← Back to blog

Hourly VPS Hosting: Pay Per Hour Cloud Servers Explained

Maria Ilinca Bostan

Hourly VPS hosting bills a virtual private server for the hours it exists rather than for a calendar month. You deploy in under a minute, pay a per-hour rate — from €0.0056 on our smallest plan — and destroy the instance when the work is done. On plans with a monthly cap, running continuously costs exactly the same as a monthly plan, so the flexibility is free rather than a premium.

What hourly VPS hosting is

A VPS is a virtual machine with dedicated RAM, dedicated disk and its own kernel, isolated from other guests by a hypervisor. Ours are KVM, which means full hardware virtualisation and genuine root access rather than a container sharing the host's kernel — you can load kernel modules, run your own firewall, mount filesystems and install a hypervisor inside it if you want to.

The hourly part is a billing decision layered on top. Three details determine whether it is genuinely useful or just marketing:

  • Provisioning speed. Hourly billing is pointless if a server takes four hours to build. Ours deploy in under sixty seconds, which is what makes a one-hour instance a sensible thing to create.
  • A monthly cap. Without one, an always-on instance costs more than the monthly equivalent. Every plan below is capped.
  • No metered extras. A low hourly rate with billed egress is not cheap. Traffic here is unmetered with no overage charges, so the hourly rate is the entire price.

The plans and what they cost per hour

PlanRAMvCPUSSDNetworkPer hourMonthly cap
Ion2 GB140 GB1 Gbps€0.0056€4.00
Pulse4 GB280 GB1 Gbps€0.0083€6.00
Core8 GB2120 GB1 Gbps€0.0139€10.00
Forge16 GB4160 GB1 Gbps€0.0222€16.00
Apex24 GB6200 GB1 Gbps€0.0417€30.00
Titan32 GB8300 GB1 Gbps€0.0806€58.00

Every plan includes full root access, KVM isolation, DDoS protection, unmetered traffic and deployment in any of twelve locations. For bandwidth-heavy work there is a parallel 10 Gbps streaming range from €9.99/month on non-blocking ports.

Who this billing model is actually for

Development and staging environments

A Forge instance used twelve hours a day on weekdays runs about 264 hours a month — €5.86 instead of €16. More importantly, an environment that costs cents is one you rebuild rather than repair when it drifts from production.

CI runners and build agents

Build capacity is bursty by nature. Four Apex instances alive for 25 minutes across forty pipeline runs a month costs under €3. Provisioned monthly, the same capacity is €120. This is the workload where hourly billing changes what is architecturally reasonable, not just what it costs.

Load testing and benchmarking

Testing on a smaller instance than production and extrapolating is how capacity plans go wrong. Deploy the exact plan you are considering, run the real workload, read the real numbers, destroy it. A day of Titan is €1.93.

Seasonal and event capacity

Sales peaks, product launches, tournament weekends, election nights. Add instances for the window, remove them afterwards, and skip the quarterly capacity negotiation entirely.

Short-lived infrastructure

A VPN endpoint for a trip, a scraping job, a one-off migration host, a demo environment for a client meeting, a temporary game server for a weekend. All of these are badly served by a monthly commitment.

Businesses with no traffic history

A new product cannot be sized from data that does not exist. Start larger than you think, measure for two weeks, then size down. Hourly billing makes an over-estimate cost hours instead of a quarter.

Who it is not for

Two honest exclusions. If your workload is genuinely fixed and you are certain it will run for a year, providers offering 20–40% annual prepayment discounts will beat any hourly rate — we do not discount for term, and for that customer it is a real disadvantage. And if your load is sustained and heavy around the clock, the question is not hourly versus monthly at all; it is whether you should be on bare metal, where €209/month buys 16 physical cores and 128 GB of ECC memory. Our VPS versus dedicated comparison works through where that line sits.

The rule people get wrong: stopped is not destroyed

A powered-off instance still holds its disk, its IP address and its allocation on a host, so it keeps billing. Shutting a server down at night does not stop the meter — destroying it does. Take a snapshot if you need the data, and treat "shut down to save money" as a myth rather than a strategy. The largest single source of waste in hourly billing everywhere is instances nobody remembered to remove.

Deploying one

  1. Pick the plan from the table above, or model it on the cost calculator. Size for peak memory; CPU can be resized more comfortably than RAM can.
  2. Pick the location. Latency to your users matters more than almost anything else you will choose. Frankfurt sits on DE-CIX, Amsterdam on AMS-IX, and there are twelve sites across Europe, Turkey and the US East Coast.
  3. Choose the OS. Ubuntu, Debian, AlmaLinux and Rocky are all available at deploy — our distribution comparison covers the differences that matter.
  4. Deploy. Under sixty seconds to a running machine with root credentials.
  5. Harden it before it does anything useful. Key-only SSH, a firewall, automatic security updates. Start with the firewall guide and the hardening checklist.

Keeping the bill predictable

  • Destroy, do not stop. Snapshot first if the data matters.
  • Automate the teardown. An instance created by a pipeline should be destroyed by the same pipeline, including when the pipeline fails.
  • Review weekly. Five minutes looking at what is running catches the forgotten instances that quietly become permanent.
  • Right-size after two weeks of data, not on day one and not never.
  • Remember the cap. Anything you genuinely intend to run all month is already priced as a monthly plan, so there is no reason to agonise over its hours.

Frequently asked questions

How is hourly VPS billing calculated?

Hours the instance exists × the plan's hourly rate, stopping at the monthly cap. A Core instance at €0.0139/hr reaches its €10.00 cap at about 719 hours, so a full month costs €10.00 and a week costs €2.33.

Does hourly billing mean worse performance?

No. It is a billing model, not a hardware tier. The same KVM isolation, dedicated RAM, SSD storage and 1 Gbps unmetered port apply whether you keep the instance for an hour or a year.

Can I resize an hourly VPS?

Plans can be changed as your requirements move; billing simply continues at the new plan's hourly rate. Sizing up is generally smoother than sizing down, so start closer to your expected peak.

Is there a minimum term or setup fee?

No minimum term, no setup fee and no contract. Either would defeat the purpose of hourly billing.

What happens to my data when I destroy an instance?

It is gone. Take a snapshot or copy the data off first — see automating VPS backups for a routine that runs without you remembering it.

Where to go next

For the cost arithmetic in more detail, hourly versus monthly billing costs out four real scenarios. For sizing, choosing the right VPS plan. For what to run on it, the use-case library covers game servers, VPN endpoints, CI pipelines, AI workloads and CDN origins. Or deploy an Ion instance now — at €0.0056 an hour, an afternoon of experimenting costs less than a coffee.

Ready in under a minute

Deploy your first server now.

No contracts, no minimums. Start on an Ion KVM VPS at €0.0063 an hour and move to a monthly bare-metal server the day you outgrow it.

$ voxa deploy --plan ion --location amsterdam

KVM VPS billed hourly, capped monthly · Dedicated billed monthly · No setup fee

Included on every plan
Free IPv4 + IPv6
Every VPS
Unmetered traffic
1–10 Gbps
DDoS mitigation
2.5 Tbps
Root / IPMI access
Included
Setup fee
€0.00
Minimum term
None