As semiconductor stocks experienced a sharp sell-off this week, some investors are shifting their focus towards quality software companies that have shown resilience despite broader market declines. The KOSPI index dropped nearly 10 percent on Tuesday following news that memory chipmaker SK Hynix was scaling back production of next-generation high-bandwidth memory chips. Meanwhile, the Philadelphia Semiconductor Index lost 7.9 percent. Both indexes, however, remain well ahead for the year, with the KOSPI up approximately 80 percent and the Philadelphia index around 90 percent in 2024.

Daniel Martin, senior investment analyst and portfolio manager at Brisbane-based Alvia Asset Partners, noted an interesting market pattern during these sell-offs. Each time the KOSPI triggered a circuit breaker this year, investor capital quietly rotated into select software-as-a-service (SaaS) companies that he considers likely winners, rather than indiscriminately selling off all tech stocks. Martin pointed out that SaaS valuations, which had reached roughly 40 times forward earnings earlier, have recently fallen to around 22 times, prompting questions about whether these stocks are nearing fair value or revealing underlying opportunities.

Martin highlighted the IGV ETF, which tracks major U.S. software names, as a barometer for this segment. The ETF had rallied into 2025 amid optimism that artificial intelligence would boost subscription businesses, but the broader SaaS market has seen notable volatility. The BVP Nasdaq Emerging Cloud Index, a key gauge of cloud software stocks, rebounded from a 33 percent drop earlier this year only to decline about 20 percent in recent weeks, illustrating a challenging environment for many investors.

Martin emphasized that not all SaaS companies are equal and urged selective analysis rather than broad purchases. He cited ServiceNow as a prime example of a resilient software firm. With more than 450 of the Fortune 500 companies as clients, ServiceNow has established high switching costs and maintains pricing flexibility. The company has shifted from a purely seat-based revenue model to a hybrid approach, introducing "Now Assist Packs," which charge clients based on AI token usage. ServiceNow’s revenue guidance for these AI-driven products increased from $1 billion at the start of the year to $1.5 billion in the following quarter, while its core seat-based revenue remained stable. Furthermore, its early partnership with Nvidia in 2023 to develop AI for workflow automation positions it favorably amid rising AI interest.

In contrast, Martin drew a distinction between software firms and chip equipment providers tied closely to hardware demand. He cautioned that equipment manufacturers, often described as “picks-and-shovels” players for data center expansions, face greater risk if capital efficiency improves and hardware needs decline. Semiconductors currently make up around 18 percent of the S&P 500, but the sector has long been known for its cyclical nature, raising concerns about sustained demand if data center rollouts slow.

For exposure to the AI-driven economy outside of tech stocks, Martin expressed a preference for energy companies, particularly in natural gas and liquefied natural gas (LNG), which power data centers and help stabilize electricity grids amid growing renewable energy generation. Woodside Energy, an Australian LNG producer with a strong project pipeline, was highlighted as a notable example.

Martin also identified companies with valuable proprietary data assets that may benefit as AI adoption spreads. These include U.S.-based S&P Global and Morningstar, as well as News Corp, the publisher of The Australian.

Regarding investor behavior, Martin warned against blindly following passive investment strategies without understanding underlying exposures. He referenced an index developed by Bianco Research in partnership with Goldman Sachs that excludes AI and AI-enabling companies from the S&P 500, noting it showed a negative price return in 2024 even as the broader market advanced. This divergence illustrates that while the AI sector has generated extraordinary gains, returns elsewhere have been more muted.