NODX a macro tailwind, but can SGX-listed semiconductor stocks deliver?
April data is best read as validation of a structural thesis that Singapore has embedded itself in the AI hardware value chain
SINGAPORE’S April non-oil domestic exports (NODX) came in at over 24.5 per cent year on year (yoy), nearly two-and-a-half times the Bloomberg consensus of 10.9 per cent and the strongest reading since February 2012.
More importantly, it was the eighth consecutive month of expansion, a run that increasingly looks less like a cyclical bounce and more like a structural re-rating of Singapore’s role in the global AI supply chain.
The question for investors is not whether the data is good, but what this actually implies for semiconductor companies listed on the Singapore Exchange (SGX).
Why this NODX print is different
Past NODX surges have often been narrow; April 2026 data is broader. Electronic NODX rose 66.7 per cent yoy, led by integrated circuits, disk media products and personal computers – all of which are linked to AI infrastructure spending.
Non-electronic NODX also returned to growth territory at 10.9 per cent yoy, driven primarily by pharmaceuticals, measuring instruments and specialised machinery. The latter is a category that captures upstream semiconductor equipment demand. When both the finished electronics and the tools used to make them are expanding simultaneously, the demand signal is not being met by inventory restocking alone.
Geographic breakdown matters. NODX to the US surged 59.6 per cent yoy, partly reflecting front-loading ahead of potential tariff escalation, and a reversal of those flows is plausible in the coming months.
But NODX to China expanded 37.8 per cent on specialised machinery and integrated circuits (ICs), and South Korea grew 71.2 per cent on demand for ICs and personal computers. Neither market is subject to the same front-loading dynamic, and their expansion suggests that the underlying pull from global AI hardware assembly is genuine.
AI capex cycle is not slowing
Some investors will ask whether April’s print reflects a peak rather than a plateau. Base effects will tighten in the second half, and some US front-loading will unwind. But the underlying demand cycle shows no signs of rolling over.
Large hyperscalers such as Amazon, Alphabet, Microsoft, Meta and Oracle have collectively committed between US$660 billion and US$690 billion in 2026 capital expenditure. That is nearly double 2025 levels, with roughly 5 per cent directed towards AI infrastructure.
TSMC raised its full-year 2026 revenue growth guidance to above 30 per cent. Micron’s US$24 billion Nand wafer fabrication facility in Singapore, which represents its largest single-country investment to date, is currently under construction. In addition, a separate US$7 billion high-bandwidth memory packaging plant is expected to begin contributing to supply from 2027 onwards.
Singapore accounts for roughly 20 per cent of global semiconductor equipment production. Its position across memory, advanced packaging and equipment becomes structurally harder to bypass as AI chip complexity increases.
Three SGX names, unequal risk-reward
UMS Integration supplies precision components for front-end deposition and tech equipment to Lam Research and Applied Materials (Amat), two of the largest semiconductor equipment manufacturers in the world, both of which beat Q1 2026 revenue estimates, supporting sustained order momentum.
UMS translated that into Q1 2026 net profit of S$14 million, up 43 per cent yoy, and guided for FY2026 to exceed FY2025. Amat has historically accounted for about 80 per cent of UMS revenue, making customer concentration the key risk to monitor.
The onboarding of Lam Research is an encouraging step towards diversification. UMS also has a track record of paying quarterly dividends since May 2012; hence, it offers income alongside structural growth exposure.
AEM Holdings provides AI and high-performance computing chip-testing solutions via its proprietary PiXL thermal management technology. The results have been striking: Q1 2026 revenue rose 36 per cent yoy to S$116.9 million, with net profit surging 329 per cent, and full-year FY2026 revenue guidance was raised about 20 per cent to S$550 million to S$600 million.
A new partnership with ASE Technology Holding, the world’s largest chip testing and packaging provider, could open access to major hyperscaler customers and expand AEM’s addressable market from US$3 billion to US$4.5 billion by 2028 (with compound annual growth rate of 22.5 per cent).
The caveat is valuation: AEM’s share price has risen more than 400 per cent year to date, leaving limited room for execution shortfalls.
Frencken Group offers the most diversified exposure of the three, with its manufacturing precision mechatronics assemblies. This includes wafer manufacturing equipment, electron microscopes and medical scanners across semiconductor, medical and automotive segments.
However, unlike UMS and AEM, Frencken reported a decline in revenue in Q1 2026 (-6.4 per cent yoy). Growth may be more gradual in the near term, given weaker orders from an undisclosed major European life sciences customer. A new Singapore mechatronics facility, expected in Q1 2027, supports longer-term capacity expansion.
Investor takeaway
April’s NODX data is best read not as a catalyst but as a validation of a structural thesis: Singapore has embedded itself at multiple points in the AI hardware value chain in ways that are durable across individual customer cycles and macro volatility.
All three potential SGX-listed beneficiaries are also part of the Monetary Authority of Singapore’s Equity Market Development Programme. UMS and Frencken are included in the iEdge Singapore Next 50 Index, and AEM may be added at the next quarterly rebalancing.
UMS offers the purest leverage to semiconductor equipment capex but carries customer concentration risk. AEM’s structural case is compelling, but its year-to-date rally has compressed its valuation gap. Frencken provides the most diversified exposure via ASML’s supply chain, with some dilution from non-semiconductor segments.
The Straits Times Index’s resilience near record highs reflects the broader market’s recognition of Singapore’s structural advantages. For investors accessing the AI supply chain theme through SGX equities, UMS, AEM and Frencken are distinct points on the risk-return spectrum. Each warrants its own due diligence rather than treatment as a homogeneous basket.
The writer is a research analyst with the research and portfolio management team of FSM Global, the B2C division of iFast Financial, the Singapore subsidiary of iFast Corporation
