The Next Software Moat Won't Be Software

By Damon Anderson, 12 March 2026. 5 minute read.

Something exhilarating is happening in the technology world. People who historically wouldn't — or couldn't — have considered building software are suddenly prototyping it. Operators, marketers, analysts and founders are opening tools like Lovable or Claude and assembling working products at a speed that would have been unthinkable even a few years ago.

The most important shift is not the tools themselves, but the collapse in the time between an idea and a prototype. Until recently, building software that could scale came with a series of unavoidable taxes: scalability, robustness and coordination. Even relatively small experiments required engineers, infrastructure and weeks of effort before you could answer a very simple question: does this idea actually have legs?

Those taxes limited experimentation. Only ideas that seemed reasonably promising were worth attempting because the cost of exploring them was so high. When that friction disappears, the number of ideas worth testing explodes. That alone begins to change the shape of innovation.

The fast that ate the slow

For most of the SaaS era, the companies that won were simply the ones that moved faster. Engineering cycles were expensive and product iteration was slow, which meant the teams capable of shipping software quickly accumulated disproportionate advantage. No longer the big that ate the small. It was the fast that ate the slow. Companies that could design better interfaces, ship more reliable systems and execute complex engineering programmes more rapidly than their competitors steadily pulled ahead.

That dynamic produced the familiar SaaS playbook: build a better product, ship faster, out-execute competitors. The assumption underpinning that model was straightforward. The moat lived in the software itself.

As the cost of generating working software falls, however, the nature of that advantage begins to shift. If prototypes can be assembled quickly, then the most valuable skill is no longer simply building software efficiently. It becomes recognising which ideas are worth building in the first place.

When disparate industries collide

Looking back at some of the most influential technology companies of the past two decades reveals an interesting pattern. The breakthrough rarely came from software alone. Instead, it emerged when two previously separate industries or systems were connected in a useful way.

Google connected search behaviour with advertising markets and created a new way of allocating marketing spend. Uber connected real-time location data with transportation logistics. Stripe connected financial infrastructure with developer ecosystems.

In each of these cases the software was important, but it was not the original insight. The insight was recognising that two worlds which had previously evolved separately could be combined in a way that created entirely new value. Once that relationship was discovered, the software followed naturally.

The advantage lies where rivers of data collide

Much of the discussion around AI today focuses on scale. The prevailing assumption is that organisations with the largest datasets hold the tickets to AI's opening act. There is certainly some truth in that view.

Networks such as Visa process hundreds of billions of transactions every year, creating a global dataset of payment behaviour and fraud signals that would be almost impossible for a new entrant to reproduce.

But scale alone rarely explains the most interesting breakthroughs. Some of the most valuable insights emerge not from the largest datasets, but from the most unusual combinations of them. When financial systems are connected with behavioural signals, or operational data is combined with communication flows, entirely new types of information begin to surface.

An example close to the work I do is analysing product telemetry alongside sales conversations. The gap between what customers say they want and how they actually behave suddenly becomes visible — and that gap is often where the most valuable signals live.

These relationships are often invisible to incumbents because their systems were designed around a fixed (and often lucrative) view of the world. The advantage therefore does not come from possessing more data, but from seeing how seemingly unrelated data might interact.

AI is rewiring the interface layer

Historically, software was defined by its interface. Humans navigated applications through screens, menus and workflows, and the user experience became the primary site of differentiation.

AI is beginning to change that dynamic.

Increasingly, AI acts as a mediator between humans and the underlying software systems. Rather than navigating individual applications, users describe an outcome and the AI coordinates the tools required to achieve it. When that happens, the importance of the interface diminishes slightly. The real leverage moves beneath the surface, into the systems the AI can access and the relationships between the datasets those systems contain.

In other words, value begins to migrate away from the interface layer and toward the structure of the underlying information.

The new moat in quirky data relationships

All of this points to a slightly uncomfortable conclusion. The next generation of software moats may be much harder to recognise than the last.

In the SaaS era the source of advantage was usually one of two things: fast disruptors with a better product experience, or large incumbents who owned vertical data stacks.

In an AI-mediated world, the advantage may sit deeper in the stack. It may lie in the unique ways an organisation connects systems, datasets and workflows that were never originally designed to interact. Those connections often compound when the right data relationships reinforce each other, creating signals that competitors cannot easily replicate.

When that happens the effect is multiplicative rather than additive. One system combined with another does not simply produce more information. It produces a new lens through which the world can be understood.

That possibility is what makes the emerging landscape so interesting. The next major breakthrough in software may not come from a better interface, a faster development cycle or even a larger dataset. It may come from someone noticing that two systems which have always existed separately might, under the right circumstances, belong together.

And if that is true, the next durable moat in software may not be software at all.