Executive Summary

Over the past several months, some of the most explosive price appreciation on the Singapore Exchange (SGX) has come from outside the traditional domains of banking and real estate. Local semiconductor-related names—specifically AEM Holdings (SGX: AWX), UMS Integration (SGX: 558), and Frencken Group (SGX: E28)—have staged massive rallies, fueled by a powerful global artificial intelligence (AI) and semiconductor supply chain recovery. For structural income portfolios, however, this surge introduces a fundamental valuation mismatch. With local cyclical manufacturers now trading at premium growth multiples that mirror or exceed the global technology giants, investors must evaluate whether they are being adequately compensated for the localized risk.

The Vanishing Valuation Discount

Historically, Singapore technology equities traded at a notable valuation discount relative to global peers to compensate investors for their smaller scale, customer concentration, and high cyclicality. Due to recent price re-ratings, this historical safety margin has completely evaporated.

The table below outlines how current local forward valuation multiples have converged with—and in some cases surpassed—the primary global drivers of the AI revolution:

Company	Value Chain Role	Forward P/E (FY26 Est.)	Dividend Yield	Q1 Net Profit Performance
AEM Holdings	Advanced testing handlers & instrumentation	36x – 46x	0.2%	+329% YoY profit rebound
UMS Integration	Precision engineering & modular assembly	~37x	1.7% – 2.2%	+43% YoY profit expansion
Frencken Group	Specialized industrial technology services	~31x	~1.0%	Solid cyclical recovery
Nvidia (NVDA)	Global AI hardware & software ecosystem standard	~41x	Minimal	Secular hyper-growth
Micron Technology (MU)	Tier-1 High-Bandwidth Memory (HBM) oligopoly	~46x	Minimal	Secular hyper-growth
Broadcom (AVGO)	Custom AI ASICs & networking silicon IP	35x – 42x	Premium Growth	High secular visibility
Microsoft (MSFT)	Enterprise AI application & cloud software layer	25x – 35x	Core Software	High recurring revenue

Vertical Analysis: Proxy vs. Pioneer

Evaluating these assets through specific industry verticals highlights a significant pricing mismatch relative to structural economic moats.

1. The Cyclical Play: AEM Holdings vs. Micron Technology

Both AEM Holdings and Micron Technology operate at premium valuation ceilings near 46x forward earnings. However, their underlying competitive positions differ fundamentally:

• AEM operates as a secondary provider of backend semiconductor testing equipment, remaining highly exposed to capital expenditure adjustments and customer concentration.
• Micron operates as a global oligopolist controlling a critical physical constraint of the AI boom—High-Bandwidth Memory (HBM)—which is a non-negotiable component for AI data centers.

At identical valuation multiples, the risk-to-moat ratio strongly favors the primary intellectual property (IP) owner over the secondary supply-chain provider.

2. The Hardware Proxy: UMS Integration vs. Broadcom & Nvidia

UMS Integration trades at 37x forward earnings primarily as an equipment contract manufacturer. In comparison, Nvidia trades at 41x forward earnings while controlling the dominant global hardware ecosystem powering AI data centers. Similarly, Broadcom trades between 35x and 42x while owning critical, irreplaceable IP in custom AI ASICs and networking silicon. Paying equivalent multiples for localized assembly rather than global hardware monopolies represents an inefficient pricing of risk.

Key Risks

• Valuation Vulnerability & Multiple Compression: Trading at 31x to 46x forward earnings leaves these local companies with zero margin for error. The current valuation architecture assumes perfect downstream execution. Any minor delay in customer product rollouts, adjustments in capital expenditure by global chipmakers, or broader macroeconomic cooling will trigger swift multiple compression.
• Customer Concentration and Substitution: While local equipment makers benefit from immediate operational turnarounds and explosive year-over-year percentage jumps from low cyclical baselines, their order books remain highly concentrated around single major clients. Furthermore, global technology supply chains maintain high flexibility; local investors face global substitution risks if downstream giants shift capital allocation.

Strategic Synthesis

The operational recoveries experienced by Singapore’s technology sector are tangible and fundamentally driven. However, in an environment where major brokerages provide seamless, low-friction access to global equity markets, geographic proximity should not dictate portfolio construction.

Rather than forcing local cyclical manufacturers to fulfill a structural growth mandate, capital is better optimized via a clean, geographic barbell strategy:

1. Singapore for Defensive Income: Anchor the local allocation strictly to assets displaying world-class dividend durability and recurring cash flows, such as major local banks and blue-chip real estate vehicles. The structural core of the local market comfortably generates 5% to 7% tax-free dividend yields.
2. The US for Pure Growth: Target secular AI exposure directly at the heart of global innovation where companies possess dominant intellectual property, pricing power, and near-monopoly software ecosystems.

Local technology equities can continue to execute well operationally. However, when their valuation multiples converge with global pioneers, they are best viewed as tactical, high-beta satellite allocations rather than core structural portfolio holdings. Separating the income mandate from the growth mandate provides a cleaner risk-adjusted framework for long-term capital optimization.