China has two AI trades right now. Here is why both are worth owning.

There are two ways to invest in China’s technology sector.

The first is through A-share hardware companies — the businesses that manufacture the chips, servers and components that AI runs on. These are listed on China’s mainland stock exchanges and captured by the CSI All-Share Information Technology ETF (SZSE:159939).

The second is through Hong Kong-listed platform companies — businesses like Tencent, Alibaba and Meituan that billions of people use every day, and that are investing heavily in AI to grow their services. These are captured by the iShares Hang Seng TECH ETF (HKEX:3067).

This year, one is up 47%. The other is down 17%. Most investors look at that and assume one is a good investment and the other is not.

That is not what is happening. The two ETFs have moved in different directions for completely different reasons. And right now, both are worth owning.

Here is why.

The world is in the middle of a historic spending boom on AI infrastructure. Google, Amazon, Microsoft and Meta are expected to spend around USD 725 billion building data centres in 2026 alone — 77% more than last year. Someone has to make the components that go into all of that. A large portion of those orders are landing with Chinese hardware companies. Orders have arrived. Profits have followed.

Those are the companies inside 159939. That is why it is up 47%.

3067 is a different story. Its biggest holdings — Tencent, Alibaba and Meituan — are not selling to data centres. They are consumer and platform businesses. Earlier this year, Meituan, Alibaba and JD.com were all offering heavy discounts to steal food delivery customers from each other. That cost them a lot of money and their profits collapsed. On top of that, all three are spending heavily on AI development right now, which is an additional drag on near-term earnings.

That is why 3067 is down 17%.

Most investors still associate “China tech” with large platform companies such as Tencent, Alibaba and Meituan. But the companies doing the physical work of building China’s AI infrastructure are mostly listed on mainland exchanges, not in Hong Kong. This is where 159939 offers a very different exposure.

159939 is highly concentrated in the “hard” end of the technology value chain. Around 85% of the portfolio is invested in businesses that make the physical components behind AI infrastructure — chips, circuit boards, servers, optical components and related hardware. By sub-sector, around 35% of the fund is exposed to semiconductors and equipment, 31% to electronic equipment and components, 22% to technology hardware, and only 12% to software and IT services.

Its holdings span three layers of the AI supply chain. At the upstream computing layer, companies such as Cambricon and Hygon Information benefit from two tailwinds: growing demand for chips that run AI models locally, and Chinese companies being required to buy domestic chips instead of imported ones as US export restrictions tighten. At the midstream infrastructure layer, companies such as Foxconn Industrial Internet and Victory Giant Technology directly receive data centre capex orders. At the downstream software layer, companies such as iFLYTEK and Yonyou Network stand to benefit gradually as AI-native applications become more widely adopted and monetised.

In simple terms, Hygon Information makes domestic chips to replace imported ones, Cambricon designs AI processors, Foxconn Industrial Internet assembles AI servers, and Victory Giant Technology supplies high-end printed circuit boards used in AI servers. This makes 159939 one of the more direct ways to invest in the production side of China’s AI buildout.

3067 is mostly made up of platform and consumer businesses. Tencent and Alibaba are the two larger holdings. Meituan — China’s dominant food delivery app — is another major position. So is BYD, the electric vehicle maker.

These companies serve hundreds of millions of Chinese consumers every day. Their AI story is different from 159939 — they are not benefiting from infrastructure spending today, but they are investing in AI now so that their businesses become more valuable over the next few years.

We do not need to speculate about whether 159939’s companies are benefiting from AI. The first-quarter 2026 results are confirmed.

All 158 companies in the index reported. Their combined revenue grew 20.7% compared to a year ago. Their combined net profit grew 74.1%. Profits grew nearly four times as fast as revenue — the strongest quarter in nearly two years.

Two things are driving this.

The first is global AI spending. Foxconn Industrial Internet assembles AI servers and its profit grew 102% year-on-year in the first quarter of 2026. Its orders are contracted and visible well into the future.

The second is China replacing foreign chips with domestic ones. The US has restricted the sale of advanced chips to China. The unintended consequence is that Chinese companies now have to buy domestic chips instead — guaranteed demand for companies like Hygon Information. Hygon has grown its profits between 20% and 75% every quarter for several consecutive quarters.

These two engines reinforce each other. Global AI spending keeps hardware orders coming. US restrictions keep domestic chip demand high. Both are growing at the same time.

3067’s difficult year comes down to two things, both of which are temporary.

The first is the food delivery price war. Meituan, Alibaba and JD.com were all heavily subsidising customers to take business from each other. The cost showed up directly in earnings: Meituan went from profit to loss, Alibaba’s profit fell 95% year-on-year, and JD.com’s fell 39%.

The second is a headwind faced by consumer technology and electric vehicle segments. In 2025, the Chinese government offered generous incentives — discounts on electric vehicles, subsidies on consumer electronics, trade-in schemes — that drove an unusually strong year of sales. Companies naturally reported strong earnings that year. When those incentives ended entering 2026, sales came off their peak and earnings comparisons turned unfavourable. The businesses have not gotten worse. The bar they are being compared against got higher.

Now look at what is happening underneath those numbers.

Tencent’s core business is operating at a 43% profit margin — once you strip out what it is spending on AI development. Alibaba Cloud grew revenue 40% year-on-year and its AI-specific revenue grew tenfold in just six months. These are not struggling companies. They are companies going through a short-term earnings dip while building something that will generate returns for years.

And the food delivery war is ending. All three platforms have publicly signalled they are pulling back on subsidies. Their losses in that segment narrowed sharply from late 2025 into the first quarter of 2026.

The market noticed. In early June 2026, the Hang Seng TECH Index jumped 4.72% in a single day. Tencent rose more than 10%. Meituan and Alibaba gained between 6% and 9%. That is investors concluding that the worst is behind them.

Based on our earnings forecasts, 159939 has a target index level of 17,224 points — around 37% above where it trades today. That assumes the hardware order cycle continues and earnings keep coming in as expected.

For 3067, our target is 7,706 points — around 67% above current levels. The larger potential gain reflects how deeply the index has been sold off. The businesses inside it have not become less valuable. They have had a difficult few months for specific, temporary reasons. Once those reasons fade, the gap between price and value should close.

Table 1: Hang Seng Tech Index earnings forecast

Table 2: CSI All-Share IT Index earnings forecast

If China’s macro recovery falls short of expectations, weak consumer could affect demand for electric vehicles and consumer electronics, reducing the earnings upside for related technology sectors.

The US could tighten technology restrictions further, which would hurt hardware supply chains inside 159939.

For 3067, the central risk is timing. The food delivery price war between Meituan, Alibaba and JD.com is showing signs of ending — but if competition persists longer than expected, the earnings recovery that underpins the 67% upside target will take longer to materialise.

Singapore-based investors also face currency risk — movements in the Chinese yuan and Hong Kong dollar against the Singapore dollar affect your actual returns. And 159939’s target assumes hardware orders keep arriving — if major customers pull back, that changes the picture.

159939 gives you access to AI earnings that are already happening. The orders are in and the quarterly results confirm it.

3067 gives you access to the recovery that has not happened yet. The platforms are in better shape than their recent results suggest, the food delivery headwinds are fading, and the AI investments they are making now will show up as revenue and profit over the next two to three years.

One without the other is an incomplete picture of China’s AI opportunity. We suggest holding both in equal measure.

For investors who want to go one step further and capture the full AI value chain — beyond information technology into the physical infrastructure that makes AI possible at scale — T. Rowe Price Funds SICAV – China Evolution Equity Fund offers a complementary angle worth considering.

Training large AI models is enormously energy-intensive. Data centres require not only chips and servers, but reliable, high-capacity power supply — a bottleneck increasingly recognised as a binding constraint on AI expansion globally. This creates investment opportunities across power equipment, grid infrastructure, and industrial machinery that sit upstream of the IT sector, and are not captured by a pure information technology index.

With 30.2% of its portfolio in Industrials and Business Services — more than three times its benchmark weight of 9.5% — the T. Rowe Price China Evolution Equity Fund is deliberately positioned to capture this layer. Holdings such as Weichai Power illustrate how the AI opportunity extends beyond semiconductors and software into the broader industrial and energy stack.

Taken together, 159939, 3067, and the T. Rowe Price China Evolution Equity Fund represent three distinct but mutually reinforcing layers of China's AI opportunity: hardware delivery, ecosystem monetisation, and physical infrastructure. Investors seeking the most complete expression of this theme may find value in holding all three.

Declaration

For specific disclosure, at the time of publication of this report, IFPL, through its connected and associated entities, and the analyst who produced this report hold a NIL position in the abovementioned securities.

This research report was prepared with the assistance of artificial intelligence tools. iFAST Financial Pte Ltd does not rely exclusively on AI for content generation. The content of this report, including all investment theses, ratings, price targets and conclusions, has been independently reviewed and verified by the research analyst(s) to ensure accuracy and professional integrity.

Investors should note that ETF investments are subject to market risk, sector concentration risk, currency risk and valuation risk. Past performance is not indicative of future performance. Investors should consider whether the product is suitable for their investment objectives, financial situation and risk tolerance.