VPS pricing is driven by four things — vCPU, memory, storage type and size, and bandwidth — with memory usually the dominant one. Entry plans run a few dollars a month, mid-range ten to thirty, and anything substantial goes well past a hundred. The monthly figure is also the smaller half of what a VPS costs.
Comparing VPS prices is easy and comparing VPS costs is not, because the advertised number covers the machine and nothing else. A VPS is unmanaged by default, which means the price excludes the backups, the monitoring, the patching, and the person who does all of it. Those are real costs and they do not appear on any pricing page.
Table of contents
- The four things you are paying for
- Why some are so much cheaper than others
- The costs the price list does not include
- Sizing without over-buying
- When a VPS is the right purchase
- Practical checks before buying
- How this fits the rest of the stack
- FAQ
The four things you are paying for
Memory. Usually the dominant factor in the price, and usually the constraint you hit first in practice. Application runtimes, database processes, and caches all want RAM, and running out produces the worst failure mode — the kernel starts killing processes and the machine behaves erratically rather than simply slowing down.
vCPU. Virtual cores, which are shares of physical cores. Two providers advertising 2 vCPU can differ substantially in real throughput depending on how heavily they oversubscribe. Look for terms like dedicated vCPU versus shared or burstable — burstable plans allow short spikes above a baseline and then throttle, which is fine for a site with quiet periods and unsuitable for sustained load.
Storage. Both size and type. NVMe is considerably faster than SATA SSD, which is far faster than spinning disk. For anything database-backed this matters more than the capacity, because disk latency is what a slow query feels like. Also check whether storage is local to the host — fast, and lost if the host fails — or network-attached, which is slower and survives.
Bandwidth. Usually a monthly allowance with an overage rate, occasionally metered from the first byte. Check the overage rate rather than only the allowance, and check whether inbound traffic counts.
Roughly, the bands are: entry at a few dollars for one core and half a gigabyte to two gigabytes of RAM, suitable for a small site or a development box; mid-range at ten to thirty for two to four cores and four to eight gigabytes, which is where real applications live; and thirty to over a hundred for anything with substantial CPU, memory, or storage.
Why some are so much cheaper than others
A two-dollar VPS and a twenty-dollar VPS with similar specifications are not the same product, and the differences are usually in things the listing does not mention.
- Oversubscription. How many virtual machines share each physical core. Aggressive oversubscription is the main way a cheap plan is made cheap, and it shows up as inconsistent performance rather than as a lower number on the spec sheet.
- Network quality. Peering, transit providers, and how close the datacentre is to your users. A cheap host with poor peering adds latency to every request.
- Storage backing. Local NVMe versus network storage versus something older.
- Support. Whether there is anyone to contact, and whether they answer in hours or days.
- Uptime commitment. Whether there is an SLA and whether it has any teeth.
- Promotional pricing. The advertised figure being a first-term rate that renews considerably higher.
The promotional pricing one is worth checking before anything else. A three-year prepaid deal at a low monthly equivalent that renews at four times the rate is a common structure, and it is only a good deal if you intend to move at the end of the term.
None of this means cheap providers are bad. It means the price difference is buying something, and knowing what lets you decide whether you need it. A development box or a hobby project genuinely does not need premium peering.
The costs the price list does not include
This is where the total diverges most sharply from the monthly figure.
Backups. Frequently an add-on at a percentage of the plan price, and occasionally not offered at all, in which case you build and test your own. A VPS with no backup is one disk failure or one bad command from total loss, and snapshots taken by the provider are not a substitute for backups you have restored.
Monitoring. Knowing the server is down before your users do requires something watching it, which is another service and another few pounds.
Patching. Security updates arrive continuously. Applying them, and occasionally dealing with one that breaks something, is recurring work with no upper bound on any given month.
Configuration. Setting up the web server, the certificate renewal, the firewall, the log rotation, and the deployment mechanism is a day of work at the start and a smaller ongoing tax forever.
Your time, which is the largest of these. Two hours a month at any professional rate exceeds the cost of most VPS plans. Four hours exceeds the cost of most managed alternatives.
The arithmetic worth doing once: take the monthly price, add backups and monitoring, then add your hourly rate multiplied by an honest estimate of the hours. That total is the comparison figure. A five-dollar VPS is frequently a two-hundred-dollar VPS once it is finished.
Sizing without over-buying
The usual mistakes are buying for a peak that has not happened and buying too small to survive the peak that does.
Start from what the application needs at rest, add headroom, and pick a provider that lets you resize. Most VPS providers can increase resources with a reboot, and some can do it without one. Being able to resize easily removes most of the pressure to guess correctly at the start.
Memory is where headroom matters most, because running out is a hard failure rather than a gradual slowdown. Leave meaningful slack — a machine sitting at 90 percent memory has no room for a traffic spike or a memory-hungry background job.
CPU is more forgiving. High CPU makes things slow; it does not usually kill them. So a burstable plan is a reasonable bet for uneven load in a way that a tight memory allocation is not.
Add swap even if you have plenty of memory. It will not save a genuinely undersized machine and it converts a sharp failure into a slow one, which buys you time to notice.
And check the resize path before you need it. Some providers can increase memory and CPU but not shrink disk, which means a resize is one-directional and an over-provisioned disk is permanent.
When a VPS is the right purchase
It is a genuinely good fit for several situations, and the point of this article is not to argue otherwise.
- You want full control of the environment — specific kernel modules, unusual software, a particular database version.
- You are running several unrelated things on one machine and want to pack them efficiently.
- You are learning system administration, in which case the work is the point.
- You have a fixed, predictable workload where a fixed-size machine is exactly right.
- Your requirements do not fit any managed platform’s model.
- You already have the operational capability and a VPS is the cheapest way to use it.
And when it is not: when nobody on the team wants to own an operating system, when traffic is spiky enough that a fixed size is either wasteful or insufficient, when the thing you are deploying is a straightforward application in a common runtime, or when the hours it consumes have a better use.
The honest comparison is not the monthly price against a managed platform’s monthly price. It is the total from the previous section against the platform’s price. Managed platforms are frequently cheaper on that basis and frequently not — it depends entirely on how much your time is worth and how much of it the server will take.
For reference on the other side of that comparison, the RunxBuild ladder starts at $4 on Dev and $6 on Basic for a service, with managed MySQL or Postgres on the same ladder, so the pieces are priced individually rather than as one machine you subdivide yourself.
Practical checks before buying
- Find the renewal price, not the promotional one, and multiply by the term you expect.
- Check whether backups are included, extra, or absent.
- Check the bandwidth allowance and the overage rate.
- Confirm the storage type and whether it is local or network-attached.
- Find out whether vCPU is dedicated or burstable, and what the baseline is.
- Check the datacentre locations against where your users are.
- Read the resize policy in both directions.
- Look for a refund window, and use it — deploy something real and measure it rather than trusting the specification.
That last one is worth the effort. Two VPS plans with identical specifications can differ substantially in real performance, and half an hour with a benchmark and a realistic workload tells you more than any comparison table.
Test disk latency specifically if the application is database-backed, since that is the number most likely to differ and least likely to be advertised.
How this fits the rest of the stack
VPS pricing is driven by memory, cores, storage type, and bandwidth, and the advertised figure covers the machine rather than the job. Add backups, monitoring, and an honest estimate of your own hours before comparing it with anything else. If most of that work is not what you want to be doing, the pieces priced individually are easier to reason about — the RunxBuild hosting calculator shows the service, database, storage, and bandwidth as separate line items rather than one number with several things hidden inside it.
Useful related references:
- High Performance VPS: What the Phrase Means and How to Verify It
- What CPU Do I Have: Reading the Answer on Linux, Windows and a VPS
- VPS SSD Storage: What You’re Actually Getting, When It Matters, When It Doesn’t
- Services on RunxBuild
FAQ
How much does a VPS cost per month?
Entry plans with one core and one to two gigabytes of memory run a few dollars. Mid-range plans with two to four cores and four to eight gigabytes run ten to thirty. Anything substantial goes past a hundred. Memory is usually the dominant factor, and the advertised figure excludes backups, monitoring, and your own time.
Why are some VPS providers so much cheaper?
Mostly oversubscription — more virtual machines per physical core, which shows up as inconsistent performance rather than in the specification. Also network peering quality, storage backing, whether support exists, and promotional pricing that renews at a much higher rate. For a hobby project none of that may matter.
What is the difference between dedicated and burstable vCPU?
Dedicated vCPU gives you a consistent share of a core. Burstable gives you a low baseline with the ability to spike above it for limited periods, after which you are throttled. Burstable suits uneven traffic with quiet periods and is unsuitable for sustained load, which is why the two can be priced very differently for the same core count.
Is a VPS cheaper than managed hosting?
On the monthly figure, usually. On the total, often not. A VPS is unmanaged, so backups, monitoring, patching, and configuration are yours, and two hours a month at a professional rate exceeds most VPS prices. Whether it is genuinely cheaper depends entirely on what your time is worth.
How much RAM does my VPS need?
Start from what the application uses at rest and leave meaningful headroom, because running out of memory is a hard failure rather than a slowdown — the kernel starts killing processes. A machine sitting at 90 percent has no room for a traffic spike. CPU is more forgiving, so err toward memory when choosing between them.