I don't want to alarm anyone, but the entire SaaS industry is about to implode. Or explode. Or both, simultaneously. That’s because AI is making software creation so cheap and fast that current SaaS pricing models are starting to look like a particularly unfunny joke. Remember when people paid actual money for calculators? It's going to be like that, but for entire software suites. Let me take you to this new reality for a moment.
The rise of the solo product visionary
Large Language Models have quickly become superhuman at generating correct code for entire applications. It's like giving everyone a team of world-class developers in their pocket. For instance, GitHub's Copilot has already shown that developers using AI are 2.2x faster at completing tasks. What could possibly go wrong?
Imagine the following: a random person with good product sense and an AI sidekick decides, "You know what? I could make a better version of Salesforce." And then they just... do it. Over the weekend. This may sound crazy, but it’s almost already possible.
The YouTube-ification of software distribution
Thanks to YouTube, your neighbour's cat can become a viral sensation overnight. Well, soon, your neighbour's hastily cobbled together AI-assisted app could be the next unicorn. There will be product discovery platforms very similar to the App Store that will make or break businesses, just like Youtube’s algorithm makes or breaks content creators.
Instead of star ratings and likes, we'll have publicly available user numbers and retention statistics. But many of the most popular apps will be fully open source and remixable, so you will be able to not only tell if an app is good before you download it, but also tweak it yourself to match your preferences.
That also means that the era of everyone accepting that most software is bad is over soon, too. We're entering a different kind of era, where most software, from recipe apps to sophisticated crypto indexes, might have a user experience that’s actually good.
I firmly believe that those with the best ideas will win this competition - those with the right vision to use technology effectively. Product taste and domain expertise will become more crucial than ever.
The other impact of this is that developers will need to pick up new skills, or stay in niches where a human touch will stay important.
What about SaaS companies?
You might be thinking, "This all sounds great for consumers, but what about the poor SaaS companies?". Let me be blunt: in 2025, SaaS companies without a clear moat will start noticing churn to fresh competition and to internal tools built with AI to serve the same need. It’s already starting: Klarna recently announced it is replacing Salesforce with internally built AI. It’ll be like watching a slow-motion train wreck, except the train is made of bits and bytes, and the wreck is your business model.
This shift won't hit everyone equally. Regulated industries will naturally be less impacted. Segments like designers and developers, who are always ready to jump on the best value proposition, will be the canaries in the coal mine, signalling the changes to come.
My advice to current SaaS companies is therefore to adopt AI in development to speed up their iteration process, keep zooming into verticals where domain expertise is key, and not to be afraid to disrupt and cannibalise themselves before someone else does.Regardless of how well they prepare though, the shift I’m talking about will go faster and faster, to the point where the willingness to pay for software will approach that of paying for watching a tiktok or youtube video.
For those of us in tech, this shift demands relentless adaptation. Ideation and execution are collapsing into a single, AI-accelerated process. We all have to adapt, quickly. The bottom line is that the future belongs to those who can dream up the next big thing and then sweet-talk an AI into building it for them.
