Market Pulse – Will We Even Need External Developers Five Years From Now?
Why AI makes the moat of strong software makers and service providers bigger, not smaller
Few questions come up more often in conversations with software entrepreneurs, investors, and clients than this one: will we even need external developers five years from now?
The concern behind it is real. If AI can already write code, review contracts, and automate tasks today, it's natural to assume that companies will simply build software themselves going forward.
The catchphrase for it was quickly found: "SaaSpocalypse." The debate was triggered, among other things, by Anthropic's announcement that AI will be able to review simple contracts. In February, tech stocks lost roughly USD 285 billion in value within 24 hours as a result. The thesis: AI makes specialized software providers obsolete.
But this pattern isn't new, and historically it's almost always the same: a new technology is first understood as a replacement before it proves to be an amplifier of professional capability.
In the 1980s, video cameras were supposed to displace cinema. Instead, they carved out a niche of their own. Production costs for theatrical films actually tripled between 1980 and 2007.
Desktop publishing was supposed to wipe out the printing trade. Instead, the number of print shops hit an all-time high of 62,000 in 1995. Individual roles disappeared, the industry did not.
This same logic now applies to the software world.
Because AI can generate code. But building software is something else. What customers can't simply replicate at the push of a button are robust engineering processes, security, code reviews, maintainability, and governance.
Or, as technology historian Jim Cortada puts it: expecting a customer to build a critical system like payroll themselves using AI is "like changing a tire on a car that's still going 60 miles an hour. Too risky."
The real moat, then, doesn't lie in the individual code snippet, but in the professional systems capability behind it. And this is exactly where AI makes the difference bigger rather than smaller.
The economics point the same way. Nobel laureate Oliver Hart argues that AI mainly lowers adaptation costs on the provider side. Custom software from professionals thus becomes cheaper and more attractive. Building in-house doesn't necessarily become more likely, in many cases it simply becomes less sensible.
For software makers and service providers, the message is therefore clear: AI is not the gravedigger of the business model. AI is the accelerator for those who already have mature delivery, engineering, and governance structures in place.
And that level of maturity is becoming increasingly relevant in the M&A context as well. Buyers no longer assess only product and revenue, but also how robust a company is technologically and operationally. Anyone who integrates AI sensibly into proven processes doesn't just build better software. They also increase the strategic appeal and the value of their own company.
The bottom line
"So the question isn't whether AI replaces external developers. The more important question is which providers translate AI into scalable engineering quality faster and more professionally than others. That's precisely where the winners will emergeover the coming years, both operationally and on the transaction side," Pascal Kopp says with conviction.
