How to ride the tech tiger in emerging markets

As the saying goes, the difficulty with riding a tiger lies not in getting on, but in getting off. After several years of strong technology-led returns, recent bouts of volatility have made investors increasingly anxious about the eventual dismount.

The pullback in AI-related hardware stocks has intensified that debate. The current investment cycle is already operating on a scale that surpasses the dotcom era. Active investors should use periods of strength to take profits and diversify. But there are also reasons to remain invested. The challenge is managing risk without losing exposure to one of the most powerful investment themes of the decade.

Putting a saddle on the tiger

Risks are undoubtedly higher than they were in 2023 or even in early 2025, as pressure mounts on ‘hyperscalers’ to monetise their enormous investment. The four largest US hyperscalers – Amazon, Microsoft, Google and Meta – are now expected to spend close to US$700 billion on capital expenditure in 2026. How quickly that spending produces an adequate return remains uncertain, particularly because the economics of AI models are still developing.

For Asian hardware suppliers, however, focusing only on the end-user economics risks missing the distinction between the companies funding the AI build-out and those supplying the equipment that enables it. Data-centre owners and AI platform providers must ultimately earn an adequate return on their capital to satisfy investors and repay their growing debt piles.

In our view, that’s where the principal risk resides. By contrast, the AI hardware companies supplying the equipment for this build-out are earning revenues today. The scale of contracted demand also matters. Long-term hardware purchase commitments across the hyperscaler ecosystem now run into the trillions of US dollars. These commitments support suppliers’ earnings visibility and help to reduce the risk around adding capacity. A note of caution: they don't eliminate the risk of overcapacity if end-demand ultimately disappoints.

Put simply, one company’s capital expenditure is another company’s revenue.

Some critics point to cyclically adjusted earnings multiples as evidence that these businesses are already expensive. Yet such measures are most useful when assessing businesses operating within an established market, with earnings that move around a relatively stable long-term trend. This semiconductor cycle combines a cyclical upswing, driven by capacity constraints. There's a structural increase in the sector’s economic importance as AI adoption broadens across consumer and enterprise applications. Valuation therefore needs to distinguish temporary peak earnings from the possibility of a permanently larger revenue pool.

The cycle is likely to move through three phases. First, exceptional profitability while capacity is constrained. Second, declining but still-elevated profitability as supply expands. Finally, margins eventually normalise closer to pre-cycle levels, albeit on a much larger revenue base. Along the way, these companies must decide how to allocate the earnings windfall among shareholders, new capacity, employees and policy-aligned investment.

We’re already seeing the opening moves in that allocation. Samsung Electronics has announced shareholder returns of roughly US$80 billion, while SK Hynix has also outlined significant capital returns. Both appear to be moving towards returning 50% of free cash flow to shareholders, alongside substantial investment in capacity and employee compensation. These announcements demonstrate both the scale of cash generation and the importance of disciplined capital allocation.

Taming the tiger

The second challenge is concentration risk. We own these three technology leaders across many of our portfolios, but our alpha has come from a broader range of companies. Exposure to the cycle need not mean relying on the same three stocks for returns.

Ownership data underline how close these holdings have come to representing the market itself. Over 90% of active emerging-market funds own TSMC, while more than 80% own Samsung Electronics and more than 70% own SK Hynix [1]. In that sense, the three tech tigers increasingly provide the market’s beta. The opportunity for active managers lies in complementing that exposure with other companies capable of delivering a differentiated return profile, while selectively taking profits to reallocate to opportunities elsewhere in the market.

The recent volatility should be viewed in context. Investors in Asia’s hardware suppliers must remain alert to concentrated returns and increasingly optimistic earnings expectations. Yet while capital expenditure guidance continues to rise and hyperscalers retain access to funding, the cycle still has support.

For active managers, the priority is to manage exposure before the market forces the dismount. For us, that means taking profits selectively, diversifying the sources of technology alpha and keeping position sizes consistent with the risks. Portfolios that remain excessively overweight tech will come face to face with the tiger sooner or later. Effective diversification will help tame the tiger.

[1] Source: Copley Fund research, August 2026

Companies selected for illustrative purposes only to demonstrate the investment management style described herein and not as an investment recommendation or indication of future performance.