vCPU vs Physical Cores: What the Difference Means for You
What a vCPU really is, how it relates to a physical core, and how to read CPU specs when you are sizing a server.
Every VPS plan lists a number of vCPUs, and almost nobody is told what a vCPU is or how it compares to the cores in their own computer. It is a simple idea wrapped in a confusing name. Here is what it actually means and how to use it when picking a plan.
Start with a physical core
A physical CPU core is a real processing unit on the chip. A modern server processor has many of them, and each core can work on tasks independently. More cores means more things can genuinely happen at the same time.
There is also a feature on most server CPUs that lets a single physical core handle two streams of work at once, which the system sees as two separate processors. These are called threads. So one physical core often presents as two logical processors.
Hold onto that, because it is where the vCPU comes from.
What a vCPU is
A vCPU, or virtual CPU, is the unit of processing power your VPS is given. In practice, a vCPU maps to one of those logical processors, usually a thread on a physical core, that the host assigns to your virtual server.
So when a plan says 4 vCPUs, it means your VPS has been allocated four of the host machine's logical processors to use. On a good host those are real, allocated slices of the physical CPU, not a number invented out of thin air. Your VPS can run four streams of work in parallel.
The short version: a physical core is the hardware, and a vCPU is the slice of that hardware handed to your virtual server.
Why this is not a trick
People sometimes assume vCPU is a watered down core. On a properly run host, it is not. The physical CPU is genuinely powerful, and the threads it exposes are real processing capacity. What matters is that the host does not oversell, meaning it does not promise more vCPUs across all its customers than the hardware can actually deliver. When a provider sizes things honestly, a vCPU gives you the performance you would expect from a slice of that processor.
This is one of the quiet differences between a cheap, oversold host and a good one. The spec sheet can look identical while the real experience is night and day, purely because of how many customers are stacked onto the same cores.
How many vCPUs do you actually need
This is the useful part. More vCPUs help when your workload does many things at once. Fewer are fine when your workload is light or mostly does one thing at a time.
A rough guide:
- A small website, a bot, or a simple service is happy with 1 to 2 vCPUs
- A busier web app, or a game server with a moderate player count, suits 2 to 4
- Heavy workloads, large game servers, several services on one box, or anything doing constant parallel processing, want 4 or more
The thing to remember is that for most real workloads, having enough cores matters more than chasing raw clock speed. A handful of vCPUs working in parallel beats one very fast core for the kind of jobs servers usually do, because those jobs are made of many small tasks happening together.
A quick word on RAM
vCPUs rarely act alone. If you add more processing so your server can do more at once, it usually needs the memory to match, because more simultaneous work means more things held in RAM at the same time. When you size up, size both together rather than maxing one and starving the other.
How to read a plan now
Next time you look at a VPS plan, the vCPU number is telling you how many parallel streams of work your server can handle. Match it to your workload rather than buying the biggest number available. Start with what fits, and because resizing a VPS is quick, you can add more later the day you actually need it.
Our VPS plans allocate real AMD Epyc vCPUs, so the number on the plan is the performance you get. If you are weighing two sizes, the VPS page lists the full specs side by side so you can match cores and memory to what you are running.