For the past three years, artificial intelligence (AI) has dominated equity markets, driving gains primarily through chip manufacturers, cloud service providers, and data-centre operators. This focus has led to substantial capital inflows into these sectors, often at the expense of other industries. However, last week saw a significant shift as key AI-related indices, including the Nasdaq Composite and the Philadelphia Semiconductor index, along with South Korea’s chip-heavy market, dropped to three-month lows.
Investor confidence in AI infrastructure firms wavered, exemplified by a surge in credit default risk for companies like Core-Weave. In Asia, several planned initial public offerings for data-centre companies, such as Singapore’s Firmus and AirTrunk, were either delayed or canceled. In Australia, NextDC, a leading data-centre operator closely linked to AI infrastructure, lost nearly 40 percent of its market value in a single trading session following community opposition that stalled its expansion plans. The sell-off extended to the broader Australian market as resources stocks, which represent an indirect form of AI exposure, faced downward pressure.
Some fund managers see this market correction as a possible turning point. Todd Hoare, head of public markets at LGT Wealth Management, suggests that the future winners may be companies that adopt AI to enhance their operations rather than those providing the underlying technology infrastructure. He notes a progression of capital allocation from semiconductor makers like Nvidia to cloud giants, chip foundries, and resource bottlenecks such as copper and electricity suppliers. According to Hoare, “AI adopters” — typically traditional, asset-heavy businesses like retailers, logistics firms, and miners — have the potential to realize significant margin improvements. Even small margin gains in these low-margin industries can translate into substantial profitability increases, which the market tends to reward generously.
Drawing parallels to past industrial booms, Hoare points to history showing that infrastructure builders often fail to capture enduring value. For instance, although railroads were critical to 19th-century expansion, many went bankrupt, whereas companies like Standard Oil capitalized on the transformative infrastructure to build dominance. Similarly, during the dotcom era, while companies such as Cisco constructed internet infrastructure, it was Amazon, Google, and Netflix that ultimately generated outsized returns.
In the Australian market, Hoare sees a natural hedge due to limited direct AI exposure outside firms like NextDC and Goodman. He highlights that the country’s market composition, with its heavy allocation to labor-intensive and asset-heavy sectors, could benefit from AI-driven efficiency improvements. However, he expresses skepticism about banks retaining AI-driven cost savings amid competitive pressures that may force these benefits to be passed on to customers.
Other market participants view the current environment as favorable for sectors beyond AI beneficiaries. Emanuel Datt, chief investment officer at Datt Capital, points to historically low valuations in small-cap stocks compared with large-cap counterparts, a dynamic often associated with strong future returns. He notes that private equity interest is mounting in undervalued, cash-generative small caps and anticipates that any AI-related market uncertainty could spur a rotation back into less correlated areas.
Conversely, Eric Marais, investment specialist at Orbis Investments, urges caution in interpreting recent market moves as a definitive turning point. His team’s analysis of tens of thousands of diversified portfolios since the release of ChatGPT in 2023 found that broad market indices continue to outperform most combinations of stocks, indicating that narrow market leadership rewards index-like positioning. Marais contends that genuine diversification requires fundamentally different portfolio construction rather than timing sector rotations.
While the recent volatility highlights the risks embedded in extended capital expenditure cycles on AI infrastructure, experts agree this does not signal the end of the AI investment trend. However, history suggests that the lasting financial gains often accrue to companies effectively leveraging infrastructure advances, rather than those merely building it.