PETALING JAYA: The artificial intelligence (AI)-led technology stock rally may have suffered a recent bout of profit-taking, but the broader uptrend appears to remain intact as investors become more selective and focus increasingly on earnings delivery rather than lofty narratives.
According to SPI Asset Management managing partner Stephen Innes, the correction seen over the past few trading sessions does not mark the end of the AI boom but rather signals a transition into a more mature phase of the cycle.
“I do not think this is the end of the AI-led tech rally, but I do think the easy phase is over,” he told StarBiz.
Innes said investors are now shifting away from indiscriminate AI enthusiasm and placing greater emphasis on companies that can demonstrate tangible infrastructure demand, earnings visibility and pricing power.
He said the next phase of AI investing would be driven by economics rather than vision, as businesses become increasingly focused on extracting productivity gains at sustainable costs.
“The market does not need to abandon the AI story. It needs to reprice the story,” he said.
“Future winners would be those capable of generating attractive returns on compute-intensive investments rather than merely promoting ambitious AI narratives.”
Innes noted AI remains one of the world’s most compelling structural growth themes, but investors are increasingly scrutinising whether massive capital expenditure commitments can translate into durable cash flows and sustainable profitability.
“Valuations still support an investment case in parts of the AI universe, but not across the whole basket.
“At these levels, investors are no longer buying a theme; they are underwriting execution, cash flow conversion and whether massive AI capital expenditures can keep turning into durable revenue rather than just another expensive arms race,” he added.
Closer to home, iFast Capital assistant manager of research Kevin Khaw Khai Sheng views the recent weakness in technology counters as a valuation correction rather than a deterioration in underlying demand.
He attributed the sell-off to investor disappointment after semiconductor giant Broadcom Inc guided for third-quarter AI chip sales of US$16bil without upgrading its full-year AI outlook, coupled with a stronger-than-expected United States jobs report that pushed bond yields higher and weighed on technology valuations globally.
The pullback dragged Bursa Malaysia’s technology index into being the worst-performing local sector on the Bursa during the period.
“Last week’s correction was a valuation reset rather than demand destruction,” Khaw told StarBiz.
He believes the weakness presents a selective accumulation opportunity as local technology fundamentals continue to improve.
According to Khaw, several Malaysian outsourced semiconductor assembly and test companies as well as automated test equipment players are guiding for second-quarter 2026 sales growth of between 10% and 50% year-on-year, supported by AI infrastructure investments and inventory replenishment.
Unlike their US counterparts, Khaw noted Malaysian technology stocks are entering the current upcycle from a much lower valuation base after the sector ended 2025 down 12% despite the Nasdaq Composite Index reaching record highs.
He expects investors to favour companies with stronger exposure to AI-related demand and front-end semiconductor activity over those reliant on consumer electronics, particularly as the global smartphone market is projected to contract by 13% in 2026.
The longer-term investment case for Malaysia’s technology sector remains compelling, he added, given the country’s estimated 13% share of the global semiconductor packaging, assembly and testing market and its role as a supplier of approximately 20% of US semiconductor imports.
Malaysia’s neutral geopolitical positioning also places it in a favourable position to benefit from ongoing supply chain diversification efforts.
Khaw identified potential US Section 232 tariffs, higher US interest rates and a stronger ringgit as key risks to monitor, although he said he would only turn cautious if punitive tariffs were imposed without exemptions for Malaysia or if major AI hyperscalers significantly reduced capital expenditure spending.
Supporting that view, Hong Leong Investment Bank (HLIB) Research believes Malaysia’s technology sector is entering the early-to-mid stages of a broader upcycle, with first quarter of financial year 2026 results confirming improving order visibility and strengthening industry fundamentals.
It said the sector is transitioning from a valuation re-rating phase into an earnings upgrade cycle that has yet to be fully reflected in consensus forecasts.
“We view much of this re-rating as fundamentally justified,” HLIB Research.
The benchmark technology index has risen about 30% year-to-date, with individual technology counters recording gains ranging from 7% to as much as 123%.
Despite richer valuations, the research house noted semiconductor upcycles historically progress from multiple expansion to earnings upgrades, suggesting further upside could still emerge.
HLIB Research pointed to several structural growth drivers, including accelerating China Plus One supply chain relocation, deeper localisation by global semiconductor equipment manufacturers, rising optical and power semiconductor content linked to AI data centre build-outs, and improving prospects for Intel Corp’s ecosystem.
It also highlighted sustained AI infrastructure spending by major US hyperscalers and continued fundraising activity by companies such as OpenAI, Anthropic and SpaceX as supportive external factors.
Within the local market, HLIB Research’s preferred technology picks are ITMax System Bhd
, UWC Bhd, Frontken Corp Bhd and Unisem (M) Bhd.
The research house remains positive on the sector heading into the second half of financial year 2026, although it expects increased volatility as investors place greater emphasis on earnings delivery.
The overarching message from analysts is that the AI investment theme remains alive and well, but investors will increasingly need to separate companies with genuine earnings power from those relying solely on AI-driven optimism froth.
