Hong Kong Exchanges and Clearing has launched the HKEX AI Index, its first index built around the artificial intelligence value chain. According to the exchange's announcement, reported by GlobeNewswire, the HKEX AI Index gives investors a single, clearly defined benchmark for allocating to AI across chips, infrastructure, models and applications. The launch also broadens Hong Kong's index suite at a time when AI spending, model releases and enterprise adoption are reshaping how markets value technology companies.

The HKEX AI Index arrives with a practical next step already attached. According to the announcement, HKEX has signed an agreement with China Southern Asset Management authorising the firm to launch an exchange-traded fund tracking the HKEX AI Index in the Chinese mainland market. That link matters because it turns a benchmark into an investable product, giving mainland investors a straightforward route into a basket of AI-related stocks rather than requiring them to pick individual winners in a fast-moving sector.

What the HKEX AI Index covers

The HKEX AI Index captures the AI value chain through four segments, according to the announcement. The chip segment supplies the computing hardware that underpins AI, while the infrastructure segment directly captures AI capital expenditure on data centres, networks and related capacity. The model segment covers the large language models and other AI systems increasingly used in enterprise deployment, and the applications segment covers the commercial products built on top of those models. Together, the four segments are meant to track how value flows from silicon to software.

The HKEX AI Index holds 40 constituents, according to the announcement. The list mixes established Hong Kong-listed technology giants with newer AI companies. It includes SMIC, Alibaba Group and Tencent Holdings alongside next-generation AI companies such as Biren Technology, Zhipu AI, MiniMax and Horizon Robotics. That mix is deliberate: the HKEX AI Index pairs the scale and liquidity of incumbent platform companies with the growth potential of specialist model builders and chip designers that are still early in their commercial journeys.

Domestic open-source large language models are becoming a preferred option for enterprise deployment, according to the announcement, and the commercial pathways for AI applications are becoming increasingly clear. That framing explains why the HKEX AI Index spans the full chain rather than focusing on a single layer. An investor who buys only chip stocks misses the model and application companies monetising those chips, while an investor who buys only application stocks misses the infrastructure spending that makes those applications possible. The HKEX AI Index is designed to hold both sides of that equation in one benchmark.

How the HKEX AI Index chooses its constituents

Eligibility for the HKEX AI Index is tied to investment and growth, not size alone. According to the announcement, eligible constituents must meet thresholds for research and development expenditure and capital expenditure, each representing no less than 3 percent of sales, alongside sales growth of no less than 5 percent over the past two years. Those filters screen for companies that are genuinely reinvesting in AI capability rather than simply describing themselves as AI companies. A firm that spends little on research or capital projects will struggle to qualify, however fashionable its branding may be.

Weighting rules add a second layer of discipline to the HKEX AI Index. No individual constituent may exceed a 10 percent weight, according to the announcement, while each stock in the applications segment is capped at 3 percent. The caps prevent a single giant from dominating the benchmark and prevent the more speculative applications layer from overwhelming the steadier chip and infrastructure segments. The HKEX AI Index is reviewed and rebalanced quarterly, so constituents that stop meeting the investment and growth thresholds can be removed and rising companies can be added as the AI landscape shifts.

The design echoes a broader shift in how enterprises treat AI agents and AI systems as infrastructure rather than experiments. Agent identity platforms are now handling millions of daily calls, as reported in coverage of Mintive's agent identity growth, while agentic platforms are moving into regulated financial operations, as reported in coverage of Casap's agentic AI funding. An index that weights research spending and capital expenditure is, in effect, a bet that the companies building and running that infrastructure will capture a durable share of AI value.

Why the mainland ETF matters for the HKEX AI Index

The planned mainland ETF tracking the HKEX AI Index is the most consequential part of the launch for many investors. Index products lower the barrier to diversified exposure: instead of researching 40 individual companies across chips, infrastructure, models and applications, an investor can buy one fund that tracks the whole basket. For mainland investors, a fund authorised through China Southern Asset Management offers a regulated channel into Hong Kong-listed AI names, including companies that are difficult to access directly from the mainland market.

The ETF also deepens Hong Kong's role as a bridge between mainland capital and global technology trends. HKEX has described the launch as another step in deepening its index business and serving investors' diversified allocation needs, according to the announcement. A dedicated AI benchmark gives asset managers a reference point for new products, gives analysts a yardstick for comparing AI portfolios, and gives companies a visible incentive to keep investing in research and capital expenditure if they want to stay in the index. Quarterly rebalancing keeps that incentive current.

There are still risks for anyone following the HKEX AI Index closely. AI valuations can move quickly on model releases, chip supply news and policy changes, and a 40-stock basket will not smooth out sector-wide swings. The 3 percent applications cap limits single-stock concentration in the riskiest layer, but it cannot remove the volatility inherent in an emerging technology theme. Investors tracking the HKEX AI Index should therefore treat it as a thematic benchmark to be weighed alongside broader market indices, not as a substitute for diversification across sectors and regions.

For now, the HKEX AI Index gives the market something it previously lacked: a Hong Kong benchmark that follows AI value from chips through infrastructure and models to applications, with rules that reward sustained investment and growth. With a mainland ETF planned to track it, the HKEX AI Index is set to move from an exchange announcement into portfolios. How it performs will depend on whether its 40 constituents can turn heavy research and capital spending into lasting revenue, but the structure of the index makes that question easy to track, quarter by quarter. Sources: GlobeNewswire announcement and syndicated coverage of the HKEX announcement.