Tap into Asia’s semiconductor growth through the Global X Asia Semiconductor ETF (HKEX:3119)

• Asia sits at the centre of global AI growth, driven by its irreplaceable manufacturing capabilities and leadership in high-bandwidth memory production.

• The Global X Asia Semiconductor ETF (HKEX: 3119) provides a single-ticket solution to access Asia’s AI ecosystem, capturing growth across Taiwan, South Korea, Japan, and China.

• The top ten constituents include industry leaders such as TSMC, SK Hynix, and Samsung Electronics, alongside key testing equipment producers like Advantest and China-based AI chip developer Cambricon. These companies have delivered strong 1Q26 earnings, with many reporting high double-digit to triple-digit year-on-year growth.

• The ETF is projected to reach approximately HKD 236 by FY2028, implying around 68% upside from current levels.

Just a few months ago, market concerns were centred on whether US hyperscalers would sustain their aggressive AI capex. Today, that narrative has shifted meaningfully. Following a strong set of first-quarter 2026 earnings across the semiconductor ecosystem, optimism has been re-ignited. The rebound of Intel, the better-than-expected growth delivered by Taiwan Semiconductor Manufacturing Company (TSMC), and continued system-level strength supporting ASML all reinforce the same message. The AI cycle remains firmly intact, with demand continuing to broaden rather than peak.

Why Asia sits at the centre of AI growth?

At the centre of this growth are Asian semiconductor companies, which occupy the most critical bottlenecks in the global AI infrastructure stack. Their strategic importance is underpinned by two powerful and enduring competitive advantages.

The first is their irreplaceable manufacturing leadership. TSMC remains the undisputed leader in advanced semiconductor manufacturing, accounting for approximately 74% of the global foundry market, according to Counterpoint Research. The company is already transitioning toward 2nm volume production, while continuing to scale its 3nm capacity. This reinforces its unmatched dominance at the most advanced nodes manufacturing.

Regardless of how AI evolves, whether through GPUs, CPUs, or increasingly software-optimised architectures, every chip must ultimately be manufactured. In practice, this means passing through a fabrication facility, and most often, through TSMC. This dynamic is becoming even more pronounced as hyperscalers accelerate the development of custom AI chips. Companies such as Google and Amazon are designing their own application-specific integrated circuits (ASICs) to reduce reliance on general-purpose GPUs. Google, for instance, has recently introduced its 8^th generation Tensor Processing Unit (TPU) to support both training and inference workloads, while Amazon’s AWS Trainium2 has been deployed for similar purposes at roughly half the price of a comparable NVIDIA H100.

Yet, while chip design is becoming more decentralised, manufacturing remains firmly concentrated. No matter who designs the chip, it still needs to be produced at scale. In that sense, the economics of AI remain unchanged, the “silicon toll” is ultimately paid to Asia.

The second key bottleneck lies in memory, specifically High Bandwidth Memory (HBM). AI performance is no longer driven solely by computational power, but increasingly by the speed at which data can be transferred and processed. This has made HBM a critical component of AI infrastructure. The market is dominated by SK Hynix and Samsung Electronics, both of which are leading the transition towards next-generation 16-layer HBM4 products.

Importantly, the nature of AI demand is evolving. While initial growth was driven by large-scale model training, the focus is now shifting towards real-time inference and agentic AI systems. This transition is broadening the demand base for Dynamic Random Access Memory (DRAM) and NAND Flash. With supply remaining tight, pricing dynamics are becoming increasingly favourable. According to TrendForce, DRAM and NAND prices are expected to rise 58% and 70% in 2Q26 in a quarter over quarter basis, reinforcing the view that the memory upcycle is structural rather than temporary.

Broad-based strength across the semiconductor value chain has been a key driver of recent market performance. Reflecting this momentum, the FactSet Asia Semiconductor Index has delivered a strong 47% year-to-date return (in SGD terms) as of 29 April 2026. This surge has also lifted tech-heavy markets, with South Korea and Taiwan overtaking the UK to become the eighth and seventh largest equity markets globally, respectively.

Global X Asia Semiconductor ETF: Your single ticket to Asia’s AI ecosystem

For investors seeking exposure to this structural trend, the Global X Asia Semiconductor ETF (HKEX: 3119) offers a comprehensive, single-ticket solution. Managed by Mirae Asset Global Investments, the ETF tracks the FactSet Asia Semiconductor Index, holding 40 leading semiconductor companies across Asia.

Table 1: The key information of the Global X Asia Semiconductor ETF

Fund Detail	Information
Full Name	Global X Asia Semiconductor ETF
Exchange / Ticker	HKEX: 3119
Benchmark	FactSet Asia Semiconductor Index
Manager	Mirae Asset Global Investments (HK)
Fund Inception	23 July 2021
Total Holdings	40 equities across 4 Asian markets
TER (Expense Ratio)	0.68% per annum
Fund NAV	HKD 533.9 million as of 29 April 2026
NAV per Share	HKD 140.5 as of 29 April 2026

One of the ETF’s most compelling features is its geographic breadth. Rather than concentrating exposure in a single market, it spans four key semiconductor hubs: South Korea, Taiwan, Japan, and China. This structure allows investors to capture different parts of the ecosystem, from advanced memory production in Korea, to leading-edge chip manufacturing in Taiwan, to the critical equipment and materials supplied by Japanese firms, and the growing domestic semiconductor ecosystem in China.

Figure 1: The ETF invests across four Asian markets

The top 10 holdings account for approximately 63% of the portfolio, with a clear concentration in the most critical enablers of AI infrastructure. Alongside industry leaders such as SK Hynix, Samsung Electronics, and TSMC, the ETF also provides exposure to Japanese companies specialising in semiconductor testing and fabrication equipment, as well as China-based AI chip developers. Together, these businesses form the “picks and shovels” of the AI gold rush, often less visible to the market, but absolutely indispensable to the entire ecosystem.

Table 2: The top 10 constituents span four key markets and cover multiple segments of the semiconductor value chain

Name of Securities	Market	Net Assets %	Role in the Semiconductor Supply Chain
SK Hynix	South Korea	11.2%	The world’s leading supplier of HBM and a key driver of next-generation HBM4. Essential for AI training, with deep integration into NVIDIA’s ecosystem.
Samsung Electronics	South Korea	9.9%	A co-leader in global HBM supply and one of the few vertically integrated semiconductor players, with capabilities spanning memory, logic, and foundry services.
TSMC	Taiwan	9.8%	The world’s leading AI chip manufacturer, controlling most of the advanced-node production and producing chips for NVIDIA, AMD, and Apple.
MediaTek	Taiwan	8.1%	A leading fabless chip designer with strong positioning in consumer AI applications, including smartphones, IoT devices, and edge computing.
Advantest Corp	Japan	5.0%	A global leader in semiconductor testing equipment, playing a critical role in the validation of increasingly complex AI chips.
Tokyo Electron	Japan	5.0%	A key supplier of semiconductor fabrication equipment, including etching and deposition tools essential for chip production.
Sony Group Corp	Japan	4.7%	A leader in image sensors that power AI vision applications, including autonomous driving, robotics, and industrial automation.
ASE Technology	Taiwan	3.3%	A major player in advanced packaging and testing, enabling integration of logic chips and HBM in chiplet-based architectures.
Suzhou Dongshan Precision	China	3.2%	A supplier of Printed Circuit Board and electronic components, supporting connectivity and hardware integration in AI systems.
Cambricon Tech	China	2.8%	A leading domestic AI chip designer in China and a key player in advancing China’s semiconductor self-sufficiency.
Source: Mirae Asset, iFAST Compilations. Data as of 24 April 2026.

 

1Q26 earnings strength reinforces the thesis

Earnings results for the first quarter of 2026 further reinforce the strength of this investment case. Among the companies that have reported, aggregate earnings grew by an impressive 157% year-on-year, representing a 24% positive earnings surprises compared to consensus estimates.

Table 3: The earnings of the top constituents for the quarter ended 31 March 2026

Name of Securities	EPS YoY growth for the Quarter Ended 31 March 2026	Key Earnings Details
SK Hynix	386%	·         Revenue reached KRW 52.6 trillion, surpassing KRW 50 trillion for the first time, with an operating margin of 72%. ·         Growth was driven by strong demand for HBM3E.
Samsung Electronics	498%	·         Revenue reached KRW 133.9 trillion and operating profit reached KRW 57.2 trillion, representing YoY growth of 69% and 756% respectively. ·         Record earnings were driven by memory business, supported by higher average selling price due to supply shortage and expanded sales of flagship System-on-Chip.
TSMC	58%	·         Revenue reached TWD 1.1 trillion, up 35.1% YoY, with operating margin expanding by 9.6 percentage points. ·         Advanced nodes (7nm and below) now account for 74% of total revenue.
MediaTek	Earnings release expected on 30 April 2026.
Advantest Corp	221%	·         Revenue reached JPY 328.1 billion, up 41.2% YoY, driven by strong demand for System-on-Chip and memory testing systems linked to AI applications. ·         FY2025 (Apr 2025-March 2026) total revenue reached JPY 1.2 trillion, 44.7% growth.
Tokyo Electron	Earnings release expected on 30 April 2026.
Sony Group Corp	Earnings release expected on 8 May 2026.
ASE Technology	88%	·         Unaudited revenue reached TWD 173.6 billion, up 17.2% YoY. The core assembly, testing, and materials (ATM) segment showed strong performance.
Suzhou Dongshan Precision	126%	·         Revenue reached CNY 13.1 billion, up 53% YoY. Optical module business revenue doubled compared to 1Q25, contributing strongly to overall growth.
Cambricon Tech	180%	·         Revenue reached CNY 2.9 billion, representing a 159.6% YoY increase. Advance payments surged by 155%, driven by newly secured chip production orders.
Source: Bloomberg Finance, L.P. Company websites. iFAST compilations. Data for the quarter ended 31 March 2026.

More importantly, forward guidance remains highly constructive. SK Hynix has indicated that demand for HBM over the next three years already exceeds available supply capacity, pointing to a sustained structural shortage. Meanwhile, TSMC has upgraded its full-year growth outlook to above 30% in US dollar terms, driven by strong demand for inference-related chips. Advantest has also highlighted accelerating capacity expansion, with plans to ramp production to 10,000 systems per year. The company expects FY2026 net income to grow 24% year-on-year, supported by rising demand for memory and system-on-chip (SoC) testing, both in terms of volume and increasing complexity as AI adoption scales. Taken together, these signals suggest that the current semiconductor cycle still has meaningful room to run.

Discounted valuation gives investors chance to gain compelling upside

Despite strong earnings and share price performance, Asian semiconductor companies still trade at a meaningful discount to their US peers. As of 27 April 2026, the FactSet Asia Semiconductor Index is trading at a forward P/E of 13.7X, which is nearly half of the 27.1X multiple implied by the MVIS US Listed Semiconductor 25 Index.

While the MVIS index does include TSMC, its composition remains heavily tilted toward US chip design and equipment leaders such as NVIDIA, AMD, and Applied Materials. Importantly, it excludes a broad set of Asian semiconductor companies that are essential to the physical production and enablement of AI hardware.

As hyperscalers increasingly shift toward custom ASIC development, the demand for general-purpose GPU designers like NVIDIA may gradually decline. However, reliance on Asian semiconductor leaders is far more entrenched. No company can realistically replicate the advanced foundry ecosystem of TSMC or the high-bandwidth memory capabilities of SK Hynix.

The moat of Asian semiconductors is fundamentally physical, deeply embedded, and extremely difficult to replicate - yet it continues to trade at a notable valuation discount.

Capturing Asia’s AI opportunity via the Global X Asia Semiconductor ETF

The AI story is not weakening; it is evolving. While applications and architectures continue to shift, the underlying infrastructure remains indispensable. Asia sits at the core of this infrastructure, providing the manufacturing and memory capabilities that enable the entire ecosystem.

For investors, the opportunity lies not just in owning the most visible names, but in gaining exposure to the full set of critical bottlenecks that underpin the AI revolution. This is where the Global X Asia Semiconductor ETF (HKEX: 3119) provides an effective solution, offering diversified access to the key enablers across the value chain.

The FactSet Asia Semiconductor Index is projected to reach approximately HKD 1,140 by FY2028, representing an implied upside of around 69% from current levels. This translates into an indicative ETF share price of approximately HKD 236.

Table 4: Earning projections of the FactSet Asia semiconductor Index

	2025A	2026E	2027E	2028E
PE Ratio	32.4	19.3	13.9	11.3
Expected Earnings Growth	44.3%	68.0%	38.3%	33.9%
EPS (HKD)	19.3	32.4	44.8	59.9
Dividend Yield	1.2%	1.3%	1.7%	2.0%
Target Price (HKD)				1,140
Upside Potential (based on fair PE ratio of 19X)				69%
Source: Bloomberg Finance L.P., iFAST Estimates. Data as of 29 April 2026.

Declaration:

This research report was prepared with the assistance of artificial intelligence (AI) 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.

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