Every wave of business software has produced its consolidator. ServiceNow rolled up the help desk. Workday rolled up HR. The pattern repeats: a fragmented services layer gets replaced by a vertical platform that does the work better, faster, cheaper.

MSPs have somehow escaped this fate.

The Managed Service Provider (MSP) market is a ~$600B giant growing at 12% annually. It comprises roughly 200,000 small operators across the US and Europe who deliver the outsourced IT backbone for virtually every SMB and mid-market company. Yet, the fragmentation is extreme: the largest player in the world holds less than a 2% market share. The median operator does $2-5M in revenue, relies on software written in the early 2000s, and is owned by an engineer in his late fifties with no succession plan.

The friction is structural. We’ve spent months talking to owners, vendors, and M&A specialists. We believe AI-native MSP rollups are the only way to play one of the largest unconsolidated services markets left in B2B.

This isn't a software play - it’s an acquisition play.

Why Nothing Has Changed

The interesting question is why this market has stayed fragmented for thirty years. Three "moats" keep new entrants out and incumbents stuck:

The Relationship Moat

‍In an 18-person MSP, the owner is the brand. They hold 80+ personal relationships built over decades. Customers don't switch because of a "better dashboard"; they only move due to catastrophic service failure or a massive price gap.

The Workflow Moat (The PSA)

‍The Professional Services Automation (PSA) system is the MSP’s operating system. Every contract, ticket, and invoice runs through it. These systems - ConnectWise, Kaseya, Halo etc - are legacy monoliths from 2005. They are so integrated and painful to migrate that one operator told us: "It would be easier to sell the business and start a new one from scratch than to switch my PSA." New PSA entrants have tried to displace the incumbents every 2-3 years for two decades. None has.

The Labor Moat

‍The cost structure is technicians all the way down. In the US, senior techs cost $80-120K and churn every year. They spend 80% of their time on repeatable "L1" tasks - password resets, printer issues, M365 access, employee onboarding, license reconciliation - 50-step SOPs that humans run simply because the software stack can’t.

Traditional PE has been running at this for a decade. Evergreen Services Group has done over 100 acquisitions and built a $1B+ revenue business. Thrive, New Charter, Alpine - all active. The model works as a financial exercise. It hasn't worked as an operational one. The synergies on offer are vendor consolidation and shared back-office. The actual delivery - what a technician does for a customer on a Tuesday morning - looks the same in 2026 as it did in 2014.

What changes with AI

The MSP software vendors - ConnectWise, Kaseya, N-able - have all shipped AI features in the last 18 months: ticket summarization, natural-language-to-PowerShell, auto-resolution of L1 issues. But adoption is bimodal. MSPs with an "AI champion" - usually the owner - hit 60-70% automation. The ones who bought the AI feature as a checkbox get nothing. 

That gap is the entire opportunity. A retrofitted AI module on top of a 20-year-old PSA saves a few hours a week in the hands of a motivated user. An AI-native delivery layer - where the agent runs L1 work without a human in the loop - changes the unit economics of the business decoupling revenue from headcount.

Why a Roll-Up Is the Right Wrapper

The obvious venture move is to back a company building the AI layer as software and selling it to MSPs. Build the product, ship to the long tail, ride the SaaS multiple.

The history of this market is unkind to that thesis. Each of the three constraints above is also a constraint on a software-only entrant. You can't reach the customer because the relationship belongs to the owner. You can't displace the PSA from outside, and every adjacent tool has to integrate with it before any MSP will even consider switching.

The way around both problems is the same: don't sell to MSPs. Become one.

Buy MSPs. Layer AI on top of their existing stack. Compound the productivity gains across the portfolio. Use the customer base as the wedge for new growth services that today's owners can't deliver. The acquired customer book gives you cash flow on day one. The relationships transfer (with retention earnouts on the seller). The AI deploys against a controlled internal environment instead of against an unmotivated external buyer. The upside compounds.

The Playbook

The acquisition target zone is $5-25M revenue, 10-50 employees.

Below $5M, the business is owner-dependent. Above $50M, you are paying enterprise multiples for a business that is already optimized. The $5-25M zone is where you find documented processes, recurring revenue, and an owner who is actually willing to sell because his exit multiple has compressed and the next decade isn't getting easier. There are thousands of these targets across the US and Europe.

Sequencing matters more than M&A velocity.

The first 18 months must focus on productivity, not workforce reduction. Positioning is everything: call it "automation," not "AI." By targeting the workflows technicians hate - billing reconciliation, license provisioning, and offboarding - you build a buyer brand in the seller community. Aggressive, day-one cost-cutting kills your M&A pipeline.

Retention earnouts are non-negotiable.

Post-acquisition churn can hit 75% if the owner walks. We look for 2-3 year retention earnouts tied to customer continuity. With the seller locked in, churn typically drops to single digits, allowing the relationship asset to transfer safely to the new platform.

Regional and vertical focus beats horizontal.

Horizontal roll-ups struggle with geographic friction. Regional and vertical specialists (e.g., healthcare in the Northeast) outperform because the relationship moat doesn't compress and the compliance requirements are specific.

The growth services thesis has to be specific.

We want to see a defined layer - compliance-as-a-service, managed security operations, AI-adoption consulting - for ACV expansion. Most MSPs say they're prepared to pay more for security tool integration. The demand is real; we want to see operators who have picked a wedge.

The India Angle

Almost every roll-up that's tried to use offshoring as the primary lever has run into the same wall. End customers don't want their offshore technician on the phone. Healthcare and government contracts often forbid it. The trust deficit is real and slow to close.

But India is exactly the right place to anchor the AI-enabled growth services arm - security operations, compliance automation, AI-adoption consulting for the 50-500 person companies that are too small for Accenture and too tech-backwards to self-serve. These are services with no incumbent relationship to defend, where the work is delivered through the platform rather than to the customer's desk, where the change-management problem is internal rather than external.

‍

Every fragmented B2B services market eventually gets its consolidator. MSPs didn't get one for thirty years because the technology to consolidate them didn't exist. It does now. The risks are real - execution, capital intensity, the threat of disintermediation from above - but the gap between what existing players are building and what's actually possible has never been wider.

We're looking for operators who've spent time inside this market - running an MSP, leading M&A integration, or building products - and have thought about how to do this differently. If that's you, we'd love to talk. Reach out at sayantan@stellarisvp.com