So I’ve been getting a lot of calls lately. Not the usual kind — not “hey Danny, my carrier just changed the comp plan again” or “can you believe what they did at the QBR.” These are different. These are the calls where someone says “so, should I be worried?” and then kind of laughs nervously, like they already know the answer.

The topic is SaaS. Specifically, the thing people are calling the SaaSpocalypse. And before you roll your eyes at the dramatic name — $300 billion in market value evaporated in a single week in February. That’s not hype. That’s a number.

Let me tell you what actually happened, what it means for MSPs who’ve built practices on SaaS resale and management, and what I think the real opportunity looks like from where I’m standing.

The $300 Billion Week

In early February, Anthropic released a set of agentic AI plugins capable of automating complex workflows across legal, cybersecurity, and sales functions. (For more on Anthropic’s channel implications, see our coverage of the Claude partner network.) Not better chatbots. Actual agents that could execute multi-step tasks, call APIs, update records, and hand off to humans when they needed to.

Investors looked at per-seat SaaS subscription economics and panicked. If an AI agent can do the job of three seats, you don’t need three seats. More than $300 billion in SaaS market value disappeared as people tried to figure out what the per-seat subscription model was worth in a world where the seat might not need a human in it.

Now, is SaaS dead? No. Of course not. Installed bases don’t evaporate because a new paradigm shows up. The companies holding ServiceNow licenses and Salesforce seats and Microsoft 365 deployments didn’t all uninstall everything in February. Life is messier than that.

But here’s what did change: the conversation with your customers changed. Because your customers read the same headlines you did. And some of them are now asking their IT partners — which is you — “should we be rethinking our SaaS stack?”

That question is either an opportunity or a threat, depending entirely on what you say next.

What’s Actually Happening on the Ground

I talked to a handful of MSP owners over the past couple weeks. The pattern is consistent.

In the SMB segment, almost nothing has changed operationally. Microsoft 365 is still Microsoft 365. Customers are still paying for seats. Nobody’s ripping out their collaboration stack because agentic AI exists. The SaaSpocalypse is mostly a financial press story at the SMB level.

In the mid-market, it’s more interesting. You’ve got ops leaders who’ve been reading about AI agents for six months and are now asking real questions about whether they need as many seats for specific functions. Legal workflows. Marketing automation. Basic customer service routing. These are the areas where agentic tools are genuinely competitive with per-seat SaaS today — not theoretically, but actually.

And the enterprise? That’s where it’s real and it’s already happening. ServiceNow has been announcing AI agent investments week after week. AWS Marketplace search activity for agentic AI tools has increased significantly over the past year. The enterprise isn’t waiting for the market to settle. They’re already deploying.

The MSP Business Model Question

Here’s the part nobody wants to say out loud: if your MSP business is built primarily on SaaS resale margin and per-seat management fees, you have a math problem.

Not a crisis. A math problem. Which means it has a solution, but you have to run the numbers honestly.

SaaS resale margin has been getting compressed for years. Microsoft and other vendors have been pushing direct and reducing partner margins on pure licensing. The MSP profitability paradox is only getting sharper. AI agents threatening per-seat volumes just accelerates a trend that was already underway.

WatchGuard’s takeaways from the CRN MSP 500 put it plainly: MSPs are moving toward vendors that help them reduce operational complexity by delivering integrated platforms with automated operations. They’re not looking for more point products. They’re looking for fewer, better things that help them run a scalable services business.

That’s the signal. The MSPs who are thriving aren’t the ones with the longest SaaS license catalog. They’re the ones who’ve figured out how to layer AI-assisted management on top of a tighter, better-integrated stack — and charge for the outcome, not the seat count. The build vs. buy vs. wait decision on AI in PSA and RMM is the practical version of this question.

What This Actually Looks Like in Practice

I’ll give you the version I’ve seen working.

An MSP in the mid-Atlantic runs about 40 clients, mostly 50-250 employees. They went through their SaaS stack with every client last quarter — not to cut tools, but to identify which workflows were candidates for agentic automation. They weren’t selling “AI.” They were selling a service rationalization engagement.

What they found: across their client base, there were three to four categories where agentic automation could genuinely replace seats in the next 12 months. Basic customer intake. Internal IT ticketing. First-pass document review. They built an automation assessment into their annual QBR. They’re now charging for that assessment as a service line.

The clients love it because they’re being told honestly where they can save money. The MSP loves it because they’ve gone from being the person who bills for seats to being the person who manages the strategy layer above the seats. That’s harder to commoditize.

The Thing Nobody Wants to Hear

If you’ve been treating your SaaS portfolio as a passive recurring revenue line — add a seat, get a margin, repeat — the SaaSpocalypse is a wake-up call dressed in a dramatic name. If you want a grounded look at how the bifurcation of the MSP market plays out when some operators adapt and others don’t, that’s the math you need to run.

If you’re still running the same strategy from five years ago, the 2019 playbook is already costing you money. Your customers are going to start asking about AI agent deployments. Some of them are going to figure out that certain workflows don’t need per-seat licenses anymore. They’re going to want someone to help them navigate that. And whoever they trust to have that conversation is going to own the relationship going forward.

The MSPs who get in front of it will come out ahead. Not by abandoning SaaS — the installed base is real and it’s not going anywhere fast. But by adding a layer of strategic value above the license catalog that makes you indispensable when the conversation about automation shows up.

It’s going to show up. It already is.

Call me when you want to compare notes. You know where to find me.