The SaaSpocalypse: How AI Wiped $2 Trillion from Software and Why Finance Is Next

Published February 11, 2026

Anthropic released some free plugins. Open source templates, basically. Workflow stuff for legal, finance, marketing.

Two trillion dollars in market cap vanished.

Not because the plugins were revolutionary. Not because Anthropic built a Salesforce killer. Because Wall Street finally did the math that everyone in tech has been doing on napkins for two years: most enterprise SaaS is a fancy UI on top of a database, and AI can be the UI now.

LPL Financial dropped 8.3% today. Raymond James fell 8.75%. Charles Schwab lost 7.4%. The contagion jumped from software to finance because some startup called Altruist built an AI tax planning tool and suddenly everyone remembered that financial advisors also charge per-seat for what is essentially data lookup with a compliance wrapper.

Wall Street has a name for this. They’re calling it the SaaSpocalypse. Which is dramatic, but the numbers back it up.

What Anthropic Actually Released

Let’s be precise because the panic outpaced the facts.

Claude Cowork launched on January 12 as an agentic tool for non-technical users. On January 30, Anthropic added plugins — 11 open-source templates covering productivity, enterprise search, marketing, sales, finance, legal, customer support, and biology research.

The plugins let enterprises package workflows, tools, and integrations into role-based AI apps. Users can tell Claude how they like work done, which tools and data to pull from, and what commands to expose. No technical expertise required.

That’s it. That’s what erased $2 trillion.

Not a product that replaces Salesforce. Not a platform that eliminates ServiceNow. A set of open-source templates that showed the market what was possible.

The Numbers Are Brutal

The initial selloff hit on February 3-5:

Thomson Reuters: Down 15%+
LegalZoom: Down ~20%
ServiceNow: Down 7.4% on Feb 3, another 6.3% on Feb 5. Down 45% over the past year.
Salesforce: Down 7-11%. Down 40% trailing twelve months.
Adobe: Down 7-11%. Down 35% trailing twelve months.
Workday: Down 7-11%. Down 25% year-to-date.
Monday.com: Lost $300 million in market cap within 30 minutes.

That last one is worth pausing on. CNBC anchor Deirdre Bosa tweeted that she’d recreated Monday.com’s project management interface using Claude. “I literally have my own monday.com… plugged into my calendar & gmail.” The stock started falling within minutes.

A single tweet. Thirty minutes. $300 million gone.

Goldman Sachs reported the US software stock basket fell 6% in a single session — the steepest loss since the April tariff-related decline. Jefferies’ equity trading desk coined it: “We call it the SaaSpocalypse, an apocalypse for software-as-a-service stocks.”

Private equity took collateral damage too. Ares Management, KKR, and TPG all lost over 10%. Apollo and Blackstone dropped 8%. Blue Owl Capital fell 13%.

Wall Street Figured It Out Last

For years, the Wall Street consensus was simple: AI helps software companies. AI is a tailwind. Companies that integrate AI charge more per seat. Growth accelerates. Everyone wins.

Anyone who has actually used these products knew this was wrong. The entire SaaS model is predicated on the idea that software is hard to build, so you rent access to someone else’s. AI makes software trivially easy to build. A CNBC anchor recreated Monday.com in an afternoon. Why would you keep paying $36/seat/month after that?

Anthropic’s plugins didn’t kill SaaS. They just showed the market what the market didn’t want to see.

Morgan Stanley now says the era of “easy growth” for SaaS is “effectively over.” Piper Sandler downgraded Adobe and slashed its price target by $140. The seat-based pricing model — the entire economic engine of modern software — turns out to be a vulnerability, not a competitive advantage. Who could have predicted that charging people monthly for something an AI can replicate for free would eventually be a problem. Besides everyone.

Microsoft: The Canary in the Coal Mine

If you want to understand how deep this goes, look at Microsoft.

Ben Reitzes at Melius Research downgraded Microsoft to Hold on February 9, with one of the most bearish price targets on Wall Street at $430. His reasoning was devastating.

Microsoft has been promoting Copilot for three years. After all that investment, all that integration, all that enterprise sales effort — Copilot has 15 million paid users. Reitzes said he was “floored.”

For context, Anthropic developed Cowork in 10 days. Most users think it works better with Excel and other plugins than Copilot does.

Reitzes warned that Microsoft may be forced to bundle Copilot as a free component to stay competitive, which would erode revenue and profits in its core productivity division. His summary: “AI is eating Microsoft.”

Microsoft’s P/E ratio has dropped to 23. That’s cheaper than IBM.

The Layoffs Confirm It

If you still think this is just a stock market overreaction, look at what the companies themselves are doing.

Salesforce cut nearly 1,000 employees in early February, spanning marketing, product management, and data analytics. CEO Carl Eschenbach stepped down on February 10. And this follows Marc Benioff’s revelation that AI had already cut Salesforce’s own customer support team from 9,000 to roughly 5,000. His exact words: “I’ve reduced it from 9,000 heads to about 5,000, because I need less heads.”

Read that again. Salesforce — the company that pioneered the SaaS model — used AI to eliminate 4,000 of its own support jobs. AI agents handled 1.5 million customer inquiries in nine months.

Workday cut 400 employees on February 4, primarily in customer support. Amazon announced 16,000 corporate cuts on January 29. Pinterest cut 15% of its workforce and was, quote, “the most explicit in asserting that AI drove” the layoffs.

In the first six weeks of 2026, 82 tech layoff events impacted 35,105 people. That’s 856 people per day.

The Financial Services Domino

Which brings us to today.

Altruist, a wealth platform for independent financial advisors, launched Hazel AI — a tool that reads tax returns, payslips, account statements, and other documents to generate personalized tax strategies. Automated. In minutes.

The financial services sector reaction was immediate. In the US: LPL Financial down 8.3%, Raymond James down 8.75%, Charles Schwab down 7.4%. In Europe: St. James’s Place down 11.25%, AJ Bell down 5.85%, Quilter down 5.3%.

The pattern is now clear. Software wasn’t unique. Any industry where humans process information according to structured rules is in the blast radius. Legal got hit first. Then project management, CRM, HR tech. Now financial advisory. Who’s next? Accounting? Insurance underwriting? Medical billing?

The Bull Case (And Why It’s Partially Right)

Not everyone is panicking.

JPMorgan’s Toby Ogg met more than 50 investors across Europe and the US and found that the sector “isn’t just guilty until proven innocent but is now being sentenced before trial.” JPMorgan argues the fears are overdone and sees a potential rebound in 3-6 months.

Bank of America’s Vivek Arya identified what he called a fundamental paradox in the market’s logic: You can’t simultaneously believe that AI capex will deteriorate because the ROI isn’t there, AND that AI adoption will be so pervasive it makes software companies obsolete. “Both outcomes cannot occur at once.”

He’s right. That is internally inconsistent. But markets aren’t logic engines — they’re fear engines. And right now, the fear is that AI agents don’t just help software companies; they replace software companies.

The $660 Billion Question

Here’s the other side of the equation that’s getting buried under the SaaS panic.

The four major hyperscalers — Amazon, Alphabet, Microsoft, and Meta — are set to spend roughly $650-700 billion on infrastructure in 2026. That’s a 67-74% increase from 2025. About 75% of it, roughly $450 billion, is specifically for AI.

The breakdown:

Company	2026 Capex	% of Revenue
Amazon	~$200B	~25%
Alphabet	~$180B	~46%
Microsoft	~$145B	~47%
Meta	~$125B	~54%

Meta is spending 54% of its revenue on infrastructure. Amazon is looking at potentially negative free cash flow to the tune of $17-28 billion.

These companies are making the biggest infrastructure bet in corporate history. They wouldn’t be doing this if they didn’t believe the disruption was real.

What This Actually Means

Here’s my take, cutting through both the panic and the dismissals.

The SaaS model is dying. The software isn’t. The per-seat subscription was never about the software’s value. It was a tax on headcount. AI doesn’t just compete with SaaS products — it makes the pricing model obsolete. Why pay per-seat when an agent does the work of five seats?

But the vendors won’t disappear. They’ll transform. Salesforce won’t go bankrupt. They’ll rebrand as an “agent platform” and charge per-agent instead of per-seat. Same landlords, new lease terms. ServiceNow will become the workflow engine underneath AI agents rather than the interface humans interact with.

The real question isn’t which SaaS companies survive. It’s who controls the agents. If your AI agent runs on someone else’s platform, reads someone else’s plugins, and stores data in someone else’s cloud — you’ve just traded one subscription dependency for another. The revolution only matters if individuals and organizations can run their own agents on their own infrastructure.

That’s happening right now in the open-source community while everyone’s distracted by stock tickers. Local models are getting good enough. Agent frameworks are proliferating. The tools for self-sovereign AI are being built in the open.

The $2 trillion wasn’t destroyed. It’s being redistributed. From middlemen who package software into seats, to infrastructure providers who supply the compute, to open-source communities that make the whole stack accessible, and eventually to individuals and small teams who can now do what used to require a 50-person department.

That’s not a crash. That’s a transfer of power. And if you’re paying attention, it’s the most exciting thing happening in technology right now.

This is Kyber Intel. We track the shift from corporate gatekeeping to individual sovereignty in AI. Follow us on X @kyberintel for daily analysis.