Carmignac: Software-as-a-Compounder

June 26, 2025

In the past, technological revolutions followed a predictable pattern: they began with breakthroughs in infrastructure, particularly hardware, before software applications followed. Think of IBM and Microsoft, for example.

By Obe Ejikeme, Fund Manager, Analyst, Carmignac

In the early 1980s, IBM and Apple were the pioneers of the personal computing revolution, with groundbreaking hardware that led IBM to become the world's largest company in 1986. When software company Microsoft went public, it had a market capitalisation equal to only 1% of IBM's at the time.

We don't need to remind you who came out on top in the long run. We are now seeing a similar innovation cycle in the AI sector. To date, nearly $1 trillion has been invested in AI infrastructure, and new players such as DeepSeek and Mistral AI, together with efficiency improvements, are driving down the cost of using AI.

Investors are now looking for concrete applications that deliver returns. Software companies seem to be in an excellent position to integrate AI into real-world business applications.

Recurring revenue models

The power of software lies in the transition from perpetual licences to recurring subscription-based revenue models. The shift to Software-as-a-Service (SaaS) subscriptions has fundamentally changed the economic DNA of these companies. Unlike traditional licensing models, which require ongoing sales efforts to secure one-time payments when licences expire, SaaS platforms generate annuity-like cash flows through automatic renewals and usage-based pricing. When customers switch from one-time purchases to ongoing subscriptions, they are effectively paying for the future value of the product up front while giving suppliers continuous visibility into revenue streams.

Software companies appear to be well positioned to integrate AI into real-world business applications.

The financial mechanisms are clear: if a SaaS company retains 90% of its customers and manages to grow new bookings by 30% annually, revenues can triple within five years. This contrasts sharply with industrial sectors, where comparable growth requires enormous investments. Adobe is an example of this principle: their transition from selling Creative Suite licences to Adobe Creative Cloud was very successful. From 2004 to 2014, Adobe's revenue growth under a licence-based model was approximately 9.5% annually. After the introduction of Creative Cloud, revenue growth nearly doubled to 18% annually. During this period, Adobe Creative Cloud accounted for more than 90% of revenue¹. This demonstrates how powerful the SaaS model is in driving recurring, or compounding, growth.

Software is scalable

The unique cost structure of software – high fixed costs for development but only marginal distribution costs – creates operational leverage that is unmatched in other industries. Once development costs have been paid, every incremental software dollar or euro goes straight to profit. For investors, this translates into recurring growth, free from the inflationary and cyclical pressures felt by capital-intensive industries.

When an enterprise software provider such as SAP decides to move to a cloud-first approach, the reorganisation can involve billions of pounds and thousands of employees. However, as such a transformation project reaches critical mass, each new customer will lead to margin improvement. At present, SAP's success in transitioning to a cloud solution is clearly visible: growth in cloud software is over 25% and operating profit is up 28%². With further technological development and the emergence of AI, we expect to see more similar patterns of transformation and growth.

Flywheel for reinvestment

Leading software companies are using their retained earnings to tap into adjacent markets. That is why we focus on companies that reinvest significantly. Microsoft's annual development budget of $19 billion funds initiatives ranging from AI co-pilots in Office 365 to quantum computing.

This “reinvestment cascade” is transforming software platforms into engines of compounding, as Microsoft's example illustrates in Figure 1.

Disruptive effects

Before ChatGPT and, more generally, generative AI emerged, SaaS was one of the fastest-growing business models. In the 2010s, SaaS companies' revenues grew by more than 300% – five times faster than the S&P 500 over the same period. This remarkable trajectory has paved the way for a crucial debate: will AI disrupt software business models, or will it usher in an era of unprecedented innovation and growth?

As SAP CEO Christian Klein noted in response to the recent Deepseek news: ‘You see the progress in the performance of large language models or chips. So it’s getting cheaper. And this is where we benefit, because we are at the forefront, bringing in AI and creating high-value applications for businesses."³ We believe that AI-driven productivity gains will benefit software companies. ServiceNow's foray into AI, for example, began in 2017 with the introduction of Predictive Intelligence, a tool designed to improve IT service management. By analysing historical incident data, the platform automated categorisation and routing, enabling it to diagnose hardware-related problems with 99% accuracy. This reduced manual intervention and ensured faster resolution times, setting a precedent for the role of AI in workflow optimisation.

One of the most promising uses of AI is what we call “agentic AI”. Unlike generative AI, which focuses on content, agentic AI emphasises autonomous decision-making and action, enabling systems to adapt, reason and perform tasks independently without human supervision. This paradigm shift has the potential to revolutionise entire industries by automating complex workflows, optimising business processes and driving scalability. This is expected to usher in a new wave of automation that could generate $500 billion in additional revenue for the enterprise software industry by 2035⁴.

Self-reinforcing trend

With companies planning to allocate more than 50% of their IT budgets to cloud/SaaS by 2025 (up from 10% in 2015), software providers appear to be becoming the cash cows of the digital economy – creating recurring revenue streams and tapping into new markets, thereby adding shareholder value. They offer the rare combination of recurring revenue, high customer retention rates and the strong expansion drive needed for exponential compounding.

Software companies are the cornerstone of the technology sector. Investors can choose from a diverse selection of software applications, such as engineering software (Ansys, Cadence Design, Autodesk), cloud computing platforms (Microsoft, ServiceNow), enterprise resource planning software (SAP, Oracle) and specialised software, such as cyber security (Palo Alto Networks) and accounting software (Intuit). This selection reflects a long-term vision focused on creating shareholder value while capitalising on structural trends in the sector, with forecasts for annual growth (CAGR) of around 13.7% until 2030⁶.

^{1) Company data Adobe, Source: Bloomberg as of 20/02/2025.
2) SAP Q4 and FY 2024 financial results, Source: SAP.nl
3) Investor’s business daily, 28/01/2025: https://www.investors.com/news/technology/deepseek-ai-enterprise-software-sap-nvidia/
4) Redburn Atlantic, Business software call my agent, 19/02/2025
5) Gartner research, as of 9/02/2025
6) Grand View Research, Software As A Service Market Size, Industry Report, 2030}