Software sell-off: Singapore’s private wealth gets more discerning on AI-era technology exposure

[SINGAPORE] Driven by the rapid evolution of generative artificial intelligence, Asia family offices and high-net-worth allocators are actively re-evaluating their underlying software exposure in public and private markets.

This is amid growing concerns that traditional software-as-a-service (SaaS) models are losing their structural moats.

Market volatility has underscored these anxieties, highlighted by the iShares Software ETF plummeting 30 per cent – a decline recorded over just a few weeks – on fears of AI displacement before staging an uneven 40 per cent rebound, though that has since tapered.

“Moves that big over a few weeks are a positioning story,” said Kenneth Goh, director of private wealth management at UOB Kay Hian. “The rebound has been uneven, led by the more embedded software names.”

Yet many industry insiders argue that the initial market panic fundamentally misunderstood how AI integrates with enterprise architecture.

“There is a tendency to overstate the idea that AI somehow replaces software. In reality, AI sits on top of software,” said Neil Synnott, regional chief commercial officer for the Asia-Pacific at IQ-EQ. “It depends on underlying platforms, data and workflows to function properly.”

Echoing this view, William Chow, deputy group CEO at Raffles Family Office, added that autonomous “agentic AI” will not replace existing enterprise software but will instead become a massive consumer of these platforms.

Family offices positioning

Rather than fleeing the software sector entirely, private wealth is staying the course – albeit with far greater risk discipline. This heightened selectivity is directly reshaping family office asset allocations.

This is visible from market flows, where James Chye, head of investment advisory solutions at Bank of Singapore, noted that there is some tactical re-engagement into software after the recent sell-off, mainly in higher-quality, large-cap names with clearer AI positioning.

However, he noted that this is not a broad-based rotation and investor allocations continue to skew towards semiconductors and AI infrastructure, where earnings visibility and capex tailwinds are stronger.

“Overall, sophisticated clients including Singapore family offices are selectively investing into the technology space via the public market and equity funds, preferring to stay diversified across the AI stack including software, semiconductors, infrastructure and enablers where earnings visibility is stronger,” said Chye.

When it comes to deploying capital, the public versus private question looms large. Chow noted that there is an active secondary market where institutional investors were offloading private credit stakes, often at substantial discounts due to AI fears and liquidity needs, and they would remain selective to make broad, systemic bets for private markets.

“Navigating this space requires rigorous due diligence to ensure the underlying debt isn’t tied to fundamentally obsolete technology. At current valuations, we find public markets offer greater flexibility and transparency (than private markets),” he said.

“Within software, we are inclined towards specialised observability platforms, proprietary databases and cybersecurity while remaining more selective in software segments where AI may compress margins or reduce barriers to entry,” he added.

To identify winners, investors must look beyond categories to structural revenue metrics. Isaac Lim, chief market strategist at Moomoo Singapore, pointed to the percentage of revenue that is usage-based versus seat-based, management speed in transitioning business models and identifying second-order beneficiaries as critical evaluation criteria.

Navigating sector segmentation

Driven by these shifts, investors are moving away from treating software as a monolithic asset class, opting instead for strict differentiation.

Charu Chanana, chief investment strategist at Saxo, said that her team has been seeing interest come back into the space, but is much more selective than in AI hardware.

“Clients are not simply buying ‘software’ as one bucket. The conversations are shifting towards which software companies can prove AI monetisation, which ones can defend pricing power and which ones risk being commoditised,” she said.

She noted that SaaS can be differentiated into three broad groups: AI beneficiaries – such as cybersecurity firms and cloud infrastructure; AI adapters such as established software companies that can embed AI into existing products; and AI vulnerable names – that have point solutions, weaker moats, high seat-based pricing, or products that AI agents could replicate or compress over time.

Abhilash Narayan, director and investment strategist at HSBC Private Bank and Premier Wealth, echoed this need for differentiation.

“While AI could disrupt the business models of certain companies which focus on a single use case or feature, we believe the large SaaS companies which provide vertically integrated capabilities and companies with access to proprietary data can actually benefit from AI-driven changes,” he said.

Exposure in private credit

This granular scrutiny is extending heavily into private credit portfolios, where software assets comprise about 25 per cent of the US direct lending market, noted industry experts.

Randy Sim, group CEO of Singapore Exchange-listed IFS Capital, observed that many investors are starting to look for opportunities in software debt in the private credit space since the sell-off, highlighting that most software debt, especially among borrowers with resilient business models, continue to perform.

Synnott added that there is growing interest in secondary and stressed situations, particularly where valuation adjustments have created attractive entry points.

This has led investors to take a much more hands-on approach to understand underlying credit quality, particularly in portfolios with higher exposure to software and tech-enabled businesses.

Nevertheless, the valuation pressure has promoted a sharp tightening of underwriting in private credit deals in the space, as clients focus on the resilience of the business model in an AI-influenced market, balance sheet strength and leverage and cash-flow sustainability.

“A rebound in listed software does not automatically reprice a loan held in the fund, and those loans are valued off the borrower’s cash flow,” Goh said. “What matters is whether borrowers can still service that debt.”

Narayan said: “Overall, while there could be further negative headlines, the asset class continues to offer an attractive yield, but manager selection, especially the focus on disciplined underwriting and liquidity alignment, remains key.”

Ultimately, the paradigm shift requires a focus on longevity over hype.

“The opportunity is not in chasing every beaten-down SaaS stock. It is in finding companies where AI improves the product, expands the addressable market, or strengthens customer lock-in,” said Chanana.

“The risk is that some software companies may see faster product cycles, lower switching costs and more pressure on margins as AI makes it easier for competitors to build alternatives,” she added.