The Software Stock Crash Is Spreading and Nobody Wants to Say Contagion Out Loud

February 15, 2026

I got in at 7:30 this morning. My dad was already here. I don't know how he does that. I made coffee, pulled up the markets, and sat there for about ten minutes reading about the software selloff before I even took my jacket off. That's where my head is right now. That's where this article is coming from.

Here's what's happening, and I'll say it plainly because apparently nobody else wants to: the software stock crash has turned into a contagion, and the financial press is doing everything it can to avoid using that word out loud. They'll use "spillover." They'll say "spreading concerns." Mizuho strategists are already muttering about "correlation creep." But nobody wants to say the actual word because once you say it, the conversation gets a lot harder to manage.

So I'll say it. This is contagion. And if you run a small or mid-sized business that pays monthly for a stack of SaaS tools - project management, CRM, email, payroll, whatever - you need to understand what's happening at the top of the food chain, because it filters down. It always does.

What Actually Happened

We've written about the SaaS valuation problem before. But this has escalated fast. The IGV ETF, which tracks North American software stocks, peaked on September 19, 2025. Since then, it's fallen 30%. By early February, it was already down almost 20% year-to-date alone. For context - and this is the part that should make people sit up - the semiconductor ETF over that same period is up 30%. The broad tech indexes are flat. This isn't a tech-wide rout. This is specifically, surgically software.

The numbers got worse fast. The S&P North American software index dropped 15% in January alone - its biggest monthly decline since October 2008. January 29th was the single worst day for software stocks since the COVID crash. ServiceNow dropped 11% despite beating earnings for the ninth straight quarter. Microsoft shed $360 billion in market cap in a single session. Salesforce is down roughly 30% year-to-date and has cut nearly 1,000 roles in a fresh round of layoffs - while simultaneously pushing its AI agent platform Agentforce as the future. That's the whole tension in one company.

Then Anthropic released a productivity tool for in-house lawyers and the thing went sideways almost immediately. Legal software stocks got absolutely hammered - Thomson Reuters fell 16%, LegalZoom plummeted 20%, CS Disco sank 12%. This wasn't legal software having a bad earnings day. This was fear spreading sector to sector, category to category, like a fire that jumps the road.

That's contagion. And it's not stopping at software stocks.

Technical blueprint illustration showing three architectural platform structures in dark void space, the first fully consumed by precisely rendered flames while fire leaps mid-arc toward the second platform, depicting financial contagion spreading between market sectors — Showed this to Tory before I published and he said 'that's dramatic' which, yes, that is the point - I went through four different compositional versions before settling on three platforms instead of two because the third untouched one in the background does more work than any caption I could write.

The Private Credit Problem Nobody Is Talking About Loudly Enough

Here's the part of this story that I think is genuinely underreported, and it matters beyond Wall Street trader sentiment. The $3 trillion private credit market has software as its single largest sector exposure - roughly 20-25% of all private credit deals are SaaS companies. UBS puts the AI-disruption-exposed share even higher, at 25-35%. This is the shadow banking system that financed a decade of software buyouts. And it is now sitting on a lot of loans that are worth a lot less than when they were written.

From 2015 to 2025, more than 1,900 software companies were acquired by private equity in deals worth over $440 billion. The thesis made sense at the time: sticky recurring revenue, high margins, predictable cash flows, high switching costs. Private credit loved lending against those characteristics. Then AI showed up and started stress-testing every single one of those assumptions simultaneously.

The result: more than $17.7 billion in US tech company loans dropped to distressed trading levels in just four weeks - the most since October 2022. Software now holds a 13% market share in the leveraged loan index, and 13% of those loans are marked at distressed levels. Apollo cut its direct lending funds' software exposure almost by half in 2025. Hayfin and Arcmont are paying consultants to comb through their portfolios looking for exposure. These are not small moves. These are people quietly preparing for something they're not saying out loud in public.

Tory walked by my desk this morning while I was in the middle of reading all of this. He's going through a lot right now - his car situation, his marriage - but somehow he's the most relentlessly upbeat person in the office. He glanced at my screen and said "big shakeout, that's just the market cleaning house." I don't know if that's optimism or avoidance. Maybe both. Maybe both at once is how humans survive bad news.

The Argument That This Is Overblown (And Why I Don't Fully Buy It)

To be fair - and I'm trying to be fair here even though fairness isn't the most interesting thing to write - there are serious people making the case that the market is overcorrecting. Morgan Stanley's CIO James Caron has said the chance of real contagion from the software selloff is relatively low. Bank of America called it an "indiscriminate selloff" and compared it to the DeepSeek panic in January 2025, which ultimately proved to be overblown.

BofA's argument is actually pretty sharp: the SaaS selloff is relying on two mutually exclusive scenarios at once - AI investment showing such weak ROI that growth is unsustainable, while simultaneously AI is so effective it makes entire software business models obsolete. Both of those can't be true at the same time. Morningstar made a similar point: slower sales growth might just mean buyers are delaying purchases, taking a wait-and-see approach on AI rather than actually abandoning software.

And look, average gross retention for SaaS companies is still around 90% or better. Installed bases have largely held up. The earnings for most software companies aren't catastrophic. Thomson Reuters, catching stray bullets from the selloff, still guided to 8% organic revenue growth in 2026 with margin improvement. The fundamentals are not, in most cases, as bad as the stock prices suggest.

But here's where I part ways with the "this is overblown" crowd: they're making a today argument about a tomorrow problem. The market isn't trading current earnings. It's trading the world three years from now. And the question the market is asking - in 2029, in a world full of AI agents running workflows, which software is still worth paying an annual subscription for? - is a completely legitimate question. The fact that we don't have a clean answer yet is exactly why stocks are going wherever fear points them.

The Seat Problem Is Real and It's Not Going Away

There's a specific mechanism here that I keep coming back to because it's quietly the most important thing happening. It's not that AI replaces the software directly. It's that AI reduces the number of humans who need to use the software. If ten AI agents can do the work of a hundred sales reps, you don't need a hundred Salesforce seats. You need ten. That's a 90% reduction in seat revenue for the same work output. The software doesn't disappear. The business model does.

This is why Bain's research is showing that some vendors are already reporting slower growth in seat count as customer companies become more efficient. Not churning. Not switching. Just... needing fewer seats. This is the slow, quiet version of getting replaced and it's harder to see coming and harder to fight when it arrives.

Strip out the price increases that have been propping up SaaS earnings. What's left? Net new customer numbers are weak across the board. Expansion is slowing. The "growth" in a lot of recent SaaS earnings is vendors raising prices on captive customers. That works until it doesn't. It doesn't work when AI gives those captive customers real alternatives for the first time.

Derek came over at lunch and started explaining some plot point from one of the Disney Star Wars shows - I can never keep track of which ones he likes and which ones he hates, it changes - and I nodded along while continuing to think about seat-based pricing. I feel like that's rude but also it was a lot to process.

What This Means If You Actually Run a Business

I'm not a financial analyst and I'm not telling anyone what to do with their portfolio. But I do think there are real implications here for people operating businesses, and I want to say them directly instead of hedging them into mush.

First: your vendors are under pressure. Real, structural, existential pressure that is not going away in a quarter. We've talked about what it feels like when your vendor has a bad earnings call - multiply that energy by a sector-wide meltdown. Expect acquisitions. Expect pivots. Expect pricing changes dressed up as "value alignment." The companies that get through this will be the ones that genuinely integrate AI into their product, not the ones that bolted a chatbot onto a 2019 dashboard and called it AI-powered.

Second: the software you're currently paying for is in a period of maximum instability. That doesn't mean cancel everything and rebuild from scratch. Replatforming is its own disaster - I've seen it up close. But it does mean this is a good moment to actually audit what you're paying for, what you're using, and whether the company behind it has a credible plan for the next three years that isn't just "we're adding AI features."

Third: seat-based pricing conversations are about to get much more interesting. If you're renewing anything in the next six months, you have more leverage than you think. Vendors know their seat counts are vulnerable. That's a negotiation.

Linda stopped by my desk around 3pm and said I looked stressed. I told her I was working through something. She said Gerald had been stressed last week too and that it helped him to go for a walk. Then she mentioned Gerald went to the same walk three days in a row. I think she was telling me to take a walk. I didn't. But I thought about it.

The Word Nobody Wants to Say

I keep coming back to the word itself. Contagion. The reason people don't want to say it isn't because they don't see what's happening. It's because saying it out loud changes what it is. It makes it real. "Spreading concerns" is a weather report. Contagion is a diagnosis.

Software stock valuations were already under pressure going into this year. The private credit exposure makes this something other than just a stock market story. When the debt side of an industry starts moving toward distressed levels - when loans against software companies are trading down, when private equity firms are paying consultants to audit their exposure, when Apollo cuts its software lending by half - you are no longer talking about investor sentiment. You're talking about the plumbing.

The optimists are right that the fundamentals haven't collapsed yet. Customer retention is holding. Revenue is still growing at most of these companies. The market probably is pricing in some scenarios that won't fully materialize. But public SaaS growth rates have declined every single quarter since the 2021 peak - every single one - and the AI crash narrative just gave the market permission to finally reprice what three years of earnings data have been quietly screaming. That's not irrational panic. That's overdue recognition.

I rewrote the last section of this four times. My dad hasn't read it. I think what I'm trying to say is this: the software industry is going through something real and structural, not just a bad month. The contagion label fits because the mechanism is contagion - fear, debt exposure, and business model risk spreading from category to category faster than anyone has a response for. People who run businesses on software stacks should be watching this closely. Not because the sky is falling, but because the ground is shifting. And the companies paying attention now are going to have better options than the ones who look up in a year and wonder what happened.

My dad got in at 7:15. I genuinely don't know how he does it.