Verdict Before the Trial: Why Investors Are Burying Software Over AI

The software market now looks like a scene from a disaster movie, where everyone runs for the exits without picking a path. Investors have staged a mass sell-off of tech stocks that only yesterday were considered a "safe haven" and a benchmark of steady growth. Analysts at JPMorgan Chase & Co. found an apt metaphor for what is happening: the software industry has been sentenced to the harshest penalty before the trial has even begun. It is a rare case where the market does not wait for proof of guilt or a drop in revenue, and instead acts preemptively, driven by pure existential fear of artificial intelligence.

To grasp the scale of the drama, it helps to recall what the last decade of prosperity was built on. The SaaS (software as a service) model thrived on selling licenses for every workstation. The more employees a company has, the more money Salesforce, Adobe or Workday earn. But generative AI now calls that very math into question. If one person using Claude or GPT-4 can do the work of five, companies will inevitably start cutting the number of paid accounts. The market is already pricing this scenario into stock quotes without waiting for quarterly reports. Investors believe AI will not merely complement software but replace it entirely, turning complex interfaces into a simple chat window.

The panic exposes a deep split in the tech sector. On one side are the hardware giants like Nvidia, which sell shovels during a gold rush and are doing very well. On the other are application developers, who suddenly find themselves in the position of carriage makers in the era of the first automobiles. JPMorgan notes that the gloomy mood on the exchange has become chronic. Any news of progress in LLMs is now read as another nail in the coffin of traditional software. Even companies that are actively building neural networks into their products get no credit from Wall Street.

Why is this happening now? Because the effect of AI has stopped being theoretical. We are seeing the first signs of corporations reallocating budgets. Money that used to go to enterprise software now goes to renting cloud capacity to train in-house models or to paying for APIs. It is a zero-sum game: to feed a "smart" AI, you have to put "dumb" software on a diet. Investors see revenue growth slowing at the cloud giants and draw a logical, if brutal, conclusion: the old model is broken.

Still, the "verdict before the trial" stance carries a measure of irony. Markets tend to overshoot, and JPMorgan hints that the current sell-off may be excessive. Software companies are not sitting idle; they are moving from selling "tools" to selling "outcomes." Instead of charging for access to buttons, they are starting to deploy AI agents that carry out specific tasks. But until this transition is confirmed by hard numbers in financial reports, skepticism will dominate. Wall Street wants to see not promises of AI integration but real protection against cannibalization of their own business.

The situation is complicated by the collapse of barriers to entry in software development. Where building a rival to, say, an accounting service once took years and dozens of programmers, AI coding now lets teams do it many times faster and cheaper. This creates a risk of decommoditization in reverse: software becomes too easy to produce, and therefore less valuable. Investors fear we are entering an era of "free" or ultra-cheap software, where margins fall to zero.

The bottom line: the market has returned a verdict on software of "guilty of inefficiency." Can companies appeal by deploying AI agents, or are we really watching the sunset of the SaaS era?