Andreessen Horowitz doesn’t invest in managed services. That’s been their stated position for over a decade. MSPs are labor-intensive, margin-thin, and built around human headcount. Not the SaaS multiples that make a VC firm famous.

So when a16z led a $25 million Series A into Treeline, an AI-native MSP launched in 2024, it said the quiet part out loud: the economics of traditional managed services just changed, and the firm is betting against 40,000 incumbents.

That’s not a dig at you personally. But you should take it that way.

The Thesis Is Simple and Brutal

a16z partner Joe Schmidt described the MSP market as hiding in plain sight for Silicon Valley. Over $200 billion in annual revenue. More than 80% of US businesses outsourcing their IT. An industry that, from the outside, looks like a 2003 business model running on a 2026 technology stack.

Treeline CEO Peter Doyle put it plainly: most MSPs are “reacting to tickets and providing labor augmentation.” That’s it. That’s the entire model. A human gets a ticket. A human resolves the ticket. You scale by hiring more humans.

Treeline’s bet is that agentic AI breaks that equation. Their platform claims to handle 98% of customer-submitted requests automatically, cuts employee onboarding time from 20 minutes to two, and reduces error rates by 95% for tickets resolved inside their system. The company has about 200 customers, 70 employees, is profitable, and expects at least 3x revenue growth this year.

Those numbers — if accurate — describe a fundamentally different cost structure than anything in the traditional MSP playbook.

This Is the “Dead Playbook” Moment

I’ve written before about the death of the generalist VAR. The logic there was straightforward: if you can’t define what you’re uniquely good at, you’re losing on price to someone with lower overhead. Treeline is the next chapter of that argument.

The headcount-per-revenue ratio has always been the ceiling on MSP profitability. Your margins are tied directly to how many technicians you need to keep the lights on for customers. Stack vendors haven’t solved this — ConnectWise, Kaseya, Datto, NinjaRMM, all of them made operators more efficient but still fundamentally labor-driven.

Treeline isn’t trying to make labor more efficient. They’re building a model where labor is mostly optional for tier-1 and tier-2 work. Technicians stay in the loop for architecture, security escalations, and complex compliance work. Everything else gets automated at a quality level they claim beats the typical human ticket resolution.

If that’s true at scale — and that’s a real if — the cost structure of the AI-native MSP looks nothing like yours. They’re not pricing against your margin. They’re pricing against your cost basis, which they’ve eliminated.

The 40,000 Number

There are roughly 40,000 MSPs in the United States. Doyle used that number in his CRN interview, and it wasn’t accidental. That’s the competitive landscape he’s describing. a16z isn’t investing in a niche product for a specific vertical. They’re investing in a model that, if it works, invalidates the majority of what those 40,000 businesses are built on.

Most of those 40,000 are small operations — under $5M in revenue, running 3-10 technicians, serving local SMBs that are largely price-sensitive. Those are exactly the customers Treeline is targeting. Not the 500-seat enterprise. The local dentist office with 15 endpoints. The 40-person law firm. The regional manufacturer still on-prem.

That’s where the volume lives. And that’s where an AI-native MSP with lower overhead wins on price before it even gets to the conversation about quality.

What Changes If They’re Right

Here’s the scenario worth planning against. Treeline (or something like Treeline — there will be competitors now that a16z is in the space) captures meaningful SMB share over the next 24-36 months. They price below the traditional MSP. Their AI handles the routine stuff. Their technicians focus on strategic work. Customer satisfaction holds.

What happens to the MSP that built their business around monthly recurring contracts with local SMBs?

You already know. You’ve seen this movie in other sectors. The efficient model wins on price, customers move, and the incumbents who didn’t adapt find themselves competing for fewer accounts at compressed margins.

The antidote isn’t panic. It’s specificity. The MSPs who survive this are the ones who’ve built expertise that agentic AI genuinely can’t replicate — deep vertical knowledge in healthcare or legal, complex security practices, strategic advisory relationships with customers who aren’t buying IT management, they’re buying a trusted partner. That’s defensible. A ticket-and-invoice operation is not.

The Uncomfortable Question

Connectwise’s acquisition of Zofiq earlier this year signaled something similar from the platform side. PSA vendors are aware of the labor math, too. The whole stack is being rebuilt around AI efficiency.

So the question isn’t whether disruption is coming to managed services. It’s here. The question is whether you’re the one disrupting or the one being disrupted.

Treeline’s $25 million just made that question harder to ignore. The smartest money in Silicon Valley just told you they think your model has a ceiling. You can disagree with them. But you’d better have something more than optimism to back it up.

Pick your lane. Go deep. Build the thing that costs $25 million and 40,000 competitors to replace. Or don’t — and watch Treeline show up in your market in 18 months.

That’s the directive. The math is already written.