If you’ve been scanning the charts lately, you might notice a glaring divergence in the tech sector. While the semiconductor giants are throwing a party that doesn’t seem to end, the application software neighborhood looks like a ghost town. Some see a “bargain-hunting” opportunity; others, like the heavy hitters at Polar Capital, see an existential graveyard.

if you’re catching a falling knife, you’d better be sure it isn’t made of digital air.

For years, application software — apps, what we use for everything from payroll to document management — was the darling of the market. They had “moats.” 

Or so we thought. The reality is that the rise of sophisticated AI tools like Claude Cowork has turned those moats into shallow puddles.

Leading fund managers are sounding the alarm that traditional software firms are facing less of a dip and more of a terminal value crisis. Why pay a premium for a legacy software suite when AI coding tools have become so proficient that clients can essentially build, replicate, or customize their own internal tools at a fraction of the cost?

Where the Money is Flowing (and Why)

The smart money is rotating out of “apps” and into “plumbing.” If you look at the portfolios outperforming 99% of the market, you’ll find a distinct lack of names like Adobe or Salesforce. Instead, the focus has shifted toward the physical and digital foundations of the AI era:

• Semiconductors: The “picks and shovels” of the gold rush.
• Infrastructure Software: The “internet plumbing” (think Cloudflare or Snowflake) that provides the foundation for enterprise systems.
• Energy Infrastructure: The massive power requirements of data centers are making utility and power gear firms the surprise winners of the tech boom.

It’s not just about competition; it’s about the balance sheet. Software companies heavily rely on Stock-Based Compensation (SBC) to keep talent. When share prices tumble, that equity becomes worthless. To keep their best engineers from jumping ship to an AI startup, these firms may have to start paying out actual cash.

Which is to say: When a sector’s primary currency (its stock) devalues, and its product becomes replicable by a chatbot, the “terminal value” of that company becomes a very uncomfortable math problem.

We are likely entering a period of “digital Darwinism” reminiscent of what happened to print newspapers in the early 2000s. The internet didn’t just compete with newspapers; it rendered their entire delivery model obsolete.

While some Wall Street strategists argue for a “mean reversion” or a “relief rally” in software, the fundamental threat of AI disruption is accelerating, not slowing down. If you’re looking for safety, look toward the infrastructure that AI needs to run, rather than the applications AI is designed to replace.