The advent of new AI-enabled business models has prompted questions around the durability of traditional software-as-a-service (SaaS) businesses, particularly those serving enterprise customers long considered to be sticky.

In the 12 months to January 2026, the median US-listed software company suffered a 16% share price decline. Notable examples include Monday.com (-55%), The Trade Desk (-54%), Wix (-43%), HubSpot (-36%), ServiceNow (-30%), Atlassian (-32%) and Intuit (-25%). Sector valuation multiples have compressed, with median enterprise value to next-twelve- months’ revenue (EV/NTM revenue) declining to 4.4-times, down from 6.0-times just 12 months ago.

In our view, the ‘fears of disruption’ are principally driven by three interrelated themes:

Firstly, Generative AI (gen AI) coding assistants and no-code application builders have reduced the technical barriers to entry in developing competing software.

Second, enterprise internal use of gen AI is slowing labour growth, diminishing the expansion opportunities for software providers who rely on selling additional seats to customers.

And, lastly, the use of AI agents to complete actions autonomously for its ‘employers’ disrupts the traditional architecture and purpose of software – a workflow on a screen and a database backend used by humans.

Importantly, not all software is judged equally. Investors are increasingly differentiating between perceived AI “winners” and “losers” based on observed fundamentals, particularly revenue growth.

Software stocks with growth rates above 22% are being rewarded with median EV/NTM revenue of 13.9-times (above long-term averages). Those with growth rates between 15% and 22% are trading on median 7.7-times (also above LTA), while those with growth rates below 15% trade on 3.5-times (well below LTA).

While higher interest rates, post-COVID normalisation and profitability focus are all factors affecting software growth, AI became a greater factor in 2025.

Below we address the three ‘fears of disruption’ and our framework for assessing the risks to our portfolio and new opportunities considered.

1. Software development is faster, but moats still matter

Tools including Cursor and Claude Code have compressed months of software development into hours. To consider the threat of disruption, we break traditional software into three buckets: features, products and platforms.

• Features: an individual function or capability, rarely sufficient to be a standalone business. These are highly exposed to fast development of new competing software via AI models

• Products: advanced software bundling many features into a self-contained solution. For quality software products, the moat is rarely speed of execution in building the code itself. The competitive advantage is often built-in business logic, complex workflows governed by industry-specific nuance and regulations, and unique data structures. Further, go-to-market capacity remains a significant constraint for emerging competitors, even if development build times significantly fall. These product businesses also have AI coding assistants at their disposal to implement AI features into its product set to compete with AI-native upstarts, albeit with the advantage of existing installed bases, brand strength and broad distribution. Yet, these businesses also face the reality that large enterprise customers will have greater capacity (and possibly the willingness) to develop competing internal tools.

• Platforms: complex and modularised systems of record, workflow backbones, and multi-sided network ecosystems with entrenched network effects are the most insulated and represent the most compelling opportunities. While gen AI can create software applications, building these advanced environments connecting real-world participants is a long and arduous journey, requiring extensive commercial negotiations and navigation of regulatory requirements and industry standards.

Our portfolio of software names, we believe, are mission-critical enterprise-grade platforms. ATI Global’s practice management software solutions (notably, LEAP) serve as the core operating system for small and mid-sized law firms, enabling end-to-end business management from matter and client data storage, workflow automation, in-built generative AI search, billing, accounting, and reporting capabilities. These capabilities are difficult to replicate and are informed by decades of customer data and feedback loops.

ATI’s InfoTrack has 2,000 unique third-party data integrations, connections and licenses secured over 30+ years to support due diligence searching. Separately, Cover Genius connects underwriters and merchants exclusively in over 300 jurisdictions globally covering 85% of GDP underpinned by hard-to-win regulatory licenses. Finally, Ofload uses software to aggregate hundreds of contracted trucking companies and offers exclusive freight services to contracted shipping companies.

2. Slowing seat growth and the shift in pricing models

A December 2025 McKinsey Global Survey found that while 63% of executives expect profits to rise over the next six months, only 30% anticipate headcount growth.

This trend has presented over the past few years, originally spurred by a shift to focus on profitability in a tougher macroeconomic environment, but now arguably accelerated by the implementation of gen AI. In response, large software vendors are increasingly leaning on price increases to sustain growth rates. For example, it is estimated that roughly three-quarters of Salesforce’s 2025 growth was generated via price.

In this context, software businesses with revenues linked to consumption, volume or outcomes (or a hybrid) are preferentially placed versus pure seat-based models. For ATI, Cover Genius and Ofload, growth rates have either sustained or increased in recent years, levered to the secular growth of their customers’ businesses and end markets – in legal searching and due diligence, embedded digital insurance, and digitally-enabled freight.

For seat-based software solutions, we expect the trend of models shifting to consumption, outcomes or value-based pricing to accelerate in the near term. The limiting factor remains that these alternative pricing models carry higher volatility and complexity in measurement.

3. AI agents and the timing of disruption

While AI-driven agentic workflows are widely anticipated, the pace of enterprise adoption remains uncertain. Current trends indicate a longer timeline than most expect.

For example, our portfolio company, Omnia Collective (an AI, data and digital transformation consultancy to government and enterprise in Australia), is still seeing strongest customer demand in areas of traditional machine learning, data engineering and analytics. This tells us that a total re-architecting of Australian organisations to suit gen AI-driven agentic workflows is still in its infancy and likely several years from the mainstream.

However, it is critical for traditional software vendors to begin experimenting with agentic solutions and setting clear roadmaps to ward off emerging threats. In the interim, switching costs remain elevated and distribution advantages prevail. For the stickiest software, price increases can be a source of funding for the additional investment required.

We consider vertical software providers deeply entrenched within highly regulated industries to be more insulated than broad horizontal providers. Horizontal software is more exposed to well-funded mega-cap competitor encroachment, while industry-specific software is more inclined to be challenged by specialised startups targeting a niche. By focusing our efforts on companies with significant revenues, large research & development teams, strong balance sheets, and high-quality leadership, we aim to back businesses who will out-innovate emerging competitive threats.

The AI and software landscape is rapidly evolving but highly uncertain. We remain vigilant in considering the risks and opportunities for our portfolio and incorporating these frameworks into our due diligence processes when evaluating new opportunities.