The end of startups
Feb 10, 2025
Historically, starting a business meant doing something involving steel, machines, factories and a wealth of capital-intensive stuff. On the side of these behemoths, many independent entrepreneurs existed serving a niche (often geographical, like your shoemaker). In a nutshell, the entrepreneurial class was made of the industrialists and the artisans. Both complemented each other while operating at a totally different scale.
Then came the startup. Low barriers to entry, abundant capital. It became a third form of entrepreneurship to the point it almost became the main character because it was easier: you didn't need to set up an actual factory, you just needed a bunch of computers and motivated workers. With relatively low investment, you could build a very sizeable business.
Obviously, investments were still very much in need, but generally speaking, the return on capital has been tremendous and has created one of the strongest economies of all times. However, recent developments in AI changed something fundamental: the marginal cost of building software is collapsing toward zero.
What used to require a team of engineers can now be prototyped by a single person in an afternoon. I'm not going to expand on this given the tons of good articles that exist about Claude Code and the likes, but entire products are being shipped by single-owned and single-operated businesses. Non-technical founders are building functional products without writing a line of code themselves.
This has called many questions lately like why would you pay for Photoshop if an agent can generate the image you want directly? Why would you use SAP if you can spin up a script that connects to your database and extracts exactly what you need? These examples sounded insane a few months back, but it looks like the market is adjusting as I'm writing this.
In essence, utilitarian software is losing its defensibility. If anyone can replicate the functionality, what remains? Moats. Network effects, brand, regulatory capture, industrial partnerships, physical infrastructure. Even interfaces are dissolving into the chatbot. When users interact with AI directly, the application layer becomes invisible. What matters is what the AI cannot easily replicate or replace.
Of course, this means opportunities for a whole new generation of entrepreneurs. They can rebuild entire businesses from their bedrooms. This will change the world back into an economy akin to how it was before software ate the world: large companies relying on massive capital investments, and numerous smaller, often single-owned businesses thriving alongside them through specialization.
This shift will call for serious recalibration in capital allocation. If startups require building a moat or erecting high barriers to entry to be defensible, then capital must flow toward more ambitious, more expensive, long-term projects. What was a moonshot yesterday becomes a portfolio necessity tomorrow. Expect VCs to look increasingly like infrastructure investors, as most investors were in the early 20th century.
For entrepreneurs, this creates two distinct paths. The first: there's never been a better time to build a small software business. Costs have collapsed so dramatically that you don't need to fundraise. A single founder can build, launch, and maintain a product that generates a comfortable living. Most of these businesses won't scale into billion-dollar outcomes, a few of them will become really big, and all of them will make some good money.
The second path: there's never been a better time to chase genuinely ambitious projects. All the dry powder that funds have accumulated will need to shift somewhere. If software alone isn't defensible, capital will flow to ventures that combine technology with something harder to replicate: consumer, hardware, logistics, physical networks, etc.
We're not fully there yet, but the asymptotic trend is clear. I think it's a net positive for entrepreneurs, society, and the economy. Making big bucks will be harder, but it will also require more substance. In this chiaroscuro, a lot of entrepreneurs will find that building something real is more accessible than ever.