Ask what a healthy MSP margin looks like and you will get a tidy number back: 15%, maybe 20% if the person is feeling generous. That number is real, and it is also a lie by omission. The average MSP profit margin sits in the mid-teens on an EBITDA basis, but the average is a barbell dressed up as a bell curve. Roughly a third of managed service providers run at a loss. Another third scrape by at low single digits. The mid-teens "average" is what you get when you blend the strugglers with a small group of operators printing 25% and up.
I run growth inside a national MSP and spent a decade before that in investment banking, a family office, and private equity, which means I read this market the way an investor reads a cap table: follow the margin, not the headline. This piece is the honest benchmark. Where margins actually land by region, why the average deceives, and the mechanics that separate the profitable third from everyone else.
Most of my numbers here trace to one report: Top Down Ventures, "The State of MSP Capital in the Age of AI," published November 2025. One disclosure up front. Top Down is a venture fund invested in exactly this thesis, MSP-adjacent AI software, so its own pilot-program figures are its own and directionally useful, not audited. Where a number comes from a primary source (Canalys, CIBC, CompTIA, GTIA), I name it. Read the projections as projections.
The quick benchmark: where MSP margins land in 2026
If you only want the numbers, start here. Average EBITDA margins for MSPs split cleanly by region, and the gaps are wide enough to matter.
| Region | Average EBITDA margin | Notes |
|---|---|---|
| North America | ~17% | ~45% of global managed services revenue (Canalys) |
| EMEA | ~12% | Early automation adopters report a 4-point lift within 12 months |
| APAC | ~10% | Valuations trail NA by 1 to 2 turns of EBITDA on governance perception |
Those regional averages come from Canalys, cited in the Top Down report. Blend them and you land near the number the market repeats: a traditional MSP averages around 15% EBITDA. The interesting part is the tail. Early operators running automation-heavy models, what the report calls managed-intelligence prototypes, report 25% to 30% EBITDA (Top Down Ventures, State of MSP Capital 2025). Same industry, double the margin.
Canalys expects global profitability to drift toward the high teens by 2028 as automation equalizes cost structures across regions. So the regional gaps you see today are partly a timing story, not a permanent feature of where you happen to operate.
Net, EBITDA, gross: you are probably comparing the wrong number
Before you benchmark yourself, get the denominator right. Three margins get quoted in MSP conversations and they are not interchangeable.
Gross margin is revenue minus the direct cost of delivering the service, mostly technician labor and the software you resell. Healthy MSPs run 50% to 60% gross, and the best push past 70%. EBITDA margin is what is left after overhead, sales, and admin, before interest, tax, and depreciation. That is the number buyers and lenders care about, and it is where the mid-teens average lives. Net margin sits below EBITDA after financing and tax, which is why you sometimes see "20% to 30% net" quoted as a target when the same business would report far less on an EBITDA basis.
When someone tells you their margin, ask which one. A 55% gross margin and a 12% EBITDA margin can be the same MSP on the same day. The report anchors its analysis on EBITDA because that is what capital prices, and adjusted EBITDA is the figure that shows up in every valuation model. If you want the deeper version of how margin feeds a multiple, I walk through it in cost-plus pricing and MSP valuation.
The barbell the averages hide
Here is the number that reframes everything. Roughly one third of MSPs operate at a loss, and another third sit at low single-digit margins (Top Down Ventures, State of MSP Capital 2025, citing the profitability distribution). Independent market data corroborates the shape: industry benchmarks have shown close to a third of MSPs losing money or barely breaking even on an EBITDA basis in recent quarters.
Put those two facts together and the "average" dissolves. Two-thirds of the industry is at or near break-even. The mid-teens headline exists because the top third, and especially the top decile running above 20%, drags the mean up. Report an average to a room of MSP owners and most of them are below it, which is exactly why the average is a bad target. It describes a population you are probably not standing in.
This matters for what you do next. If you are in the bottom third, the fix is rarely "grow revenue." Adding low-margin clients to a low-margin business just scales the problem. The profitable third did something structural: they priced better, sold a different mix, automated the routine work, or fired the clients bleeding them. Revenue scale and margin are different games, and the barbell is the proof.
The automation dividend: 400 basis points with no new revenue
The single biggest lever the report documents is automation, and the math is specific enough to test against your own P&L. In Top Down's pilot programs, MSPs adopting agentic automation saw a 60% reduction in average ticket resolution time and a 25% decrease in labor cost per device under management. The combined effect: roughly 400 basis points of EBITDA uplift with no top-line growth (Top Down Ventures, State of MSP Capital 2025). These are the fund's own pilot numbers, so treat them as directional, but the direction is what matters.
Work it through on a concrete business. The report models a $10M revenue MSP with 75 employees automating 30% of its Tier-1 workload.
| Line | Before automation | After automating 30% of Tier 1 |
|---|---|---|
| Revenue | $10.0M | $10.0M (same clients, plus ~20% more capacity) |
| Staff | 75 | 75 (flat) |
| Labor as share of opex | ~50% | ~50%, spread across more clients |
| EBITDA margin | ~15% | ~18% |
| Annual EBITDA | ~$1.5M | ~$1.8M |
The same 75 people support around 20% more clients, labor costs stay flat, and the margin lifts from 15% to 18%. That is about $300K of additional annual profit with no new capital and no new headcount (Top Down Ventures, State of MSP Capital 2025). Scaled across a $595B industry, the report puts the total prize north of $20B a year, but the number that should interest you is the $300K on your own books.
There is a flywheel underneath the one-time gain. The report models that once automation covers roughly 20% to 25% of recurring workflows, each additional 5-point gain in automation buys about 2 more points of margin and roughly 10% faster deployment cycles (Top Down Ventures, State of MSP Capital 2025). So the first 30% of Tier 1 is the hard part. After that, the margin curve gets steeper, not flatter, which is why the early adopters running 25% to 30% EBITDA are pulling away rather than being caught.
The catch is that the tools doing this are consolidating fast, and betting on the wrong vendor is expensive. I mapped where the automation money has landed and who is likely to get absorbed in the MSP tool stack breakdown.
Labor is half your cost base, so it is the whole game
Why does automating tickets move the margin so hard? Because labor is roughly 50% of operating expenses for most providers (CompTIA 2024, via the Top Down report). When half your cost base is people, a small efficiency in how those people spend their time compounds straight to the bottom line. CompTIA's figure is blunt: a 25% reduction in labor expense yields 150 to 250 basis points of free-cash-flow margin improvement even with flat revenue.
The reason there is so much slack to capture is that a lot of MSP labor is not skilled work. Around half of MSP technical labor is still routine L1 and L2: password resets, endpoint provisioning, patch compliance (CompTIA 2024). Top Down estimates that 40% of the average technician's workflow will be performed by AI agents by 2028, and 62% of channel partners already plan to hire "AI service managers" within twelve months (GTIA 2025).
The point for margin is simple. Your most expensive cost line is people doing work that increasingly does not require a person. Close that gap and the margin follows. Ignore it and the operators who do not will out-price you on the same contracts.
Margin quality now sets your valuation, not revenue scale
For most of the last decade the way to a big exit was to get big. Private equity financed more than 150 MSP roll-ups between 2010 and 2024, stacking revenue on around 4x EBITDA of debt and selling the scale (Top Down Ventures, State of MSP Capital 2025, citing PitchBook and William Blair). That trade is fading. Median buyout returns compressed even as revenue grew, and the marginal acquisition now adds less than the last one. The next vintage of returns comes from margin expansion, not from buying revenue and hoping the multiple holds. I unpack why in the roll-up era is ending.
Capital has split into two camps. CIBC's 2025 data shows AI-native software trading at 10x to 12x forward revenue while traditional SaaS sits at 5x to 8x, and the report frames the same divide inside managed services: a "hope camp" selling projected synergies versus a "proof camp" selling documented savings. Buyers now reward proof per dollar invested. An MSP with verifiable automation metrics, tickets closed by agents, hours saved, margin recovered, trades at a premium to one selling potential.
Where does that land? Canalys expects scaled, AI-enabled MSP platforms to cluster around 9x EBITDA as regional margins converge. The lesson for an owner is that two MSPs with identical revenue can be worth very different money, and the gap is margin quality plus the evidence behind it. For the full list of what moves the multiple, see MSP valuation drivers.
The report goes one step further and prices the maturity of your automation directly into the multiple. It splits the transition into three phases: augmentation (AI on your own internal work), integration (AI inside client workflows with a human checking it), and autonomy (you guarantee an SLA around the AI's performance). A Phase 1 operator earns roughly half the multiple of a mature one, Phase 2 sits near the market average, and Phase 3 commands a premium (Top Down Ventures, State of MSP Capital 2025). Same margin on the P&L, different value on the way out, because a buyer pays more for a margin that is durable and documented than for one that might reverse next quarter.
What separates the profitable third
The operators printing 20% and up did not get lucky on region or size. Four things separate them, and all four are choices.
Pricing discipline comes first. The profitable third charge for outcomes and hold their rates at renewal instead of absorbing cost creep. If your pricing has not moved in three years while your costs have, the margin gap is self-inflicted. There is a subtler trap here too. As automation makes each ticket cheaper to deliver, the instinct is to pass the saving to the client. Pax8 data in the report points the other way: clients pay more for verified outcomes, uptime guarantees, and audit readiness than they do for time worked (Top Down Ventures, State of MSP Capital 2025). The reframe is that automation is not cheaper labor to be discounted, it is better assurance to be priced. I go deep on the number itself in how much an MSP should charge and on the structures in MSP pricing models.
Service mix is second. Compliance and security-focused managed offerings carry gross margins 5 to 8 points higher than traditional infrastructure management, because clients treat compliance as risk insurance and pay a premium for it (Canalys 2024, via Top Down). Moving one client from foundational IT to a security or compliance practice lifts revenue and margin at the same time.
Automation is third, and it is the 400-basis-point lever from earlier: take the routine L1 and L2 load off human hands and let the same team carry more clients. The report also documents new pricing shapes built around this, from per-agent billing (one MSP replaced three help-desk staff with a set of autonomous agents, cutting cost 40% while doubling response speed) to shared-savings models where you keep a slice of the efficiency you deliver.
Client selection is fourth and the least glamorous. A third of the industry loses money partly because it keeps clients that cost more to serve than they pay. Firing the bottom of your book is the fastest margin move that requires no new tooling. As for what to measure, stop tracking revenue growth in isolation. Track adjusted EBITDA margin, labor cost per device under management, and the share of tickets closed without a human touching them. Those three tell you which third you are in and whether you are moving.
The Bottom Line
The average MSP profit margin is a mid-teens number that describes almost nobody. The industry is a barbell: a third at a loss, a third barely above water, and a top slice running double the mean. Benchmarking yourself against the average tells you nothing useful. Benchmarking against the distribution tells you where you actually stand and how far the ceiling is.
The good news is that the levers are known and mostly within your control. Price for the value you deliver, shift the mix toward security and compliance, automate the routine labor that eats half your cost base, and cut the clients you are subsidizing. Do those and the margin follows, and in this market the margin is what the next buyer is paying for. Revenue got you a seat. Margin quality is what gets priced.