Cloud Migration Cost: What You'll Actually Pay in 2026

Cloud migration usually costs a 20 to 500-user company $15,000 to $75,000 as a one-time project, plus a monthly cloud bill that runs for as long as you stay in the cloud. The project is the down payment. The run rate is the real number, and it's the one most quotes leave out.

That's the real problem with how cloud migration cost gets talked about. For any company weighing managed cloud services, you ask what it costs, someone hands you a range between $15,000 and a million dollars, and you're no closer to a decision than when you started. A range that wide isn't a price. It's a shrug.

Here's the part that trips up most companies. The migration is a one-time event with a start and an end. The cloud bill isn't. It shows up every month, forever, and it's almost always higher than the number you signed off on. Worldwide public cloud spending is on track to hit $723 billion in 2025, up 21.5% in a single year, according to Gartner. A lot of that growth is companies paying more than they planned to.

So what does cloud migration actually cost?

For a company with 20 to 500 users, a full migration typically runs $15,000 to $75,000 as a one-time project, plus $1,500 to $12,000 a month in cloud fees after cutover. Small file-and-app moves sit at the bottom. Regulated or custom environments sit well above it.

Those are two separate numbers, and blending them is where budgets go wrong. The project cost covers planning, moving your data, standing up the new environment, testing, and cutover. The run rate is what you pay every month once you're live. One is a project. The other is a subscription you signed up for without reading the recurring line.

Here's a rough map by environment size. Treat these as planning ranges, not quotes. Your real number depends on how much data you're moving and how much of your old setup you drag along with it.

• 3 to 5 servers, file shares and basic apps. A one-time project of $15,000 to $30,000, then $1,500 to $4,000 a month after cutover.
• 10 to 20 servers running line-of-business apps. A one-time project of $30,000 to $75,000, then $4,000 to $12,000 a month.
• 25 or more servers with custom or regulated workloads. A one-time project of $75,000 to $150,000 or more, then $12,000 to $30,000 or more a month.

Notice the run rate doesn't shrink over time on its own. It creeps up. New workloads, more storage, forgotten test servers nobody shut off. Left alone, a cloud bill grows. That's not a flaw in the cloud. It's what happens when you rent instead of own and nobody's watching the meter.

Why does every quote you get look so different?

Because two vendors can move the same servers in completely different ways, and each way carries a different price. The industry calls these the 6 Rs. You don't need the jargon. You need to know which one you're buying, because it decides both your upfront cost and your monthly bill.

Think of it as a spectrum. On one end, you move everything exactly as it is and change nothing. Cheap to do, but you carry your old inefficiency into a place that charges by the hour. On the other end, you rebuild your apps to run natively in the cloud. Expensive and slow, but the monthly bill comes out lowest. Most SMBs land somewhere in the middle.

• Rehost, also called lift-and-shift. You move it as-is and change nothing. Lowest upfront cost, but the monthly bill can climb if legacy waste moves with it.
• Replatform. Small optimizations on the way over. Medium upfront, and it often runs 20% to 40% lower than a straight rehost.
• Refactor means rebuilding the app to run natively in the cloud. Highest upfront, 2 to 5 times a rehost, but the lowest bill long term.
• Repurchase. Drop the old app and buy a SaaS version instead. Cost varies, but you trade it for a predictable subscription.
• Retain. Leave it on-premise for now. No migration cost, no change to your bill.
• Retire the thing nobody uses. Shut it down and it removes the cost entirely.

The trap is rehost. It's the cheapest quote, so it wins the deal. But lift-and-shift takes a bloated legacy server that was wasting capacity in your closet and moves that exact waste to a place that bills you hourly for it. You were overpaying in electricity. Now you're overpaying in cloud credits. Same waste. New invoice.

That's why the cheapest migration quote sometimes produces the most expensive cloud bill. Ask any vendor which R they're quoting. If they can't answer plainly, that's your answer.

What's the one cost nobody puts in the quote?

The double-run overlap. It's the stretch where the old system and the new one both run at the same time while you test and validate, and you pay for both. It's the single most common reason migrations blow their budget.

Here's how it happens. The plan says you'll run old and new side by side for two weeks, confirm everything works, then flip the switch. Clean. Then testing surfaces a problem. An approval slips. A department isn't ready. Two weeks becomes six. For a month and a half, you're paying your old hosting or hardware costs and your full cloud bill at the same time. Nobody budgeted for that second rent check.

We've watched a planned 3-week overlap run to 9 weeks because one line-of-business app wouldn't behave in the new environment. The migration itself came in on budget. The overlap is what stung. That's the pattern almost every time, and it's exactly why a migration plan that holds up matters more than a low sticker price.

There's a real lever here, though. Microsoft's Azure Hybrid Benefit gives you 180 days to run on-premise and Azure workloads at the same time at no extra licensing cost. Used well, that turns the most expensive part of the timeline into a planned, funded window instead of a surprise. Most companies don't know it exists until after they've paid for the overlap twice.

What drives your number up or down?

A handful of things push your number around more than compute and storage ever will. Compute is the part everyone models. It's rarely what wrecks the budget.

• How much data you're moving. Moving data in is manageable. Pulling it back out later, egress, runs $0.05 to $0.09 per gigabyte and adds up fast if apps constantly reach across environments.
• Licensing. Software licenses can cost more than the cloud infrastructure running them. This is where Azure Hybrid Benefit earns its keep, cutting Windows and SQL Server costs by up to 85% versus pay-as-you-go if you already own the licenses.
• How much legacy junk you carry over. Every idle server and oversized VM you migrate becomes a recurring charge. Retire what you can first.
• Downtime tolerance. A migration that can't take the business offline for a weekend costs more than one that can. Zero-downtime cutovers need parallel infrastructure, and parallel means paying twice.
• Integration and dependencies. The app that quietly talks to four other systems is never a clean lift. Untangling it is labor, and labor is the biggest line item in most migrations.

Notice what topped that list. Not servers. Data, licensing, and the stuff you drag along. Flexera's latest survey found managing cloud spend is now the number one concern for organizations at 82%, edging out security, and that roughly 29% of cloud spend is wasted. The waste doesn't come from the move. It comes from what you moved and never cleaned up.

How do you build your own estimate?

Start with the run rate, not the project. The project cost is a one-time hit you can negotiate. The run rate is what you'll live with for years, so model it first.

Pull your current server list. For each one, map it to its cloud equivalent in the Azure Pricing Calculator or the AWS Pricing Calculator. That gives you a monthly baseline. Then do the honest part, which most people skip. Add storage growth, add egress if your apps are chatty across environments, and apply Hybrid Benefit if you own your licenses. TechTarget has a solid walkthrough on calculating migration cost before you move if you want the long form.

Then build the project number on top. A workable rough model looks like this.

• Labor to plan, move, test, and cut over, usually the largest slice.
• Data transfer and any temporary storage during the move.
• The double-run overlap. Budget for double your planned timeline, not your planned timeline.
• A 10% to 20% contingency, because something always surfaces after cutover.

Two numbers. Run rate first, project second. If a vendor quote only shows you one of them, you're holding half the picture. Ask for both in writing before you sign anything.

When is cloud migration not worth the cost?

When your workloads are steady, predictable, and running hot on hardware you already own. Cloud wins on variable and growing workloads. It loses on stable ones that run hot around the clock, and paying rent on something you could own outright rarely pencils out.

This is the part cloud vendors don't lead with, and we're a cloud provider, so read that how you want. The most public example is 37signals. The company behind Basecamp pulled its workloads back out of the cloud and cut its bill from $3.2 million a year to $1.3 million, saving roughly $2 million annually. Their workloads were big, steady, and predictable. Textbook case for owning instead of renting.

You're probably not 37signals. But the logic scales down. If most of what you run is a stable file server and a couple of line-of-business apps that haven't changed in five years, the honest math on on-premise versus cloud tradeoffs might point at a hybrid setup, not a full move. Cost is one of the most common reasons companies pull workloads back out, and many say tighter upfront modeling would have kept them in. The lesson isn't avoid the cloud. It's run the numbers honestly before you commit, not after the first surprise bill.

What does this look like with a partner vs doing it alone?

The migration is the easy part to price. The run rate is the part that needs someone watching it, and that's where a managed partner changes the math. A one-time project has a clear end. Cloud cost governance doesn't.

Doing it alone, the common failure isn't the move. It's month four, when nobody owns the bill, test servers are still running, storage has crept up, and the cost has quietly drifted 30% past what you modeled. Flexera found 59% of organizations now run a dedicated team just to manage cloud spend, because left unmanaged, it drifts. Every time.

A managed approach handles the move and then keeps watching the meter. Right-sizing, shutting down idle resources, applying licensing benefits, catching the creep before it compounds. Consilien runs migrations to fully managed private cloud where the run rate is governed, not just billed. We translate the technical decisions into what they cost your business, because a cloud bill is a business decision wearing a technical costume. That's the whole point of doing this with someone who does it for a living.

The two numbers to walk away with

Cloud migration isn't one cost. It's a one-time project and a monthly run rate that outlives it, and any conversation that only covers one of them is incomplete. Get both in writing. Budget the double-run overlap before it surprises you. And apply the licensing benefits you already qualify for, because leaving Hybrid Benefit on the table is money you're handing over for no reason.

Run the honest numbers first. If cloud fits your workloads, it's one of the better business decisions you'll make. If it doesn't, better to know now than at month four. If you're weighing a move and want a real number instead of a range, speak to a cloud expert who'll show you both figures before you commit to either.