year, the total assets under management of ETP products tracking the Hang Seng Index Series had exceeded US$130 billion, roughly 1.5 times the size of Hong Kong’s domestic ETP market.
The Hong Kong Stock Exchange has independently developed an AI and semiconductor index, signaling that it is no longer content to merely provide a trading venue; instead, it seeks to extend its “listing‑location advantage” upstream into index construction and downstream into product licensing.
The AI index ETF authorization granted to Southern Fund marks the first milestone in this chain. Yet its starting point is a market that has already been deeply anchored by hundreds of billions of dollars in passive capital.
The compilation details are worth examining closely. The AI Index does not simply rank stocks by market capitalization; instead, it sets clear fundamental thresholds: constituent stocks must meet the following criteria: R&D expenses and capital expenditures as a percentage of sales must each be no less than 3% over the past two years, and sales growth must be at least 5%. In terms of weighting, no single constituent stock may exceed 10%, and individual stocks in the application segment are capped at 3%; the index is rebalanced quarterly.
The message conveyed by this set of rules is that the Hong Kong Stock Exchange seeks a benchmark with “substantial AI investment,” rather than a basket that merely chases conceptual hype. Meanwhile, the Semiconductor Index uses a sampling criterion based on the “30 largest companies,” emphasizing size and representativeness, thereby establishing a methodological division of labor that complements the AI Index’s fundamental‑oriented approach.
However, the index construction rules themselves also harbor inherent tensions. The 10% cap on each constituent’s weighting means that large-cap stocks like SMIC, Lenovo, and Xiaomi cannot dictate the index’s direction; instead, the “average performance” of the 40 constituents will determine the index’s overall outcome. In an environment of sharp divergence within the AI sector, a well-balanced allocation may not outperform an active strategy that concentrates bets on a handful of top performers.
The semiconductor index is constructed using a size-based sampling approach, which essentially reflects the current market landscape rather than forecasting future industry leaders. When the largest constituent stocks happen to be experiencing capital outflows, the index’s short-term performance will inevitably suffer. This is not a flaw in the methodology but rather an inherent pro-cyclical characteristic of thematic indices.
For investors, until ETF products are officially launched and have been operating for a while, the new index serves primarily as a benchmark for observing the structure of Hong Kong‑listed technology stocks, rather than an immediately deployable allocation tool. Southern Fund’s authorization marks a starting point, but the timing of product fundraising, subsequent asset size, and tracking error are the true measures of the index’s practical utility.
2 From 18C to the thematic index: HKEX’s technology asset service chain is gradually becoming more complete
Zooming out, this move by the Hong Kong Stock Exchange is embedded within a two-tier structure.
The first layer is the ongoing shift in the underlying structure of the Hong Kong stock market. The share of technology stocks in total Hong Kong stock trading volume has risen from 15% a decade ago to 44%, with trading value increasing by more than sevenfold; in the first half of 2026, science and technology enterprises accounted for 63% of the total IPO fundraising scale on the Hong Kong stock exchange, compared with just 16% during the same period last year.
According to data from the Hong Kong Stock Exchange, from December 2025 to May 2026, AI value-chain companies raised a total of HK$97.9 billion through new share offerings, accounting for 55% of the total IPO proceeds on the Hong Kong stock market during the same period, with an additional 70 AI value-chain companies currently awaiting listing.
Chapter 18C of the Listing Rules is an innovative listing regime designed specifically for special-purpose technology companies, which was first proposed in a consultation paper issued by the Stock Exchange of Hong Kong in October 2022 and officially launched in March 2023.
According to data from the Hong Kong Stock Exchange, under the 18C regime, as of August 31, 2026, 23 specialized technology companies have raised a total of HK$59.426 billion through refinancing, surpassing the combined IPO proceeds of HK$40.358 billion, demonstrating that the ongoing financing channel is now fully operational.
At this point, the Hong Kong Stock Exchange’s official compilation of an AI and semiconductor index represents a formal institutional acknowledgment of this shift. It complements the 18C index: while 18C addresses “how to bring companies in,” the thematic index tackles “how they should be priced and allocated once they’re listed.”
The second-layer backdrop is even more complex. Sino-U.S. technological competition persists, semiconductor export controls show no signs of easing, and achieving self-reliance in AI computing power has become a policy priority, while Hong Kong is expected to assume the role of “an offshore pricing center for China’s technology assets.”
According to Wind data, in 2025, southbound funds recorded net purchases exceeding HK$1.4 trillion for the full year, hitting a record high. Since July 2026, the scale of southbound net buying has continued to expand, with cumulative net purchases surpassing HK$100 billion since then, serving as a stabilizing force during market downturns.
A source familiar with the securities industry told Wan Dian Research that assessments of the new indices should be differentiated across three time scales: at the institutional level, the Hong Kong Stock Exchange has filled a gap in standalone hard‑tech benchmarks, which is the right direction and aligns logically with the 18C framework; at the product level, the AI index has already seen its first launch, while the semiconductor index remains to be seen—this divergence in pace itself conveys important signals; and at the market level, the recent pullback has, paradoxically, created an opportunity for investors to build positions at lower levels in products linked to these new indices. For investors aiming for long-term allocation, this presents a favorable window to gradually accumulate shares in stages and wait to fully establish positions once the related products are officially launched.
Institutional dividends are long-term, product implementation is medium-term, and market sentiment is short-term. Confusing the three can easily lead to overlooking short-term practical obstacles under the halo of institutional significance. A more pragmatic approach is to treat the new index as a long-term benchmark for assessing the structure of Hong Kong‑listed technology stocks, and to wait until the linked products are in actual operation and their scale and tracking data are available before determining whether it is worthy of being used as an allocation tool.
3 Beyond the long-term institutional benefits, the new index still needs to overcome this hurdle.
However, before the long-term institutional benefits take effect, short-term pressures at the constituent stock level have already begun to surface.
The “leading stocks” newly included in the index happen to be concentrated in the sectors currently experiencing the most significant capital outflows and valuation contractions, thereby imposing realistic constraints on the index’s short-term outlook.
Taking Biren Technology as another example, Wind data shows that after reaching a peak of HK$70.75 in June, the company’s share price has been on a steady downward trend. As of October 26, its Hong Kong-listed shares closed at HK$23.70, marking a cumulative pullback of over 65% from the peak.
As the stock price came under pressure, the company conducted two discounted share placements within less than a year of its IPO: in July, it raised approximately HK$7 billion by placing shares at HK$46.20 each, and in October, it placed 130 million shares at HK$31.08 each. Together, these two rounds have netted the company over HK$17 billion from the market.
Notably, there appears to be a certain “unspoken understanding” between the lifting of share lock-up and the placement of shares at Wall仞 Technology. On July 1, the cornerstone investors’ lock-up period expired, and the company announced a share placement three days later; on October 6, the lock-up period for the placed shares ended, and another placement followed two days after that.
As a result, the market has developed an expectation that “once restrictions are lifted, shares will be dumped,” with a lock-up expiration window still ahead in January 2027, covering 33.47% of the company’s total share capital held by founders and pre-IPO shareholders.
Biren’s predicament is far from unique; it reflects a common paradox faced by domestic AI chip companies: technological catch-up requires sustained, massive investment, while the capital markets’ patience rapidly erodes when stock prices decline.
Zhongji Xuchuang represents another type of pressure. As a leading player in the optical module sector, it is one of the most direct beneficiaries along the AI computing infrastructure value chain, yet its stock price has also been subject to sharp volatility.
Wind data shows that on October 22, Zhongji Xuchuang’s share price fell nearly 8.7% in a single day, erasing all of its year-to-date gains and hitting a new low since April. Its financing balance has dropped from a peak of HK$31.577 billion in mid-July to around HK$16.2 billion, shrinking by nearly 50% over the past three months.
The continued withdrawal of margin financing reflects a reassessment of the market’s enthusiasm for AI investments. Demand for optical modules is closely tied to capital spending by North American cloud providers; whenever there are any signs of a slowdown in the global pace of AI infrastructure development, companies that “sell shovels” tend to be among the first to feel the impact of shrinking valuations.
The broader market environment is equally grim. According to a Goldman Sachs prime brokerage report, from October 8 to 14, hedge funds engaged in net selling of global AI stocks for five consecutive trading days, wiping out within a week the cumulative net purchases made since early 2025. The Nasdaq-100 index posted its largest weekly decline since April 2025, while chip stocks closed lower for eight straight trading sessions.
The market’s focus has shifted from “what can AI do?” to “when will the investment pay off?”: Ultra-large cloud providers have pledged capital expenditures on an unprecedented scale, yet their monetization pathways remain unclear. In response, Broadcom’s CEO warned that funding challenges could become the biggest hurdle to AI expansion.
This context has a direct implication for the HKEX’s new index: a substantial proportion of its constituent companies derive their valuations from the assumption that AI demand will continue to expand. If global AI-related trading enters a prolonged cooling phase, even if linked products are launched, they may still face the reality of sluggish capital inflows.
4 A heavyweight card—can Southern Fund turn it into a royal flush?
In summary, Southern Fund’s AI ETF authorization is a positive signal, but the long-term viability of an index product ultimately hinges on post-listing asset growth and secondary-market liquidity.
The Hong Kong Stock Exchange has entrusted the licensing of its AI-indexed mainland ETFs to Southern Fund, rather than to other leading institutions—a choice that warrants close scrutiny. The company’s general manager, Yang Xiaosong, described this collaboration as “an important step in implementing our internationalization strategy and meeting investors’ needs for diversified asset allocation.”
From a strategic perspective, this indeed represents a differentiated positioning for Southern Fund in the cross-border index‑product space: the HKEX AI Index comprises 40 constituent stocks, spanning chip design, AI infrastructure, and the full stack of models and applications, thereby distinguishing itself from existing domestic AI‑themed ETFs in terms of both constituent composition and narrative framework. Securing exclusive licensing and enjoying a first‑mover advantage means that, for now, Southern Fund faces no direct competitors under the “Hong Kong‑listed AI” sub‑category.
However, the flip side of these opportunities lies in the numerous constraints at the execution level. First and foremost is the issuance window. According to Interface News, in the third quarter of 2026, the number of newly launched public funds reached a new high since 2022, yet the average fundraising size was only about RMB 477 million—the lowest since 2005. Meanwhile, more than 70% of equity ETFs were established just at the minimum threshold. Among the 15 non‑initiated funds launched since October, 10 had assets under management below RMB 300 million, accounting for 67%.
Among Southern Fund’s own product lineup, the Southern ChiNext Software ETF, launched on October 12, raised only about RMB 220 million, just above the RMB 200 million threshold. Under these issuance conditions, even if an AI‑index ETF were to launch, whether its initial fundraising can reach a meaningful scale is the first practical challenge Southern Fund must confront.
In Wan Dian Research’s view, an ETF that barely meets the size threshold will struggle to generate effective liquidity in the secondary market, no matter how meticulously its index is constructed. Meanwhile, at present, only Southern Fund offers AI-index–linked products on the Hong Kong Stock Exchange, and more issuers are needed to help “grow the pie.”
The liquidity ecosystem of a thematic ETF often requires multiple institutions to deploy simultaneously in order to take shape; while exclusive licensing provides a competitive advantage, it may also be a disadvantage when it comes to ecosystem development.
More importantly, from a timing perspective, the HKEX AI Index was launched precisely at a moment when the technology sector was undergoing a significant correction, with constituent stocks such as Biren Technology and Acbel Precision Industry facing downward pressure on share prices and capital outflows.
If Southern Fund were to launch an AI‑index ETF with significant marketing efforts at this juncture, it would need to strike a balance between “seizing strategic opportunities” and “avoiding issuance at elevated price levels.” This is not to say that the product should not be launched; rather, the timing of the offering, the anticipated scale, and the approach to investor communication all require greater prudence than during periods of heightened market enthusiasm.
From a longer-term perspective, the true value of Southern Fund’s recent authorization may lie not in the fundraising size of the first ETF, but in whether it can build sustainable operational capabilities across its cross-border index product lineup. The HKEX AI Index rebalances its constituents quarterly and sets fundamental thresholds for R&D investment and capital expenditures, with its constituents including AI‑related companies listed on both the A‑share and Hong Kong stock markets.
This means that the product’s day-to-day management involves adjusting its constituent stocks across markets, accounting for exchange-rate factors, and controlling tracking error arising from differences in trading rhythms between the two regions. Southern Fund has already built up some experience in the cross-border ETF space, but the operational ramp-up period for a completely new underlying asset is unavoidable.
The first product’s tracking error data, sample‑adjustment execution efficiency, and post‑launch ongoing marketing capabilities will determine whether this authorization becomes a long-term cornerstone of Southern Fund’s cross‑border product lineup or merely another limited‑scale thematic ETF.
Overall, Southern Fund has secured a significant advantage in this authorization deal, but the outcome will hinge on three key factors: whether it exercises sufficient patience in choosing the right issuance window; whether its product positioning can withstand scrutiny from industry trends; and whether its operational capabilities can sustain the long-term management of a cross-market thematic ETF.
The Hong Kong Exchange’s index construction provides a foundation for the product, but whether the product can move from “being authorized” to “being genuinely adopted by investors” depends not on the licensing agreement, but on the execution pace that Southern Fund will follow going forward.
For more exciting content, follow TMTPost on WeChat (ID: taimeiti), or download the TMTPost App.