Software no longer has value
What does it mean for tech valuations?
April 20, 2026
At the dawn of the mobile revolution building "an app" was enough to raise fund raising pre-revenue and even sell a company. In 2026 anybody can build an app without any technology background at all. You no longer need a wonder dev, or a team of engineers to build an app, and even to build substantially complex software.
Sure a skilled hand (or hands) is still required for substantially complex technology problems, but the reality is most companies building software don't have substantial complex problems. The vast majority are building CRUD apps (Create, Read, Update, and Delete data), simple web / mobile interfaces to their own business logic. AI is really good at building and maintaining these. If software is easy/quick to write, it is trivial for a competitor to replicate it. First mover has rarely meant much, but it means even less when a kid with a Claude Max account can dupe your app in a weekend.
Companies can no longer rely on "having written software" as a source of value outside of extremely technical problems*. If anything it makes having good technology table stakes and not optional. We saw a blitz of companies in 2024 get funded / sold based on the fact that they had built some harness / interface to OpenAI only to get absolutely obliterated in 2025 as their "agentic" tool was suddenly a few prompts away from being built. So beyond software how do you value a company?
The most obvious is business fundamentals. People may be scratching their heads wondering why I am even talking about anything else, but in many of the boom/bust startup waves we've seen over the past 25 years things like revenue and margin have at times not mattered at all. That definitely changed in 2025 and it continues in 2026. VCs actually care that you can make money delivering the product/service you are offering. Sure there are still big risky growth bets, but not the way we saw in the 5 years prior. Investors want to see a current business that works.
Beyond that, trust/brand is now even more valuable. We've already seen the 2025 absolute boom of fully AI built apps has lead to consumers ignoring the vast majority of them. People are weary of trusting their information to unknown apps and companies. So a company with a brand / following and consumer trust has even more value it can leverage now that the cost to build tech is so cheap.
Data has immense value, even more than before. It is easier than ever to leverage data with AI doing the heavy lifting. I recall building a model for billing code prediction by hand in 2023, a little CTO gets to pick their favorite-project-project. Writing torch code, testing neural network design, iterating on my dataset to get better and better scores. In 2026 its a Claude prompt or two away. Data also helps Claude/Codex/Antigravity to build better apps faster: it's context for solutioning, it's test data to allow for simulations before a customer launch, it's food for RAG chat bots. The big frontier AI labs are still in a race for data that doesn't exist in books they can buy or scrape on the open web, so if you have it, use it!
Is expertise a value? I think it still is. Acqui-hires still make sense, especially for AI talent. There is still secret sauce. There are still things that even the largest multi trillion parameter models don't understand. But every person, leader, and executive should ask themselves what about their job can AI not do today, and repeat this question every 3-4 months. It is changing rapidly.
I guess I am a little biased but I do think we will see a rise in AI powered services business, healthcare services, IT services, you name it. Suddenly you can use AI as a massive lever to improve margins and scale (not replace!) humans. Can we deliver even better health care with better margins? Absolutely, and we are. It is hard to know what the market will favor (beyond frontier AI companies and AI ecosystem plays) in the coming years, but it is absolutely clear AI will be required to enable any of it.
* Extremely technical problems like mega scale apps or building frontier AI models. The cost of training a frontier AI model, the amount of theoretical research (even if it is extremely AI assisted), etc is all still tough to replicate, so it makes sense that these companies still have "technology" value. But rest assured these will be a few prompts away one day.