Dedicated CPU VPS: What You Are Really Paying For

What a dedicated CPU VPS actually reserves, how CPU steal tells you whether you need one, and a live side-by-side of shared and dedicated plans.

By NoBull Networks Published 6 min read

A dedicated CPU VPS is a virtual server with processor cores kept for you alone. Other customers on the same machine cannot use them. That matters when your work keeps the CPU busy for hours, or when a slow reply costs you money. For a typical website it is often money you do not need to spend. This post explains what "dedicated" means, how to check whether you need it, and what we sell under that name.

Shared vCPU vs dedicated CPU#

A vCPU is the virtual processor your server sees. On a shared plan, the host's scheduler (the part of the host that decides whose work runs next) hands out time on physical cores to every server on the node. Most of the time nobody notices, because most servers are idle most of the time. On a dedicated plan, specific cores are pinned (permanently assigned) to your server, and no other customer's work runs on them.

Shared vCPUDedicated CPU
Who uses the coresYou and other tenants, in turnsOnly you
Short burstsFast, when the node has headroomFast
Hours at full loadDepends on the neighbours and the host's limitsSteady
CPU stealSome is normalShould be near zero
Price per coreLowerHigher

How to tell if your VPS needs a dedicated CPU#

Linux measures the problem for you. Steal time is CPU time your server asked for but the host gave to another server at that instant. Run vmstat 1 10 and read the last column, st. Our YABS performance guide covers the full test, and its rule of thumb holds: single digits under load are fine, and sustained double digits mean the node is oversold or you have outgrown sharing.

Steal is only half the picture. Ask how your work uses the processor:

  • Bursty and interactive. A website, an API, or n8n spends most of its life waiting and then works hard for a second. A fast shared core handles that well.
  • Sustained. Video encoding, build servers, and CI runners (machines that compile and test code on every change) run the CPU flat out for long stretches. Sharing hurts here.
  • Multi-tenant. A cPanel or DirectAdmin server hosting other people's sites cannot let one busy customer slow the rest. Our control panel guide puts resellers and agencies on a dedicated plan for that reason.
  • Latency-sensitive. A production database or a real-time system cares about its slowest responses, not its average. Dedicated cores make the slow ones look like the rest.

Is dedicated vCPU the same as a dedicated core?#

Not always, and this is where the fine print hides. Modern server processors run two hardware threads on each physical core, and many providers count one thread as one vCPU. A "dedicated vCPU" can therefore be half a physical core, with the other half sold to someone else. That is not dishonest, but it is worth asking any host, including us, which one you are buying.

Our VDS page describes the D line as AMD EPYC threads pinned to your server, with DDR5 ECC memory and NVMe RAID-10 storage reserved for you alone. Virtual dedicated server resources explained walks through each line of that spec sheet.

What you give up on a shared plan, and what you keep#

A shared plan is only as good as the host's discipline. We stop selling on a node once it reaches 60% average CPU load, on every line, and the reasoning is in how our stock control works. That is why our Performance VPS line handles bursty work well without dedicated cores: a published Geekbench 6 run on its AMD EPYC 4565P at 4.5 GHz scored 2,323 single-core. The cap keeps the headroom there; it does not reserve a core for you. Dedicated cores do.

What a dedicated CPU costs, side by side#

Our plan codes use the number for the size class, so a P40 and a D40 carry the same memory. Here are both lines at each shared size, read live from our plan data:

PlanCPURAMStorageTransferMonthly
P40 Performance VPS2 shared vCPU8 GB100 GB16 TB$32.00
D40 VDS2 dedicated8 GB160 GB32 TB$64.00
P50 Performance VPS3 shared vCPU12 GB150 GB24 TB$48.00
D50 VDS3 dedicated12 GB240 GB48 TB$96.00
P60 Performance VPS4 shared vCPU16 GB200 GB32 TB$64.00
D60 VDS4 dedicated16 GB320 GB64 TB$128.00
P70 Performance VPS6 shared vCPU24 GB300 GB48 TB$96.00
D70 VDS6 dedicated24 GB480 GB96 TB$208.00

The VDS costs more per core, and it also carries more storage and transfer at the same size. The difference buys isolation. If your steal time is low and your work is bursty, the Performance plan at the same size is the better deal.

Moving from a VPS to a dedicated CPU#

Start on a shared plan unless you already know your workload runs hot for hours. Watch steal and CPU load in the Cloud Panel usage graphs during your busiest week. If the CPU is pinned high and steal is climbing, move to the VDS at the same size class so your memory stays the same. If memory is the constraint instead, read how much RAM a VPS needs first; a dedicated core will not fix that.

FAQ#

What is a dedicated CPU VPS?
A virtual server whose cores are reserved for it and pinned so no other customer's work runs on them. It keeps virtual machine conveniences like rescue mode, reinstalls, and snapshots.

Is a dedicated CPU VPS the same as a dedicated server?
No. A dedicated server is a whole physical machine. A dedicated CPU VPS reserves part of one, so you get predictable compute without buying or managing the hardware.

Dedicated vs shared vCPU: which is faster?
For short bursts on a well-run host, often neither; they run on the same processors. The difference shows under sustained load, where dedicated cores stay steady and shared ones depend on the neighbours.

Do I need a dedicated CPU for a game server?
Usually not for one server. Our game plans run on nodes capped at 60% load. For a fleet of servers, or a game server beside other busy services, a VDS makes sense; game server hosting vs a VPS covers the trade.

See it for yourself. Performance VPS on AMD EPYC, stock-controlled so the specs on the card are the specs you get.