For many years, software-as-a-service (SaaS) firms, which provide the digital utilities we use every day, such as Microsoft 365, have been prized for their consistent high returns. They are known for their high margins and predictable revenue and have become the foundation of many high-quality investment portfolios. People saw them as “fortress” businesses that could survive any economic conditions. However, in the first few months of 2026, that reputation has been shattered. A massive wave of selling has hit the sector, leaving many of the world's most highly regarded firms trading at their lowest valuations in years.
This drop comes from a growing fear about the AI revolution. Many now worry that AI might replace some of these companies entirely. The sell-off has been referred to as the “SaaSpocalypse”. It has created a major divide in the market. One side believes the SaaS industry is in permanent decline. The other sees a rare chance to buy superb businesses at a discount. The challenge is to tell the difference between a real risk and a distraction. That requires an understanding of where a company actually gets its strength and whether a piece of software is just a tool that anyone can copy.
Software-as-a-service firms fear AI could dismantle their business model
We are moving from a world where software was just a tool to one where software acts as an autonomous agent. This is easily the biggest change in the way business is done since the internet first arrived and might yet rival the industrial revolution in terms of impact. The fear for software firms is that AI might soon dismantle their competitive position. Things came to a head when one of the world's largest AI companies, Anthropic, released its latest tools, Cowork.
Try 6 free issues of MoneyWeek today
Get unparalleled financial insight, analysis and expert opinion you can profit from.
Sign up to Money Morning
Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter
Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter
Part of the strength of software-as-a-service businesses is perceived to be the difficulty of writing and deploying code. Cowork threatened this by offering natural language programming. Essentially, anyone who can speak clear English can now code software into existence without having to learn how to code. For a few years now, AI has been a useful tool in helping coders improve their code, but Cowork goes beyond this. It can build a customised tool to manage specific work in comparatively little time. The reaction was panic.
The argument goes that if a non-coder can speak a project-management tool into existence, then why would they pay for an expensive software system designed by a third party? This threat of zero-barrier entry is what caused the SaaSpocalypse. Investors looked at the huge profit margins of SaaS businesses and instead of seeing it as a strength, as they have for years, they saw a target. There is a fear that white-collar work, such as processing invoices, will become commoditised. If the work of a lawyer or an accountant is mostly about following a process, then the software helping them is suddenly very vulnerable to a cheaper and faster AI alternative.
However, this fear might overlook how businesses actually work. The biggest threat from Cowork might not be to the software owners, but to the people using it. We are seeing a huge disruption in jobs because AI is great at replacing process-heavy tasks, such as auditing or conveyancing. Perhaps the most extreme example was fintech company Block, which recently announced that it would cut its staff from 10,000 to fewer than 6,000, with CEO Jack Dorsey saying, “intelligence tools have changed what it means to build and run a company”. Most probably, there will be many more similar announcements in the coming months and years.
But while the threat to jobs is clear, better or cheaper software doesn't always replace an established one. This is where the market seems to misunderstand the advantage of being the established player in a market. The strength of a business such as Sage, for example, which provides accountancy software to thousands of small businesses in the UK, isn't just about its range of features. Rather the software serves as a system of record; the single, trusted source of truth for the most sensitive data a business has.
Understanding nuance will bring profits
For the typical Sage user, the software is a tiny part of their costs, but their entire internal process is built around it. Switching to a cheaper AI-built alternative involves far more than “vibe coding” a replacement. Vibe coding is where a user describes the “vibe” or desired outcome of a program in plain English, and the AI writes the code, but relying on this for critical infrastructure involves a massive risk to the business. If an AI-generated tool makes a mistake, the consequences can be catastrophic. Even before the advent of AI coding tools such as Cowork, software testing has been a longer and more important process than the creation of the software itself.
There is also the so-called maintenance trap. Once software is being used it needs to be improved over time to meet the changing needs of customers. Building a prototype with Cowork is almost free, but maintaining that code and keeping it secure is a different problem altogether. Most businesses would rather outsource that responsibility to another company that stays on top of relevant law than carry the risk of mistakes themselves. For many, a third-party guarantee is worth more than the savings from using AI instead.
One area where software-as-a-service businesses might suffer is where they charge on a per-user or per-licence basis. Using Block's sharp reduction in staff as an example, if AI makes a worker twice as productive, a company might decide it needs 50% fewer workers. That means 50% fewer software licences to pay for. This is a problem for SaaS businesses that make money based on the number of people using their platform. To counter this, we will probably see a shift away from per-licence pricing toward models based on outcomes or usage. The winners will be the ones who can move from being a tool for individuals to becoming an enterprise-level system. The recent sell-off shows that the market hasn't distinguished between simple per-licence software and the mission-critical infrastructure that is built into the rules of a business.
The disruption is real, but it has layers. A new AI start-up can build a better sales tracker, but it can't easily build years of trust or a global network of trained users. In professional services, the skill that AI replaces is often just the manual part of the work. The value the software provider offers is the governance, the audit trail and the legal protection. As “vibe coding” becomes a more viable way to develop software, it allows more people to build their own tools. These unmanaged tools could, if not implemented with care, create a mess of data silos. That chaos often leads businesses back to the safety of a professional suite where the data are clean and there is a human to call when things go wrong.
The job of an investor is to find where the protection remains intact despite these technical challenges. AI is definitely pushing down the cost of making software, which theoretically could hurt profitability. However, it is also pushing down the internal costs for big software companies. A business such as Constellation Software, for example, can now maintain its products with much more efficiency. This could lead to rising profit margins even if growth slows down. The market panic has been so bad that it has treated every company the same. This lack of nuance creates opportunity for investors. If you look past the headlines about the death of software, there remain enduring moats built around mission-critical parts of a business or the long-term systems of record. Making sense of this really comes down to the difference between a simple software tool and the essential infrastructure a business cannot live without. When a firm handles the vital but repetitive tasks for an organisation, it becomes the main protector of its data and its processes.
Intuit has successfully integrated AI into its business
The businesses well placed to survive the 'SaaSpocalypse'
The real opportunities are found in the businesses that control the primary system of record. These are much more valuable than the companies that just provide a tool for people to do their work. Sage Group (LSE: SGE), for example, is a perfect example of a business protected by its place in a company's daily workflow. It provides the accounting and payroll software that small businesses across the UK rely on. Many firms are required by law to use these tools for tax compliance. For most of them, the cost of Sage is a tiny part of their overall bills, but the risk of switching to another provider is massive. Staff are already trained on the system and moving years of data is a technical nightmare. In early 2026, Sage reported that its new AI tools were saving customers between five and ten hours of work every week on administrative tasks. This makes the product even more vital and should keep the steady subscription income flowing.
Experian (LSE: EXPN) is the world's largest credit bureau and manages 1.3 billion data updates every month. In a digital world that is currently struggling with fake identities made by AI, Experian's verified credit histories have become a vital filter for the global banking system. Its tools for analysis and decision-making are built into the lending platforms of major banks. The firm saw its organic revenue growth accelerate to 8% in late 2025. As banks need accuracy when they make lending decisions, they are unlikely to swap a trusted data source for unverified AI models. It has a vast dataset that third-party AI models are unable to train on and this control over data allows Experian to protect the business while its internal AI models can improve the quality of the analytics it provides.
Constellation Software (Toronto: CSU) is the owner of hundreds of tiny, niche software companies that handle essential tasks, such as bus routes or hospital billing. The cost of the software to the customer is usually low, but its importance is huge. The chance of a new AI start-up copying thousands of these specialised workflows is slim. And by using tools such as Cowork, Constellation is in a good place, cost-effectively to improve the software it sells. Moreover, the firm is currently using its cash to buy up tiny software businesses that have seen their prices fall due to fears about AI. By applying its own efficiency rules to these new buys, Constellation is turning the market's panic into growth.
Oracle (NYSE: ORCL) has moved past its old image of just being a database firm and is now a vital part of the AI sector. It makes up the digital plumbing of the global economy, handling everything from airline bookings to national banking systems. It has a huge advantage because it holds the key private data that AI needs to function. It is also putting serious money into the physical side of the business, with billions going into high-end data centres. This spending is massive and it has caused some tension in the market, but moving these essential tasks to a different provider is so difficult that customers tend to stay put.
Intuit (Nasdaq: INTU) is a similar business to Sage, but with different geographic exposures. It has successfully integrated AI into its business with its tax-agent model reportedly helping to find extra deductions that business customers were previously missing. These kinds of savings can quickly repay the modest cost of its QuickBooks service and has led to a customer retention rate of around 85%. For small businesses, QuickBooks Online serves as the main ledger and payroll system. While the tech for simple accounting has become cheaper, the difficulty of moving old financial data creates a very lasting relationship. Intuit is using AI to move into live services, further embedding it in customers' workflows. This helps protect the firm against any loss of income from having fewer people use the software. As with Sage, Intuit's software remains a necessity for smaller and mid-sized companies.
Relx (LSE: RELX) has moved on from being an old-fashioned publisher of journals such as The Lancet to becoming a provider of data analytics. By 2026, revenue from paper journals dropped to just 4% of its total sales as the business moved almost entirely online. Its legal and medical data sits behind private paywalls and is verified by experts. This archive is arguably where the value lies even should AI agents be used to analyse the data. However, there is a threat from general AI models that might become good enough for basic legal research. Relx needs to make sure its high-end tools are distinguishable from cheaper, more basic AI options.
Wolters Kluwer (Amsterdam: WKL) is similar to Relx and works in areas where large portions of digital revenue come from AI-powered tools. Its model uses experts to check the work to ensure that its medical and tax insights can be defended in court. The company's UpToDate platform is linked to improved patient outcomes in hospitals, which partly justifies its high prices. Even so, the firm has to spend a lot of money to stay ahead and it is investing heavily in product development to beat off new AI challengers. Profit margins are expected to hit 28% in 2026, but the long-term cost of staying ahead is something investors will have to watch.
Oracle is now a vital part of the AI sector
SAP (Frankfurt: SAP) is in the middle of a massive project to move its vast global customer base over to the cloud. This is a slow and complicated process, but the numbers show that it is starting to pay off. The company's main push is a program called RISE with SAP. It is in an advantageous position because its systems serve as the operational backbone for major organisations. However, this is a double-edged sword as its customers are also in a better position to develop and implement cheaper customised systems.
Salesforce (NYSE: CRM) is a leader in its field, but is facing a problem in that its customers need fewer white-collar workers. The company's Agentforce product has grown very fast from a small base. But if AI agents allow one person to do the work of several, firms will need to pay for fewer licences from companies such as Salesforce. Its traditional way of making money from fees paid per licence is vulnerable to AI.
Adobe (Nasdaq: ADBE), owner of Photoshop, faces a direct threat. AI content-generation could make traditional licences less valuable. Professional designers still rely on Adobe, but casual users can now use AI tools for basic design tasks and, in time, the value of creative work could diminish as AI tools improve and companies reduce budgets. This creates a problem where the company must use AI to remain relevant while possibly hurting its own subscription numbers. The challenge for Adobe is to prove that it is worth the cost when simpler AI alternatives are ubiquitous.
The best software-as-a-service buys for UK investors
Experian looks like one of the safest bets
For UK investors, Sage and Experian look like some of the safest bets. Sage is built directly into the regulatory rules that British small businesses have to follow. This makes its software an essential service rather than just a tool you can choose to replace on a whim. As the firm moves toward automated accounting, it is more likely to grow its margins than lose its customers. Similarly, Experian has a massive data barrier that AI cannot easily copy. In a world where AI-generated fraud is becoming a bigger problem, verified data is incredibly important. This creates a safety net for banks.
Outside of the UK, Oracle is a strong way to invest in AI infrastructure without paying the extreme prices found elsewhere in tech. By controlling the database layer and putting billions into physical data centres, it remains a major part of corporate infrastructure across the world. Intuit also seems well-placed to handle this shift. Its AI features are already saving real money for its customers. Finally, Constellation Software continues to be expert at buying mission-critical niche businesses. It is exploiting market fears to buy up small businesses at a discount. It is using the panic to fuel its own growth.
AI is undoubtedly going to change how we work. We are entering a time where tasks that rely on a set process are being automated at a fast pace. However, the recent sell-off in the software sector has been blind. It has ignored the fundamental difference between a simple tool and something that is vital to the business. If you can separate the businesses that own and offer mission-critical services from those that merely sell a tool, you can find top-tier companies at prices that don't reflect their long-term value.
This article was first published in MoneyWeek's magazine. Enjoy exclusive early access to news, opinion and analysis from our team of financial experts with a MoneyWeek subscription.