CIO Insights: Ctrl+Alt+Delete on software?

10 minute read

Even as the situation in Iran drags on and geopolitical risks remain elevated, markets are starting to look through the near-term noise. Beneath the headlines, structural themes, most notably artificial intelligence, continue to shape investor positioning and sector dynamics.

This shift is clearly reflected in US software equities, where the sector has corrected over the past few months, as much as 37% from recent highs. The sell-off reflects concerns around AI-driven disruption, pricing pressure, and the durability of the traditional Software-as-a-Service (SaaS) model. Beneath the surface, however, lies a more nuanced picture.

In this month’s CIO Insights, we look at whether the software sell-off reflects the fundamentals or has run ahead of them, and which parts of the sector have a stronger footing.

Key takeaways

• The software sell-off has been driven by a valuation reset due to investor concerns, rather than a deterioration in earnings. While stock prices have dropped from recent highs, the earnings picture tells a different story. Analyst earnings forecasts have softened, but nowhere near to the same degree as the correction in prices. In other words, market pricing has moved faster than the fundamentals. This is a key distinction because multiple compression – or a fall in how much investors are willing to pay for each dollar of earnings – can reverse quickly once sentiment shifts. A confirmed earnings deterioration, however, is a far slower and harder hole to climb out of.
• Not all software is equally exposed to AI disruption. Larger, more established names have the scale and product depth to absorb disruption stemming from AI, rather than be displaced by it. Certain sub-sectors are also more insulated: cybersecurity, for example, faces growing structural demand as AI widens the environment for cyber attacks. Even within more defensible parts of software, the picture varies significantly at the company level, which makes it difficult to identify the winners and losers ahead of time.
• History suggests that companies can evolve with disruption – and the market may be too negative on software as a whole. Past technological shifts – from physical retail to e-commerce, or from broadcast to streaming – did destroy some incumbents. But they also produced new winners who adapted. AI’s impact on the software sector could follow a similar pattern. For long-term investors, that means staying invested in the sector with a preference toward parts with the clearest structural footing, rather than exiting the sector entirely.

(See our Glossary at the end for a breakdown of the terms used in this article.)

A price reset, not a fundamental collapse

US software stocks have sold off sharply since late last year. The proximate trigger was a series of AI product announcements, most notably from Anthropic in January. This made concrete a disruption that’s already playing out in parts of the sector: autonomous AI agents replacing the human workflows that per-seat software subscriptions are built around (we’ll go into this in more detail in a later section). The sell-off that followed was one of the sharpest non-recessionary drawdowns for the sector in decades, pulling software's weight in the S&P 500 down to 8% from 12%.

Analyst estimates paint a different picture. Against a backdrop of growing AI disruption concerns, forward 12-month sales estimates for an expanded basket of software stocks have risen steadily since late September – up around 11% and wavering little throughout the sell-off.

Gartner, a market research firm, forecasts global software spending will be up 14.7% this year, accelerating from 11.5% in 2025¹. Forward 12-month earnings estimates moved more, climbing around 20% before partially retracing, and as of late April remain roughly 12% above where they started, as shown in Chart 1.

Over the same period, that broad index’s price fell by over 35% – though it has recovered some of those losses in recent weeks. When prices fall significantly more than earnings and sales estimates, the driver is the valuation multiple (what investors are willing to pay per dollar of future earnings), not the earnings themselves.

The broad software sector's forward 12-month price-to-earnings (P/E) ratio has fallen from around 38x in July 2025 to approximately 22x today – roughly 30% below its long-run average of ~34x , and now within touching distance of the S&P 500 at ~21x, as shown below in Chart 2.

Yet even after downward revisions, the market consensus still puts the broad software sector's forward 12-month earnings per share (EPS) growth around 33% this year – above the roughly 25% for the broader S&P 500. In short, software now trades at roughly the same multiple as the S&P 500, even as its earnings growth continues to outpace it.

This gap between price and fundamentals also matters because the two scenarios have very different recovery profiles. Multiple compression driven by sentiment can reverse quickly – a few solid earnings quarters, or clearer evidence that AI disruption is narrower than feared, could be enough. An actual earnings deterioration is a far slower and harder hole to climb out of.

With first quarter reporting now underway, the current disconnect in prices and fundamentals will be harder to justify if companies deliver in line with expectations and maintain their forward guidance. Indeed, Microsoft's latest results are consistent with this view: overall revenue grew 18% year-on-year, beating estimates, while its Productivity and Business Processes segment – which includes Microsoft 365 and Dynamics – grew 17%². That suggests the fundamental deterioration the market has been pricing in hasn’t materialised.

Not all software is created equal

While the sell-off has been indiscriminate, some companies and sectors are far better positioned than the market's broad reaction suggests – whether because of their scale and product depth, the nature of what they sell, or the structural demand they serve. 

Below, we look at two areas where the market's broad reaction appears to have overshot the underlying fundamentals: large-cap software, and cybersecurity.

Large-cap software fundamentals are still holding up

The core fear driving the sell-off is existential: that autonomous AI agents will replace the human workflows that per-seat software subscriptions are built around, eroding the revenue base entirely. Traditional software is priced on seats: a company pays for every employee using a CRM, a project management tool, a finance platform. If AI agents can execute those tasks without a human in the loop, the addressable market for per-seat licensing shrinks sharply. For companies whose revenue depends on the number of humans doing knowledge work, that is a genuine structural threat.

Larger, more established names – for example, Microsoft, Salesforce, or Oracle – sit in a more defensible position than the broad sell-off would imply. They have the engineering resources to embed AI natively into their own platforms, turning a potential disruptor into a product feature. Deep enterprise relationships and multi-year contracts also create switching costs and afford them greater revenue visibility. Additionally, diversified product suites make wholesale replacement by an AI agent far more difficult than displacing a single-function tool.

The data reflects this: since end-September 2025, forward earnings estimates for large-cap software companies are up 14%, compared to 9% for the broader software universe. This suggests earnings expectations have proved more resilient at the top end of the market than across the wider field. If a disruption were already materialising at the revenue level, you would expect forward revenue estimates to be falling.

Chart 3 shows this isn’t the case; those estimates are up almost 12% for large-cap software since the sell-off began, well above the 7.5% gain for the S&P 500. At this stage, the market is pricing in a disruption that the fundamental data hasn’t yet confirmed.

Cybersecurity has structural support, but also some near-term headwinds

Within the broader software universe, cybersecurity occupies a structurally different position. Security spending scales with the size and complexity of the digital infrastructure being defended, rather than headcount. And that infrastructure grows with advancements in AI: more cloud workloads, more data repositories, more AI agents operating with their own identities and access privileges, creating potential vulnerabilities.

The threat environment is escalating in tandem. CrowdStrike's latest Global Threat Report found that attacks by AI-enabled actors rose 89% in 2025 versus the previous year³. And 82% of detections were malware-free – meaning that attackers gained entry through valid credentials and trusted identity flows rather than traditional attack tools, making detection harder and the demands on security platforms greater.

Global decision-makers with procurement budgets are seeing the same picture: Chart 4 below shows the results of a 2026 World Economic Forum survey of nearly 900 respondents across 99 countries⁴. Of those, 87% said AI vulnerabilities had increased over the prior year.

Cybersecurity companies also benefit from moats that make displacement harder compared to elsewhere in software. The data they have accumulated over years of defending real environments is not something a newer tool can replicate easily. And once embedded, these platforms are hard to remove: regulators expect continuity, auditors demand documentation, and the cost of getting it wrong is too high for most organisations to risk a switch. Simply put, these are not tools that organisations swap out lightly.

That structural demand case sits alongside a near-term wrinkle. Forward revenue estimates for the cybersecurity sector have been revised down over recent months, as broader enterprise caution around IT procurement has weighed on even non-discretionary spending. In our view, this near-term revenue pressure reflects the budget cycle, rather than the overall direction of demand for cybersecurity solutions.

Differentiation within sectors means picking winners is harder than it looks

Even within the more defensible parts of the software universe, the picture can vary significantly at the company level. Table 1 below shows 6-month revenue and earnings estimate revisions for a selection of large-cap software names grouped by subsector.

Within cybersecurity, for example, analyst revisions to revenue and earnings expectations are already moving in different directions across names; and in some cases, within the same name. The same pattern is playing out in enterprise SaaS, where some names are seeing upgrades while others face cuts. In other words, a structural tailwind for a sector doesn’t necessarily translate into consistent outcomes at the company level – and the factors driving divergence are not always visible in advance.

What past episodes of tech disruption tell us about the current sell-off

As the adage goes, "history doesn't repeat itself, but it often rhymes." Past episodes of technological disruption show that pricing a new technology as an existential threat to an entire sector tends to overshoot. The more common outcome is a restructuring of the impacted sector: weak players with indefensible positions get displaced, while those with genuine structural advantages adapt and often come out ahead. Two recent episodes illustrate the pattern.

Episode 1: How brick-and-mortar retail didn’t crumble with e-commerce 

When Amazon began its rapid expansion in the 2000s, a prevailing narrative emerged that physical retail was finished. Parts of it were: brick-and-mortar retailers like Sears and Toys R Us filed for bankruptcy in the 2010s amid rising e-commerce penetration, with online sales growing from roughly 4% of US retail in 2010 to more than 16% by 2025⁵. The advantage of relying solely on physical shelf space diminished.

But physical retail didn’t disappear. The winners were retailers that turned their physical footprint into an omnichannel advantage – operating seamlessly across stores and online. Take Walmart: even as Amazon scaled rapidly over that period, revenue rose from about $408 billion in FY2010 to $713 billion as of the latest fiscal year⁶. The key was integrating its store network, distribution infrastructure, supplier relationships, and digital channels into a single retail system; stores became assets for pickup, delivery, returns, and fulfillment rather than a legacy liability.

Episode 2: How legacy media adapted to the shift to streaming 

When Netflix launched its streaming service in 2007, cable and broadcast dominated how Americans watched television. Over the following decade, that dominance eroded steadily as viewers shifted to on-demand platforms. By May 2025, streaming had surpassed broadcast and cable combined in total US viewing time⁷ – a milestone that would have seemed implausible when Netflix first launched.

But media companies with deep content libraries found a way through. Disney launched Disney+ in 2019 and crossed 100 million subscribers within 16 months⁸, while Warner Bros. Discovery scaled HBO and its broader content library to reach around 100 million subscribers globally. Alongside Netflix and Amazon’s Prime Video, a small group of scaled platforms came to dominate streaming – each built on content libraries that newer entrants could not quickly replicate. Revenue models compressed, but companies with strong content and distribution adapted and found new ways to monetise their content.

Our take on the software sell-off 

The market seems to have sold off software broadly, without distinguishing between companies that are structurally exposed and those that aren’t. The fundamentals – at least so far – don't support the scale of that reaction. History shows us that past disruptions followed a consistent pattern: broad pessimism at the outset, a restructuring of the sector, and a split between companies that adapted and those that didn't. AI’s impact on software is unlikely to be different. 

For long-term investors, that means staying invested in the sector, with a preference toward parts with the clearest structural footing – in particular, large-cap software and cybersecurity – rather than hitting Ctrl+Alt+Delete on software altogether.

Authors

Stephanie Leung, Chief Investment Officer

Stephanie and her team oversee the full spectrum of investment products and portfolios offered at StashAway. She brings more than two decades of investment expertise across multiple asset classes. Prior to joining StashAway in 2020, she managed investment portfolios at institutions such as Goldman Sachs and multi-billion dollar family offices in the region.

Justin Jimenez, Head of Macro and Investment Research

Justin has more than a decade of experience in economic and investment research, and contributes to shaping the investment office's views on the global economy and asset classes. Prior to joining StashAway in 2022, he was an economist at Bloomberg.

Glossary

Software-as-a-Service (SaaS)

A business model where software is delivered over the internet on a subscription basis. Customers pay a recurring fee, usually monthly or annually, to access the product.

Switching costs

The financial and operational burden of replacing one product or service with another. When switching costs are high, customers may stay even if an alternative exists, because the cost of migrating can outweigh the benefit.

Addressable market

The total pool of potential customers or revenue available to a product or service.

Revenue visibility

How predictable a company's potential future revenue is. Long-term business contracts, for example, give clearer visibility into future income.

Moat (business)

A durable competitive advantage that protects a company’s business from rivals. Moats may include proprietary data, technology, or customer relationships.

Multiple (valuation)

How much investors are paying for each dollar of a company's earnings, revenue, or other financial metric. For example, if a company earns $1 per share and the stock trades at $25, the earnings multiple is 25x.

Earnings per share (EPS)

A firm's total profit divided by the number of shares outstanding. If a company earns $10 million in profit and has 5 million shares, the EPS is $2. 

References

1. Gartner. (2026). Gartner Forecasts Worldwide IT Spending to Grow 10.8% in 2026, Totaling $6.15 Trillion. Retrieved from: https://www.gartner.com/en/newsroom/press-releases/2026-02-03-gartner-forecasts-worldwide-it-spending-to-grow-10-point-8-percent-in-2026-totaling-6-point-15-trillion-dollars
2. Microsoft Corporation. (2026). Microsoft Fiscal Year 2026 Third Quarter Earnings Press Release. Retrieved from: https://www.microsoft.com/en-us/investor/earnings/fy-2026-q3/press-release-webcast 
3. CrowdStrike. (2026). 2026 Global Threat Report. Retrieved from: https://go.crowdstrike.com/rs/281-OBQ-266/images/CrowdStrike-2026-Global-Threat-Report.pdf
4. World Economic Forum. (2026). Global Cybersecurity Outlook 2026. Retrieved from: https://www.weforum.org/publications/global-cybersecurity-outlook-2026/
5. Federal Reserve Bank of St. Louis. (2026). E-Commerce Retail Sales as a Percent of Total Retail Sales (ECOMPCTSA). Retrieved from: https://fred.stlouisfed.org/series/ECOMPCTSA
6. Walmart Inc. (2026). Walmart Reports Q4 and Full Year FY2026 Results. SEC Filing (8-K), 19 February 2026. Retrieved from: https://www.sec.gov/Archives/edgar/data/104169/000010416926000032/earningsreleasefy26q4.htm
7. Nielsen. (2025). Streaming Reaches Historic TV Milestone, Eclipses Combined Broadcast and Cable Viewing for First Time. Retrieved from: https://www.nielsen.com/news-center/2025/streaming-reaches-historic-tv-milestone-eclipses-combined-broadcast-and-cable-viewing-for-first-time/
8. The Walt Disney Company. (2021) Disney+ Tops 100 Million Global Paid Subscriber Milestone. Retrieved from: https://thewaltdisneycompany.com/press-releases/disney-tops-100-million-global-paid-subscriber-milestone/