China's Structural Bull Market Case Still Holds: From Valuation Repair to the AI Engine

Macro Confirmation: Growth Stabilises, Deflation Fades, Policy Support Holds

At the macro level, what matters isn’t month-to-month noise in the data — it’s confirmation of the underlying structure of the economy. 1Q GDP grew 5.0% year-on-year, confirming that growth remains resilient. More significant still, the Producer Price Index (PPI) turned positive for the first time in March and continued rising in April, suggesting the deflationary backdrop that has weighed on corporate nominal earnings for years is starting to lift. For blue-chip-heavy broad indices, prices are shifting from a drag to a tailwind — a key link in the earnings recovery chain.

The composition of manufacturing strength is worth unpacking too. High-tech manufacturing value-added grew 15.1% year-on-year in May, with chip output up 22.9% and industrial robot output up 27.9%; equipment manufacturing accounted for nearly 80% of industrial growth. Supply-side momentum is running well ahead of aggregate demand. This pattern confirms that economic momentum is shifting from traditional sectors toward higher-value-added, R&D-intensive segments — a direct reflection of China’s “new quality productive forces” policy push, and a bottom-up support for tech-sector earnings growth. A flat headline number shouldn’t be read as a weak recovery; the underlying structure tells a more meaningful story than any single figure — one of resilient growth, fading deflation and continued structural upgrading.

On monetary policy, the central bank has moved into a data-watching window and pulled back from explicit RRR- or rate-cut guidance, but the easing bias hasn’t changed. The constraint is clear: commercial banks’ net interest margin has fallen to a record low of 1.40%, which is the binding limit. Going forward, policy is more likely to move asymmetrically — guiding down medium- and long-end rates such as the 5-year-plus Loan Prime Rate (LPR) to lower financing and mortgage costs, while protecting bank margins and profitability as much as possible. Our base case remains one asymmetric rate cut in the second half. It’s worth clarifying that the tighter stance among major overseas economies stems from imported (energy) cost-driven inflation rather than overheating domestic demand. China’s room to ease is determined by its own fundamentals and isn’t meaningfully constrained by the external monetary cycle. That raises our confidence that policy support will hold.

On the external front, the tail risk of an extreme escalation in tariffs has narrowed following the first China-US meeting under the current administration in May. A newly established bilateral trade and investment mechanism has institutionalised the path toward détente; even if some measures are later folded into Section 301 actions, this framework should still soften the marginal impact of policy uncertainty. Combined with China’s increasingly diversified export destinations and reduced reliance on any single market, the external variables that weighed most heavily on China equity risk appetite over the past few years have narrowed considerably — freeing the market to refocus its pricing on fundamentals and earnings themselves.

An Earnings-Driven Re-Rating

This rally in Chinese equities is being driven primarily by earnings delivery, not simply by multiple expansion. A-share forward P/E ratios remain within a reasonable historical range, and as earnings come through, further upside opens up. This is the shared foundation for the two technology allocation ideas we lay out below.

Supply-side strength has already been confirmed by the data: as noted, high-tech manufacturing value-added grew 15.1% year-on-year in May, and equipment manufacturing contributed nearly 80% of industrial growth, with supply running well ahead of aggregate demand. At the same time, AI demand is converting into earnings faster. Global AI capex remains elevated — North America’s leading hyperscalers are running combined 2026 capex of roughly USD 725 billion, flowing through the supply chain to downstream players. The memory price index is up more than 70% this year, with the supply gap expected to persist into 2027, lifting both volumes and prices — and directly boosting sector earnings — upstream.

Demand-side data tells the same story: weekly token calls for domestic large language models have at times overtaken those in the US, surpassing 4.6 trillion calls a week, and accelerating commercialisation is giving the sector clear earnings visibility. The investor base is also improving — insurance funds, wealth management products and social security funds are all steady long-term buyers of A-shares, and combined with the continuing easing bias, this is improving the composition of the A-share investor base and lifting the market’s valuation floor.

In terms of implementation, leadership rotates rapidly between AI sub-sectors, making timing costly. Capturing earnings-driven sector beta systematically through a high-purity, broad-based technology vehicle is a more robust approach than concentrating in a single stock or sub-sector. That leads to the two allocation ideas in this note: an onshore broad-based technology vehicle centred on domestic semiconductors and AI hardware, and an offshore technology leader centred on platform-economy earnings recovery.

AI Policy and Domestic Substitution: The GF China Securities All-Share Information Technology ETF

The first layer of support for the onshore technology theme comes from policy. AI+ was written into the government work report as a standalone priority for the first time, and AI agents are now mandated for accelerated deployment. Information technology carries the dual designation of both an “emerging pillar industry” and a “future industry,” with policy support running through the entire 15th Five-Year Plan from 2026 to 2030. On the fiscal side, a 4% deficit ratio, special treasury bonds directed toward technology innovation, and new “compute-power coordination” infrastructure all provide a firm backdrop for compute buildout and domestic substitution.

More significant is the breakthrough at the systems level. Facing external constraints on advanced process nodes, domestic manufacturers have replaced geometric scaling with time-domain scaling — using logic folding and system-level interconnects to let mature process technology support high-end compute demand. In optical interconnects specifically, China already leads globally. This means domestic compute self-sufficiency isn’t necessarily hostage to any single process-node bottleneck.

Export controls have, if anything, accelerated this process. The structural tightening of high-end chip supply to China has forced full-stack self-sufficiency across models, compute and equipment, with domestic penetration rates continuing to rise. Semiconductor equipment is the segment with the greatest leverage to domestic substitution: the IPOs of CXMT (DRAM) and Yangtze Memory Technologies (NAND) confirm that domestic AI memory capacity is now a reality, and are pulling through further demand for domestic equipment and materials — a dynamic that points directly to a broad-based domestic technology vehicle centred on semiconductor equipment and AI hardware.

Semiconductor equipment carries the greatest leverage to domestic substitution because it is essentially a picks-and-shovels business: regardless of which memory or logic foundry ultimately wins, the etching, thin-film and metrology equipment and materials needed for capacity expansion will all be supplied domestically. The CXMT and Yangtze Memory Technologies IPOs don’t just confirm that domestic AI memory capacity has arrived — they signal that a capex cycle led by domestic foundries is now underway, one that should pull through domestic equipment and materials demand with a high degree of certainty and durability. Compared with betting on a single chip-design company, a broad-based vehicle spanning equipment, materials, manufacturing and design can capture the overall dividend from domestic substitution while avoiding the single-stock risk that comes from a concentrated technology path or customer base — this is the core rationale for using the GF China Securities All-Share Information Technology ETF as our core onshore allocation.

This isn’t just a forward-looking thesis — first-quarter results already bear it out. Constituents of the CSI All-Share Information Technology Index posted combined revenue growth of approximately 20.7% year-on-year in 1Q26, with net profit up around 74% — profit growth nearly four times revenue growth, the strongest single-quarter performance in almost two years. Structurally, global compute spending and domestic substitution are working as twin engines: Foxconn Industrial Internet (AI server assembly) posted 1Q net profit growth of roughly 102% with strong order visibility, while Hygon Information Technology (domestic CPU/GPU) has logged profit growth of 20% to 75% for several consecutive quarters, driven by the rigid substitution demand that export controls have created. Both engines are firing in the same direction, and that is earnings delivery that is already happening.

China’s competitiveness in the AI value chain isn’t limited to a single link — it spans the full chain from underlying energy supply to top-layer applications. The push toward full-stack self-sufficiency is materially improving earnings visibility for companies across that chain. The table below maps where Chinese companies stand across the five layers of the AI stack:

Layer	Core Rationale
Application Layer · AI Monetisation Outlet	Domestic AI is commercialising faster through super-app ecosystems. Doubao, Qwen and similar models are deeply embedded in content and e-commerce use cases, and continue to expand into robotics and autonomous driving, with monetisation steadily strengthening.
Model Layer · Foundation and Inference	Domestic frontier models have entered the top global tier, with lower training costs and faster commercial deployment, and the capability gap with leading overseas players continues to narrow.
Infrastructure Layer · Data Centres and Compute	Cloud and data centre capacity is expanding rapidly, supporting both training and fast-growing inference compute demand.
Chip Layer · Silicon and Accelerators	Domestic chips accounted for roughly 41% of chip sales within China in 2025, with a clear and accelerating trajectory toward self-sufficiency; export controls have, if anything, sped up this process.
Energy Layer · Grid and Foundation	Power supply is abundant and cost-competitive, leaving AI-related energy constraints significantly lower than overseas. China’s 2030 energy system roadmap — raising non-fossil power’s share to around 50% and substantially expanding new-style energy storage and renewable hydrogen — provides policy-backed demand support for the long-term expansion of compute capacity and the surrounding supply chain.

Source: iFinD, government work report, TrendForce; compiled by iFAST Research

Data as of June 2026

In terms of implementation, the GF China Securities All-Share Information Technology ETF (159939.SZ) offers close to 100% sector purity across the full semiconductor equipment and AI hardware chain. Core holdings include NAURA Technology (semiconductor equipment), SMIC, Hygon Information Technology, Cambricon Technologies and Foxconn Industrial Internet — the core names in domestic compute and manufacturing — making it a direct vehicle for systematically capturing this earnings-driven sector beta. Based on constituent-weighted earnings forecasts for the CSI All-Share Information Technology Index (000993.SH), our base case puts the index’s 2028 target at approximately 17,224 points, implying roughly 81% of potential upside over three years from current levels.

CSI All-Share Information Technology Index	FY25	FY26E	FY27E	FY28E
PE Ratio (x)	68.0	59.0	48.0	40.0
Earnings Growth	45%	70%	23%	20%
EPS (RMB)	0.69	1.16	1.43	1.72
Base-Case Target Level (Fair PE of 55x)				17,224
Three-Year Potential Upside				+81%

Source: iFinD; iFAST Research estimates

Data as of 30 July 2026. E denotes estimates; targets are based on fair-PE assumptions under the base-case scenario.

On valuation dynamics, as earnings deliver quickly through the forecast period, the index’s forward PE falls from roughly 59x in FY26E to around 40x in FY28E; under a 55x fair-PE assumption, that puts the 2028 target at approximately 17,224 points, implying roughly 81% of potential upside over three years from current levels. More importantly, the returns behind the domestic semiconductor and AI hardware theme represented by GF’s ETF are driven by structural earnings growth that is delivering quarter by quarter — orders are already booked, and quarterly results keep confirming it. This represents the hardware-delivery layer of the AI opportunity that is already happening.

Consumption’s Soft Spot and the Offshore Tech Recovery: The Hang Seng TECH ETF

In contrast to the structural strength in onshore technology, consumption is the clearest soft spot in China’s economy today — and the key to understanding the offshore technology sector. Total retail sales of consumer goods fell 0.6% year-on-year in May 2026 to RMB 4.11 trillion, the first monthly year-on-year decline since 2023 and a marked deceleration from April. Big-ticket spending was hit hardest: autos fell 16.1%, appliances 15.6%, building materials 13.6%, furniture 8.7%, and gold and jewellery 8.9%, largely on the back of fading subsidies and high base effects. Necessities, services and online consumption held up relatively well — catering grew 0.6%, cosmetics 2.5%, apparel 3.8%, and services retail grew 5.4% for January–May. The structural drag stems mainly from the property market’s erosion of household wealth: the consumer confidence index fell to 89.0 in April, its lowest since the second half of last year, with households showing an elevated preference for saving.

This weakness feeds through to offshore technology via a clear chain: consumption → platforms → offshore technology. Softer domestic demand directly weighs on the platform economy, with e-commerce GMV, local services and online advertising growth all slowing together. Combined with the earlier cutthroat price competition between platforms eroding corporate profits, offshore Chinese technology earnings and sentiment have stayed under pressure and traded weak. This is the direct reason Hong Kong-listed tech has clearly lagged Seoul and Tokyo.

Both drags on Hang Seng TECH are fixable, temporary factors. The first is the food-delivery price war: Meituan, Alibaba and JD.com have been fighting for market share with subsidies, hitting profits directly — Meituan swung from profit to loss, Alibaba’s net profit fell around 95% year-on-year, and JD.com’s fell about 39%. But all three have publicly signalled they are pulling back on subsidies, and losses in this business have narrowed noticeably from late 2025 through the first quarter. The second is a year-on-year pullback in consumer electronics and new-energy vehicles against 2025’s high subsidy-driven base — the businesses themselves haven’t deteriorated, the comparison base has simply risen. Meanwhile, core business profitability hasn’t weakened: excluding AI spending, Tencent’s core operating margin is around 43%, and Alibaba Cloud’s revenue grew roughly 40% year-on-year, with AI-related revenue up around tenfold within six months. The market has already started to confirm the turning point — the Hang Seng TECH Index rose as much as approximately 4.72% in a single day in early June 2026, with Tencent up more than 10% and Meituan and Alibaba up 6% to 9%.

We believe the pressures above are already largely priced in, leaving limited further downside. Across both earnings and valuation, the conditions for a medium-term recovery are building. First, valuations sit near historical lows, and a stabilising offshore yuan (CNH) should help both the translation of offshore earnings and foreign investor sentiment. Second, the platform price war is easing, unit economics are improving, and platform losses are narrowing, turning the earnings trajectory upward. Third, the 8 June quarterly index review added frontier large-model names such as MiniMax and Zhipu AI, tilting the Hang Seng TECH Index’s composition further toward the AI theme. Fourth, if consumption-boosting and anti-“involution” policies are stepped up further in the second half, a pickup in platform GMV, advertising and monetisation would provide additional upside. The market broadly expects 2Q GDP growth of around 4.5%, and the second half is the main window to watch for this recovery.

In terms of implementation, the Hang Seng TECH ETF (3067.HK) covers leading internet platforms as well as the newly added frontier large-model names, making it a direct vehicle for capturing this offshore recovery theme. Based on constituent-weighted earnings forecasts for the Hang Seng TECH Index, our base case puts the index’s 2028 target at approximately 7,706 points, implying roughly 60% of potential upside over three years from current levels.

Hang Seng TECH Index	FY25	FY26E	FY27E	FY28E
PE Ratio (x)	18.3	17.8	15.4	13.4
Earnings Growth	2.8%	2.6%	15.8%	14.6%
EPS	251.54	258.08	298.86	342.49
Base-Case Target Level (Fair PE of 22.5x)				7,706
Three-Year Potential Upside				+60%

Source: Bloomberg Finance L.P.; iFAST Research estimates

Data as of 30 July 2026. E denotes estimates; targets are based on fair-PE assumptions under the base-case scenario.

The index currently trades at a forward PE of roughly 13x to 18x; under a 22.5x fair-PE assumption, that implies a 2028 target of approximately 7,706 points, or roughly 60% of potential upside over three years from current levels. We’d flag one important caveat: this larger upside reflects the index having been heavily sold off previously, rather than any deterioration in its constituents’ fundamentals. It depends on both a re-rating back toward fair value and confirmation of an earnings turning point centred on the consumption and platform recovery — a slower-than-expected pace on either front would pressure this higher-beta position. Unlike GF’s ETF, where the hardware earnings are already happening, Hang Seng TECH is capturing a recovery that has not yet happened: platform fundamentals are already better than recent results suggest, headwinds in businesses like food delivery are gradually fading, and the AI resources being invested today should convert into revenue and profit over the next two to three years.

Taken together, GF’s ETF and Hang Seng TECH are not substitutes for one another — they are two sides of the same AI opportunity. GF captures hardware earnings that are already happening; Hang Seng TECH captures a platform recovery that has not yet happened. Without either one, the picture of China’s AI opportunity is incomplete.

Recommended Names and Positioning Framework

Based on the analysis above, we recommend holding the GF China Securities All-Share Information Technology ETF (159939.SZ) and the Hang Seng TECH ETF (3067.HK) in equal weight. The former captures hardware earnings that are already happening — orders are booked, and quarterly results keep confirming it. The latter captures an offshore platform recovery that has not yet happened — platform fundamentals are already better than recent results suggest, headwinds in businesses like food delivery are fading, and current AI investment should convert into revenue and profit over the next two to three years. Without either one, the picture of China’s AI opportunity is incomplete.

For investors seeking broader exposure across the full AI value chain — extending from information technology to the physical infrastructure that underpins AI at scale — the T. Rowe Price Funds SICAV – China Evolution Equity Fund offers a complementary perspective worth considering. Large-model training is extremely energy-intensive: data centres need not just chips and servers but reliable, high-capacity power supply, and electricity is increasingly seen as a hard constraint on AI expansion globally. That constraint creates investment opportunities in power equipment, grid infrastructure and industrial machinery — segments that sit upstream of the information technology sector and aren’t captured by a pure information-technology index.

The T. Rowe Price fund is deliberately positioned in this layer: industrials and commercial services make up 30.2% of its portfolio, more than three times its 9.5% benchmark weight, and holdings such as Weichai Power reflect how the AI opportunity extends from semiconductors and software into the broader industrial and energy system. Investors can allocate across the vehicles below based on their own risk tolerance and market views:

Recommended Names	Ticker	Market / Domicile	Benchmark Index	Role in Portfolio
GF China Securities All-Share Information Technology ETF	159939.SZ	Shenzhen Stock Exchange	CSI All-Share Information Technology Index	Hardware earnings realisation
Hang Seng TECH ETF	3067.HK	Hong Kong Stock Exchange	Hang Seng TECH Index	Platform monetisation
T. Rowe Price Funds SICAV – China Evolution Equity Fund	LU2187417386	Luxembourg (UCITS)	MSCI China All Shares Index	Physical infrastructure

Source: Public disclosures from fund managers and index providers; compiled by iFAST Research

Taken together, the GF China Securities All-Share Information Technology ETF (159939), the Hang Seng TECH ETF (3067) and the T. Rowe Price fund correspond to three distinct but mutually reinforcing layers of China’s AI opportunity: hardware earnings realisation, platform monetisation, and physical infrastructure. Investors looking to express this theme more fully could consider holding all three; the relative weighting of each should reflect individual risk tolerance, investment horizon and preference for active management.

Risk Factors

•        Consumption and platform recovery falling short of expectations: The valuation recovery in offshore technology depends heavily on an earnings turning point in consumption and the platform economy. Retail sales turned negative year-on-year in May for the first time since 2023. If consumption-boosting and anti-“involution” policies transmit more slowly than expected, or the food-delivery price war eases more slowly than expected, the recovery in platform GMV, advertising and monetisation could be delayed, putting pressure on Hang Seng TECH’s earnings recovery path and valuation re-rating.

•        AI commercialisation and technology iteration falling short of expectations: Earnings forecasts for the onshore technology theme assume continued delivery of AI application commercialisation and steadily rising penetration of domestic compute and semiconductor equipment. If AI model iteration or commercialisation slows, or the pace of domestic substitution stalls, earnings forecasts for the CSI All-Share Information Technology Index could face downward revisions, pressuring the valuation floor.

•        Export controls and supply-chain risk: Further expansion of US export controls on advanced chips and semiconductor equipment to China could disrupt the pace of domestic substitution and earnings forecasts for related constituents in the near term. GF’s ETF target also assumes continued hardware order flow; a pullback in capex by major customers would change this picture.

•        Regulatory and listing constraints: Alibaba (9988.HK) and Baidu (9888.HK) were added to the US Department of Defense’s list of “Chinese Military Companies” in June 2026, and US federal government procurement contracts will be barred from sourcing from these companies via third parties starting in 2027. Both are significant constituents of the Hang Seng TECH Index, and investors should be aware of these medium-to-long-term constraints.

•        A persistent offshore discount and high sector volatility: The valuation discount of offshore Chinese technology relative to onshore stocks stems from geopolitical and liquidity factors, and could persist for an extended period rather than narrowing as our base case assumes. In addition, volatility in this sector has risen significantly after its sharp rally, and the concentrated sector exposure means periodic drawdowns could far exceed those of broad-based indices; all the vehicles discussed in this note are suited to investors with a higher risk tolerance.

•        Active management and sector concentration risk: The T. Rowe Price fund is an actively managed product, and its performance depends on the manager’s stock selection and allocation decisions, which may diverge from the benchmark. Its significant overweight to industrials and energy fits the AI physical-infrastructure theme, but these sectors are more cyclical, and a slower-than-expected capex cycle or pace of power demand could weigh on fund performance.

•        Currency risk: The vehicles discussed in this note are denominated in RMB, HKD and foreign currencies respectively. For investors settling in Singapore dollars or other home currencies, movements in the RMB and HKD against that currency will directly affect actual returns.

•        Taiwan Strait geopolitical tail risk: Potential volatility in cross-strait relations is a hard-to-predict tail risk, is not part of our base case, and should not be read as a base-case forecast. That said, if geopolitical tensions escalate beyond expectations, already-cautious institutional foreign inflows could come under periodic pressure. Investors should factor this in as a low-probability, high-impact scenario.

Disclaimer

Important Notice

This report was prepared by the iFAST China Research Team and is intended for general informational reference only; it is published openly for general investors. This report does not constitute, and should not be construed as, an offer, solicitation, or investment advice to buy or sell any security, fund, or investment product, and does not take into account the investment objectives, financial situation, or specific needs of any particular recipient.

The information, opinions, forecasts, and valuations contained in this report are based on public information that our team believes to be reliable, together with our independent judgement; however, the Company makes no express or implied warranty as to their accuracy, completeness, or timeliness. Forward-looking statements in this report are subject to uncertainty, and actual outcomes may differ materially from them.

Past performance is not indicative of future performance. The value of investments and the income derived from them may fall as well as rise, and investors may not recover their original investment principal; where an investment is denominated in a foreign currency, exchange rate movements may also adversely affect its value.

Before making any investment decision, investors should independently and carefully assess the relevant risks, and where necessary, seek independent professional financial, legal, or tax advice tailored to their individual circumstances.

The Company and relevant licensed entities within the iFAST Group (including iFAST Platform Limited (IFPL), among others) are regulated by the relevant regulatory authorities in their respective jurisdictions. For further information about the Company and this report, please visit the official iFAST website.

Analyst Disclosure

As at the date of publication of this report, the author, Ian Li, CFA, and his immediate family hold NIL positions or interests in the securities or related investment products discussed in this report; the author’s compensation is not directly linked to the specific views expressed in this report.

AI Usage Disclosure

AI-assisted tools may have been used in the drafting, data compilation, and chart-preparation stages of this report; all related outputs have been independently reviewed and verified by our research team.