The software sector is having a truly terrible month. We're talking October 2008 levels of terrible—the kind of selloff that makes people dust off their Lehman Brothers comparisons and wonder if something fundamental has broken.

The iShares Expanded Tech-Software Sector ETF (IGV) is on pace for its steepest monthly drop since the financial crisis. And Thursday's session made things considerably worse, with the sector plummeting roughly 6% in a single day. That marked the worst trading session since March 2020, when the world was figuring out what "lockdown" meant.

The catalyst? Microsoft Corp. (MSFT) decided to have its worst day since that same pandemic panic, dropping more than 12%. The irony is that Microsoft actually beat Wall Street's expectations in its latest earnings report. But investors weren't interested in what the company achieved—they were focused on what it warned about. Specifically, slowing growth in Azure cloud services and guidance that suggested AI monetization might not be the smooth upward trajectory everyone had priced in.

The damage has spread well beyond Redmond. High-flying software names that were market darlings just weeks ago are getting hammered. Palantir Technologies Inc (PLTR), Oracle Corp (ORCL), and AppLovin Corp (APP) are all down about 20% for the month. These aren't small, obscure companies—these are names that investors had confidently bid up as AI beneficiaries.

So the question traders are asking themselves is whether this is just a technical correction—some profit-taking after a big run—or something more fundamental. And Wall Street's answer is starting to lean toward the latter.

The Existential Question: Can Software Companies Survive AI?

Analysts are increasingly suggesting that this selloff isn't just about near-term disappointment. It might be about AI fundamentally changing how software gets built, sold, and consumed. And not necessarily in a way that's great for traditional software companies.

Thomas Shipp, head of equity research at LPL Financial, put it bluntly in a Thursday note: "Can software survive AI?"

That's quite a question to ask about one of the market's most reliable growth sectors. But Shipp's logic is straightforward. Software has historically been one of the best businesses around because it's inherently scalable. You build it once, and then each additional sale costs almost nothing. That's why subscription models work so well and why investors have been willing to pay premium valuations.

"The software business has been one of the best businesses historically," Shipp said. "It is inherently scalable, as once software has been developed, there is a very low incremental cost each time it is sold."

Those premium valuations could return, Shipp noted, but only if software companies avoid getting displaced by AI. And that's the big "if."

"That said, software companies that are not a system of record do have some risk of displacement," Shipp said. "Most, if not all, software producers, will need to offer their own AI enhancements to maintain their market share going forward."

But there's another risk that Shipp flagged, one that's even more insidious. What if generalized AI tools make knowledge workers so much more productive that companies don't need to hire as many people? Fewer employees means fewer software licenses. Suddenly, the user count that justified all those recurring revenue projections starts looking like it might plateau.

The More Radical Take: Demand Itself Could Be Suppressed

If you think Shipp's concerns sound bearish, wait until you hear from Jordi Visser, head of AI Macro Nexus Research at 22V Research. His view makes the displacement argument look optimistic.

In a note to clients last week titled "Why Buying 'Cheap' Software Is the New AI Bubble Trade," Visser argued that markets are approaching a major turning point. He sees 2026 as the year when AI shifts from centralized data centers into enterprises, devices, vehicles, and machines. That migration, he believes, will broaden returns away from software and toward physical-world infrastructure.

But the really provocative part of Visser's thesis isn't about where AI goes—it's about what it does to software demand.

"When software can be generated, modified and orchestrated by agents at near-zero marginal cost, 'build' stops being a strategic project and becomes an everyday action," Visser wrote. "At that point, the entire buy-side market for software isn't disrupted — it's suppressed."

Suppressed, not disrupted. That's an important distinction. Disruption implies that someone else wins—that market share shifts from incumbents to insurgents. Suppression means the market itself shrinks because the underlying demand wasn't as real as everyone thought.

Visser compared the phenomenon to appetite suppression in patients using GLP-1 drugs. Just as those medications reduce food cravings, agentic AI could eliminate what he calls "software noise"—all the tools and platforms companies bought because building custom solutions was too expensive or time-consuming.

The implications, Visser argued, go beyond finance into philosophy. For decades, markets operated on the assumption of infinite appetite—for calories, for stimulation, for software. AI challenges that assumption by removing friction and revealing what demand actually looks like when the artificial triggers disappear.

"The Great Deflation," in Visser's framework, is the market realizing that demand for many high-value industries was inflated by the very inefficiencies they claimed to solve. Companies sold software to fix problems that AI can now eliminate entirely, making both the problem and the solution obsolete.

Visser cautioned that this unwind won't happen in a few quarters. It's a long-term trajectory shift, not an immediate collapse. But the direction, he believes, has been set.

For investors, the dividing line may no longer be between AI winners and losers. It might be between companies built for an era of unlimited appetite and those positioned for a world where "enough" is finally defined. And if that's the case, the current software selloff might be less about temporary fears and more about a permanent repricing of what these businesses are worth in an AI-native world.