When Engineering Stops Being the Moat
Over the past few months I've had a front row seat to something I don't think the software industry has fully appreciated yet. I watched a mature enterprise platform being re-engineered by a single senior engineer using AI development tools. The work would previously have required a sizeable engineering team and many months, if not years, of development. It wasn't that AI magically solved every problem or removed the complexity of the platform. Years of embedded business logic, integrations and operational knowledge still needed to be understood. What struck me was something much more fundamental. The economics of building software are beginning to change.
It reminded me of what's happened in the automotive industry over the past decade. For years, European manufacturers built an extraordinary competitive advantage around internal combustion engines, transmissions and manufacturing excellence. Those capabilities took decades to develop and created a formidable barrier to entry. Building a globally competitive automotive company wasn't simply a question of capital. It required engineering expertise that few organisations possessed. Electric vehicles didn't remove the need for engineering, but they fundamentally changed where engineering mattered. As drivetrains became simpler and software became a greater source of differentiation, one of the industry's biggest historical moats became less important. Companies such as BYD were suddenly able to compete on far more equal terms because the barrier to entry had shifted.
I'm increasingly wondering whether AI is doing something similar to software. For decades, one of the greatest competitive advantages in enterprise software was the ability to build. Large engineering organisations, years of product development and significant investment created platforms that were incredibly difficult to replicate. The depth of engineering itself became part of the moat. Today, AI-assisted development doesn't eliminate the need for great engineers, but it does change the amount of engineering effort required to build, iterate and modernise software. Capabilities that once demanded large teams and significant investment are increasingly becoming achievable with much smaller teams operating at an entirely different pace.
This raises what I think is the more interesting question. If engineering effort becomes less scarce, where does competitive advantage move? History suggests that when one source of advantage becomes more accessible, value rarely disappears. Instead, it migrates somewhere else. My suspicion is that software companies will increasingly differentiate themselves through customer understanding, workflow integration, proprietary data, trust, ecosystem participation and their ability to help customers realise value from the technology they've built. In other words, many of the capabilities that sit outside the code itself may become more valuable than ever before.
What makes this particularly interesting is what it could mean for innovation more broadly. Software innovation has naturally concentrated around places with the deepest pools of engineering talent. Silicon Valley became Silicon Valley for many reasons, but access to exceptional engineering capability was undoubtedly one of them. If AI continues to reduce the effort required to build world-class software, that advantage begins to diffuse. The next generation of category-defining software companies may not emerge because they can assemble the largest engineering teams. They may emerge because they understand a customer problem better than anyone else, can iterate faster than their competitors and no longer require hundreds of engineers to compete globally.
I don't think we're witnessing the end of software engineering. If anything, the very best engineers are becoming even more valuable because AI amplifies what they're capable of achieving. What I do think we're witnessing is a shift in where competitive advantage comes from. Just as electric vehicles changed the economics of automotive engineering, AI has the potential to change the economics of software development. And if the economics are changing, then the sources of value creation are likely to change with them.
Which leads to the question I find most interesting.
Are we simply making engineers more productive?
Or are we witnessing the moment when engineering stops being the moat?
