Summary

Artificial intelligence (AI) is one of the biggest technological developments in history. Since 2025, the market has worried that AI will displace software-as-a-service (SaaS). Are leaders in SaaS such as Xero, TechnologyOne and WiseTech about to lose their competitive advantage? Chris Smith, Ausbil’s Head of Equities Research and Portfolio Manager, shares why these firms are adopting and adapting to become AI winners, and talks about why AI is showing opportunity across many sectors, including resources, energy, real estate and industrials.
Key points
  • We see AI accelerating growth in economic productivity and as the next logical phase of the revolutionary shift to the cloud, radically increasing capacity across almost every sector.
  • SaaS has revolutionised the distribution of software and services with cloud computing, with a number of players emerging as sector leaders with significant and defendable ‘moats’ – despite recent market fears.
  • The ‘on-premises to cloud’ transition rewarded companies which recognised that software delivery was changing and positioned themselves on the right side of that shift. AI presents a new capability for the existing software stack.
  • We believe AI does not selectively advantage start-ups over incumbents. In fact, scaled incumbents who lean into AI are in an advantaged position.
  • We see three types of opportunities for companies that lean into the boom in AI, including:
    1. new areas of value creation in AI and AI-related businesses;
    2. existing software and technology companies, including SaaS, that embrace andmonetise AI features for their customers and improve productivity; and
    3. non-AI companies that are beneficiaries of the AI revolution.
How is AI impacting companies, for example SaaS?
We are seeing in real time AI and its partner, ‘big data’, rapidly change the world in almost every part of our lives. AI1 is working at so many levels and well beyond what we can see, changing the world as we know it and offering significant opportunity for investors who know where to look, and what for.

While large language models have been around for a decade, it could be argued that AI’s true inflection point in the psyche of the general populace was the November 2022 launch of OpenAI’s ChatGPT (Chat Generative Pre-trained Transformer), a large language model-based chatbot that has proved a watershed in the potential for AI to change the world. As an example of how quickly ChatGPT was adopted, while it took 55 months for Spotify to reach 100 million users, ChatGPT reached the same number of users in less than 2 months (Figure 1).

Figure 1: The ever-faster growth to 100 million users 

Source: Huang, S., & Grady,P. (2024, September 20). Generative AI's Act Two. Sequoia. 

With the emergence of ChatGPT it became widely apparent that computers were now smart enough to help humans to solve problems without having to write code, instead, at the straightforward command of our voices and keyboards. But does this mean that code, which took years but can now be written in days with AI, is the end of technology companies as we know them and, importantly, the death knell for SaaS companies?

The recent market panic of early 2026 which some call SaaSpocalypse points at the fear that AI could spell the end for established leaders in SaaS. There are a number of compelling reasons we believe this is not the case, and that we see AI as not the end of SaaS, but the beginning of new productivity and products which can make SaaS and almost any other leader in the market even stronger. Most of our confidence stems from developing a fundamental understanding of AI and how it works – the popularised iceberg example gives a logical framework to help understand why AI is more likely to strengthen the competitive advantage of leaders in SaaS rather than weaken them (Figure 2).

Figure 2: The much-discussed AI iceberg makes a lot of sense

Source: Ausbil, adapted from Finbarr Taylor diagram – April 2026.

It is a risk to think of AI as simply the code, user interface and features that can be programmed to almost any task, including launching new platforms to compete with existing companies, such as SaaS, which have built and tested their offering over long periods of time. Assuming that AI, the tech we can see, is all there is – like the small portion of the iceberg – is taking the risk that nothing below the surface matters anymore or has little value for existing market leaders. For example, what is the value of decades of accrued proprietary client data or customer trust built over many decades? Can AI simply replace this? Are customers actually willing to shift to the next new idea with complex and critical processing like payroll, taxes, operations, billing, customer service, etcetera? Let’s take a closer look below the iceberg at what AI may not be able to replicate easily for new entrants.

Distribution. In a world where coding costs trend toward zero, distribution becomes a very valuable and scarce asset. In services, established brand and trust means that loyal customers are less likely to consider a new offering, especially incumbent customers that understand the depth of process and workflow, experience and IP that underpins the offering. In sensitive areas like accounting, invoicing, compliance, payroll, tax and accounting, distribution and incumbency matter a lot (like Xero); so too in freight (WiseTech); and in enterprise resource planning for complex councils and government organisations (like TechnologyOne).

Network effects. Customers in leading Australian SaaS businesses like Xero, Technology-One and WiseTech benefit from every other customer in the network effects of increased efficiency, shared IP growth, security, process excellence and productivity improvement. Such a mutually beneficial environment is hard to replicate in a startup and adds to the protective moat for these types of businesses.

Switching costs. In critical services like freight management, accounting and enterprise planning, customers are hesitant to switch if there is risk of disruption, especially in critical operational areas. Hesitancy to take on switching risks and costs further favours incumbent businesses.

Data moats. Years of proprietary customer data cannot just be replicated, modelled or found by AI as it is heavily guarded by the owners of that data. AI models rise and fall with the quality of their data, and no amount of learning and estimation can replace real world operational data often collected over decades, and the inherent corporate knowledge that results. However, AI can be richly deployed by incumbents across proprietary data to drive further improvements in experience, risk management and product improvement. Proprietary data provides valuable context for incumbents that an AI challenger cannot adequately replicate.

Brand, trust, knowledge and expertise. Critical services and software are a hard sell for established players, and an even riskier and harder sell for new entrants. Companies and decision makers typically suffer inertia in favour of critical systems that are working – especially in areas were companies like Xero, WiseTech and TechnologyOne are operating. A lot of this comes down to trust infrastructure that has been built over many years, and the demonstrated knowledge and expertise which as accrued between clients and their SaaS providers.

Regulatory and compliance infrastructure. Years of investment in compliance, regulation, corporate learning, system development, proprietary data and training have created regulatory and compliance services with leading SaaS platforms that are far greater than the sum of their coding. If anything, AI can be deployed by leading incumbent SaaS players to further strengthen the regulatory and compliance moat, even if the moat is largely based on trust and brand – in many cases, the asymmetric downside risks of change (cost, training, conversion, testing, and user acceptance) far outweigh the marginal cost benefits of a cheaper new offer for leading SaaS players in critical areas like Xero, WiseTech and TechnologyOne.

Ausbil has taken a systematic approach to looking critically at incumbent SaaS players to determine where they sit on the critical elements we have discussed on AI and AI success. Figure 3 illustrates by way of example three SaaS companies that we believe are well placed to deploy AI to further improve their leadership in their respective marketplaces.

Figure 3: SaaS – what constitutes winning in the time of AI
Source: Ausbil, 31 March 2026. Rating scale developed by Ausbil to illustrate our house view of the differences by company in each category. Ratings data is based on assumptions which are not guaranteed. Ausbil may or may not hold positions in these three SaaS names from time to time.

The ‘on premises to cloud’ transition rewarded companies which recognised that software delivery was changing and positioned themselves on the right side of that shift. Xero is a good example of this, whereas MYOB, which had market leadership before the shift to cloud computing, failed to make the shift to the cloud and lost their leadership to others.

Rather than ‘miss the boat’, SaaS leaders like those in Figure 3, are integrating the new capabilities of AI into their existing software stack. The companies that win will be those that embed AI into products with strong distribution, proprietary data, and deep workflow integration, and many of the other characteristics shown in the iceberg illustration (Figure 2 above). The companies that lose will be those whose products are made redundant by AI, or who lack the trust, scale and capability to compete in an AI-augmented market.

What the market may not have realised in the indiscriminate selldown of SaaS companies in early 2026 was the fact that the best names enjoy ‘moats’ and competitive advantages not just from AI scalability, but also from the non-AI features of their businesses, such as distribution (which involves trust, think of the legacy of Salesforce with its client base); expertise (think demonstrated leadership that is hard for new players to displace, like Xero for accounting); network effects (like total-solution software offered by Microsoft); switching costs (the cost of shifting from an existing leader in SaaS to an unproven startup); data moats (think of Tesla’s current advantage in historical trip data); data context; and regulatory and compliance infrastructure (think the persistent dominance of banks in the modern world of tech).

SaaS companies like Xero, TechnologyOne and WiseTech are each launching and driving monetisation using AI agents this year. Block has also launched its Goose AI agent and is embedding AI across both its Square (Seller) and Cash App businesses in the next 12 months, helping to drive revenue, and materially reduce costs (noting that XYZ has just announced an effective 40% reduction in headcount).

Life360, which develops and operates a mobile application focused on family safety and communication, has been growing its user base such that one-in-ten people in the US now use the app. With the acquisition of Fantix in 2025, Life360 is using AI across its platform to drive revenue, enhance advertising with AI-driven customer insights, path-to-purchase effectiveness data, and personalisation for optimised advertising campaigns to raise ad revenue, directly benefitting earnings and earnings growth.

In terms of these companies becoming victims to the AI revolution, we would strongly urge caution as there is far more to these companies than just code and software which is likely to ensure client bases remain and grow.
 
Where is AI adding opportunity?
Any company with a competitive advantage should be able to extend this with the application of AI. Given the immensity of the potential for AI and machine learning, not just to replace people but also to extend reasoning tasks beyond the human experience, it is difficult to identify the full complexity of this impact on companies. By focusing on how AI is already impacting earnings and earnings growth, we can focus on opportunity and risk for the listed companies in our investable universe.

While the long-term trajectory of AI remains less clear in terms of who will ultimately win the agent layer, there are a number of key trends that are emerging, as both disruptive and additive themes. The first of these themes is the power demanded by AI and data centres as in the build out of this technology.

Data centres provide the processing power and drive that is needed by AI, and for the deployment of AI by users in its various tasks. Processing demand from rising AI application is set to become so intense that by 2027, a server rack that had 8 GPUs using 41 KWh in electricity only five years earlier will have 567 GPUs using 600 KWh of energy and will be the size of a filing cabinet (Figure 4).

Figure 4: The emergence of ‘hyperscalers’, data and its power density



Source: Goldman Sachs (2025, August 29). How AI Is Transforming Data Centers and Ramping Up Power Demand.

The intensity of processing in hyper-scaling data centres, including cooling the heat generated by information processing, is becoming a major piece in the energy and electricity grid complex. The massive extra load AI is adding to data centre energy consumption is driving major capex growth in data centres and in the grids through which they communicate and source power.

Figure 5: AI is now a major driver of electricity demand for processing and data centre cooling


Source: Goldman Sachs (2025, August 29). 

According to Goldman Sachs, AI processing will account for ~28% of data centre energy demand by 2027, from negligible demand levels prior to 2020 (Figure 5). This is beneficial for companies generating energy, electricity, and the companies and resources needed for grid and data centre expansion, like copper, rare earths and lithium for energy storage.

However, from a sustainability perspective, with all this added data and energy demand, AI is also generating more carbon emissions. Carbon emissions from data centres will add +1% to global emissions by 2030, according to estimates from Goldman Sachs, which ultimately will need to be reduced through increased investment in renewable energy, and improved energy efficiency, which is something AI is already helping to improve in a virtuous cycle.
Are there AI tech companies on the ASX for Australian investors to gain exposure?
The Australian listed market does not have an investable semiconductor sector, other than some very small micro-cap companies that may loosely fit the description. The ASX does, however, have a range of technology companies that are engaged in AI and related businesses, as well as companies that are beneficiaries of the advances in AI technology.

In terms of AI and machine learning, which place heavy demands on data processing, other than SaaS players, ASX-listed companies like Megaport, Macquarie Technology, NextDC, Infratil’s CDC, and Telstra through InfraCo are all enablers of data centre connectivity and cloud storage. Moreover, real estate companies including Goodman Group and Charter Hall have built major global businesses in data centres and smart / robotic warehouses, driven largely by the explosive demand from AI and machine learning, and the benefits this brings to the operational efficiency of their tenants.
What types of non-AI companies are beneficiaries of the AI revolution?
In terms of non-AI companies, there is a rapidly growing list of companies that are investing in and profitably deploying artificial intelligence in ways that may directly improve earnings and earnings growth. AI is not just being deployed for improvement, optimisation or cost reduction; it is also creating new revenue lines for companies.

In the resources and energy sectors, the build out of data centres and hyperscalers is dependent on access to energy and commodities like copper, aluminium, rare earths and lithium, making resources companies a major beneficiary of the AI revolution. The fundamentals around demand from AI, including data centres, grid expansion, battery storage, AI demand and energy demand is seeing beneficiaries in the resources sector. This is expected to benefit companies like Lynas Rare Earths and Pilbara Minerals in critical minerals; and BHP, Rio Tinto and Sandfire Resources with large businesses in copper, essential for the tech and power buildout of AI.

In the industrials sector, Brambles, founded in 1875, is applying AI across its global logistics and container network to optimise space and shipping. Brambles is one of the most sustainable logistics and supply-chain companies globally, offering reusable pallets made from certified sustainable lumber. Brambles has been investing in digital and tracking technology for years, and benefits are starting to come through, but AI has seen the pace of improvement rapidly contribute to efficiencies, including digitisation, optimisation with the proprietary BRIX platform, enhanced analytical tools improving customer outcomes, and autonomous tracking, all contributing to cost savings and underlying earnings growth.

Online services have been another beneficiary of AI. Companies like REA Group in online real estate classifieds and CAR Group in car classifieds are achieving product and efficiency enhancement from the application of AI. The online digital environment lends itself well to the application of AI, particularly given the data-driven consumer interactivity that helps drive models for optimisation, decision support, targeted advertising and intelligent content customisation.
Why do we believe AI is an opportunity and not a bubble?
We believe that the AI revolution is a profound and fundamental advance in technology. Looking at the committed capex for US hyperscalers, we do not believe that AI is a bubble. When you combine the 2025-2027 estimates for capex, hyperscalers will have invested more on AI capex then the entire market cap of the ASX 200 (Figure 6). However, they will have generated an even larger level of operational cashflows which can more than fund this expansion in activity.


Source: Morgan Stanley equity analysts see hyperscaler capex approaching $740 billion in 2026 and $910 billion in 2027, as at Feb 2026.

As a house, we do not see the AI revolution as directly comparable to the bubble that precipitated the Tech Crash of 2001. The advent of the internet that drove the boom in technology investing at the end of the 1990s was a new technology where, across all sectors, it was not yet apparent what actual earnings accretion would be generated for existing and new companies
deploying this technology.

In the AI era, companies are able to leverage AI in ways that are immediately accretive to earnings and earnings growth. As we have seen with SaaS, leaders in this sector can apply AI to further enhance their competitive advantages which also benefits the outlook and resilience for earnings. We are already seeing large scale AI monetisation on the ASX, with companies like WiseTech, Xero and TechnologyOne, all of which have launched AI agents and are rolling them out to their customers.

AI has been under development and varying deployment for years. Data, online digitisation and cloud computing, which pre-date big data AI, are all revenue and earnings generative to good businesses. The growth in AI is both evolutionary and revolutionary; the nature of the technology is such that it can form the core of a business, or companies can incrementally experiment at low cost with how it can transform their earnings. We are observing the rapid emergence of AI applications that have the potential to improve business efficiency and earnings across almost every sector. The AI revolution is not just an opportunity in SaaS, but it is an opportunity for every leading company to expand and enhance its business and earnings, and with real benefits for investors like Ausbil that are focused on earnings and earnings growth as the key driver of returns.
References
1. A common definition of AI: In seeking to define the broad swathe of capabilities AI offers, most rely on the
comprehensive definition derived by the Federal Government of the United States , which includes:
a) Any artificial system that performs tasks under varying and unpredictable circumstances without significant
human oversight, or that can learn from experience and improve performance when exposed to data sets.
b) An artificial system developed in computer software, physical hardware, or other context that solves tasks
requiring human-like perception, cognition, planning, learning, communication, or physical action.
c) An artificial system designed to think or act like a human, including cognitive architectures and neural networks.
d) A set of techniques, including machine learning, that is designed to approximate a cognitive task.
e) An artificial system designed to act rationally, including an intelligent software agent or embodied robot that
achieves goals using perception, planning, reasoning, learning, communicating, decision making, and acting.
Download PDF