The MSP Roll-Up Era Is Ending

For fifteen years the smart money in managed services ran one play: buy small MSPs, bolt them together, sell the bigger thing for a higher multiple than you paid for the parts. The MSP roll-up built most of the national platforms you now compete against, and the deal counts say it is still accelerating. Trackers logged 466 North American MSP transactions in 2025, up about 20% on the year before, with private equity on roughly 69% of the disclosed ones. Evergreen Services Group crossed its hundredth cumulative acquisition and a billion dollars of ARR through its Lyra platform by the middle of last year.

So why do I think the era that produced all of that is ending?

I spent a decade in investment banking and private equity before I started running growth inside a national MSP, and I track this consolidation landscape closely. The volume is real. What is changing underneath it is the thing the volume was built on: the math that made aggregating revenue pay. My main source here is Top Down Ventures' State of MSP Capital 2025 report from last November. Top Down is a VC fund invested in the AI-software side of this thesis, so read their projections as projections. The underlying numbers I lean on trace to Canalys, CIBC, William Blair and PitchBook, and I have flagged which is which.

The three epochs of MSP capital

Capital has moved through managed services in three distinct waves, and each one rewarded a different behaviour.

The middle epoch is the one that shaped the market you operate in. From 2010 debt was cheap, recurring revenue was predictable, and private equity worked out that a fragmented industry of founder-led shops was perfect raw material. By 2020 William Blair counted around 13 billion dollars of dry powder aimed at the sector. That capital built Thrive, Evolve IP, Fully Managed, NexusTek and the rest of the platform names that now show up on every acquirer list. If you want the full picture of who assembled these platforms and where their money came from, I mapped the buyer landscape separately in who buys MSPs.

The pattern held because MSP private equity consolidation had a genuine edge to harvest: a regional shop bought at five times earnings, folded into a platform that trades at eleven or twelve, is worth more the moment the ink dries. That gap is the whole game. And that gap is closing.

The third epoch opened with the AI shock of 2022 and 2023. Once generative tools could actually take routine technical work off a technician's plate, the binding constraint on an MSP's margin stopped being how many shops you had bolted together and became how much of the work you had automated. That is a different question with a different answer, and it does not reward the same behaviour the middle epoch did. Capital noticed before most operators did, which is usually how these turns go.

Why the roll-up math stopped working

More than 150 MSP roll-ups were financed between 2010 and 2024, backed by firms like ABRY, Evergreen Services Group and Berkshire, building platforms such as Thrive, New Charter and Lyra. Average deal leverage sat around four times EBITDA, according to Top Down's read of the period. For most of that run the model compounded beautifully. Then three things happened at once.

The first is marginal efficiency. The tenth acquisition into a platform delivers less than the second did. Early on you consolidate tooling, centralize the NOC and SOC, and strip duplicate back office, and the savings are large. By the fiftieth bolt-on the easy integration gains are already priced in, and each new one adds complexity roughly as fast as it adds scale. Top Down's own framing is blunt: the marginal efficiency of each additional acquisition is now smaller.

The second is returns compression. By 2023, median buyout returns in the sector had compressed even as portfolio revenue kept growing, per Top Down's analysis. Growing the top line stopped translating into the returns the model was underwritten on. When revenue climbs and returns fall, the problem is not the operators. It is the thesis.

The third is crowding. When 75 or more active platforms are all chasing sub-ten-million-dollar targets, the multiple arbitrage that made the MSP roll-up strategy work gets bid away. Everybody knows a five-times shop reprices at eleven inside a platform, so sellers and their advisors price that in, and entry multiples drift up toward the exit multiple. Arbitrage only pays while it is scarce. This one is not scarce anymore.

None of this means consolidation stops. It means buying revenue stopped being the part that earns the premium.

The hope camp and the proof camp

Public markets already sorted MSP-adjacent businesses into two piles, and the gap between them tells you where value moved.

CIBC's 2025 work puts AI-native software at 10 to 12 times forward revenue against traditional SaaS at roughly 5 to 8. Top Down borrows a useful label for the same split inside managed services. One pile is the hope camp: roll-ups projecting synergies they have not yet delivered, asking to be valued on the scale they plan to reach. The other is the proof camp: AI-enabled platforms that can show documented savings, tickets closed by agents, hours removed, margin recovered. Firms with verified efficiency gains trade at roughly double the multiple of firms selling potential.

The phrase Top Down uses is proof per dollar invested, and it is the right lens. Investors are no longer paying for the story of scale, they are paying for evidence that a dollar of capital turns into a durable point of margin. A platform that can put a number on tickets its agents close and hours its automation removes is underwriting the multiple with data. A platform that can only point at a growing revenue line is asking to be trusted.

Read that as an owner and the message is direct. A platform that grows by acquisition alone is selling a promise about a future it has not built yet. That is the hope camp, and the market has started pricing it like one. This is the honest, uncomfortable part of the picture: for most of the last decade, aggregating revenue was enough to command the premium, and now it is table stakes rather than the story. What earns the premium is evidence.

The Amazon aggregator parallel

If a fragmented market plus cheap debt plus a roll-up thesis sounds like a guaranteed win, we ran that exact experiment in a different industry and watched it break.

Between 2020 and 2022, Amazon FBA aggregators raised billions to roll up independent third-party sellers. Thrasio and its many imitators pitched investors on the same logic that powers MSP consolidation: buy dozens of small revenue streams, centralize supply chain and marketing, capture synergies, sell the aggregated whole at a higher multiple. Then the synergies mostly failed to appear, interest rates rose, the debt caught up, and the category collapsed. Thrasio, once valued in the billions, filed for bankruptcy in 2024. I wrote up that unwind in detail in why the Amazon FBA aggregators collapsed.

The MSP market is healthier than FBA ever was. Recurring contracts, real switching costs and genuine operational integration make managed services a far better roll-up substrate than a pile of Amazon listings. I am not predicting an identical collapse. The parallel that matters is narrower and harder to dismiss: buying revenue is not the same as building an operation, and a model that only aggregates eventually runs out of the thing that made aggregation pay.

What replaces aggregation: margin expansion

The next vintage of returns comes from margin expansion through automation, not multiple arbitrage. That is Top Down's central call, and the operating data behind it is the most concrete part of their report.

Their pilot programs point to a 60% reduction in average ticket resolution time and a 25% drop in labor cost per device under management. Because labor is roughly half of an MSP's operating expense, that converts to around 400 basis points of EBITDA uplift with no new revenue at all. A traditional MSP averages something like 15% EBITDA; early automated prototypes in their fieldwork report 25 to 30%. You do not get that by buying another shop. You get it by making the shop you own materially cheaper to run.

Put it on a real balance sheet and the point lands harder. Take a ten-million-dollar MSP with 75 staff. Automate 30% of the Tier-1 workload and the same team supports about 20% more clients while labor cost stays flat. In Top Down's worked example, EBITDA margin lifts from roughly 15% to 18%, around 300,000 dollars of extra annual profit, with no new capital and no acquisition. Roughly half of MSP technical labor is still routine L1 and L2 work, per CompTIA, so the raw material for that shift already sits inside most operations. The gains are a projection, and the fund making it has a stake in the outcome, but the direction is hard to argue with.

This is why the character of the capital itself is changing. The old roll-up sponsor was a financial engineer: source targets, structure debt, integrate back office, exit on multiple. The successor is closer to an operational technologist, underwriting a deal on how fast an MSP can convert manual workflows into agent-executed ones. Top Down reframes the old Rule of 40 for this, adding automation-driven margin expansion alongside revenue growth. The venture money is following the same logic into the tooling layer, which I cover in venture capital funding for MSP buyers and in more depth on valuations in MSP valuation in the AI era.

The through-line is simple. Value used to accrue to whoever aggregated the most revenue. It is moving to whoever can document the most intelligence per dollar of cost.

What this means if you're selling to a platform in 2026

If you are weighing a sale, the single most useful thing to internalize is that the buyer's model changed, so what the buyer pays for changed with it.

The headline multiples still look familiar. A sub-five-million-dollar-EBITDA shop realistically clears 3.5 to 5 times from a regional consolidator. Higher-quality recurring-revenue MSPs run 4.5 to 8. Platform-tier businesses over 500 million in enterprise value have been printing around 11 times and up, with cyber-heavy or scale-leader exits reaching into the teens and, in a few cases, near 20. Recurring revenue percentage still drives the base multiple. That part has not moved.

What moved is the swing factor on top. Top Down maps a three-phase adoption curve, and it weights valuation hard. A shop still using AI only for internal productivity, the augmentation phase, gets roughly half the multiple of a mature operator. Put AI into client workflows with human validation, the integration phase, and you price near market. Offer SLA guarantees around AI performance, the autonomy phase, and you reach premium multiples. Two MSPs with identical revenue can now sit a full turn or two apart on the strength of documented automation, and that is a projection from a fund with a stake in it being true, so weigh it as one.

You can see the same shift in how the smarter platforms want to price their own client contracts. Static per-seat subscriptions are giving way to what Top Down calls contracted usage corridors, minimum commitments with outcome bonuses on top, and to per-agent and per-outcome models where the MSP gets paid for assurance rather than hours worked. A buyer building toward that pricing wants to acquire operations that already generate the underlying proof, because that is what the new contracts are sold on.

The practical read is not investment advice, it is market structure. The buyer sitting across from you in 2026 is underwriting margin expansion, not just revenue you bring. So the metrics that used to live in a diligence footnote, tickets closed without a human, hours removed per technician, cost per endpoint, are becoming the number the offer turns on. If any of that is real in your business, document it before a process starts rather than trying to narrate it mid-diligence. For the wider preparation picture see MSP exit readiness, how the terms get built in deal structures for an agency exit, and where automation actually lands on the P&L in the MSP profit margin benchmark.

The bottom line

The roll-up era is not ending because consolidation is stopping. Deal volume is climbing and will keep climbing while thousands of founder-led shops still need a buyer. It is ending in a narrower, more important sense. Aggregating revenue stopped being the thing that earns the premium.

For a decade, scale was the story. Buy the parts cheap, assemble them, exit on a bigger multiple, repeat. That arbitrage is being bid away, the returns started compressing while revenue kept growing, and the market has begun paying for documented efficiency instead of aggregated size. The hope camp gets priced like hope. The proof camp gets paid.

Whoever you sell to in 2026 is underwriting a different model than the one that built the last decade of platforms. They are buying margin they can prove, not revenue they can add. Understand that shift and you are reading the market the way the capital already reads it, which is the only vantage point from which a sale ever goes well.