Short answer: A vCPU is a virtual processor assigned to a VPS, normally one hardware thread of a physical CPU core (two vCPUs per core with hyper-threading). On shared plans vCPUs are time-sliced with other tenants; on dedicated plans they are reserved for you.
vCPU explained
Providers count vCPUs differently. Most map one vCPU to one hardware thread; a few map it to a full core. Clock speed and CPU generation matter as much as the count: two vCPUs on a modern AMD EPYC or Ryzen host can outrun four on an old Xeon.
On shared-CPU plans the host oversubscribes threads, so sustained load from neighbours shows up as steal time in your server, and many hosts throttle customers who run at 100% for long periods. Dedicated-vCPU plans pin threads or cores to your virtual machine and cost two to four times more; they are the right choice for databases, game servers and CI builds.
See it in the data
Frequently asked questions
Is a vCPU the same as a core?
Usually not: a vCPU is typically one thread, and a physical core runs two threads with hyper-threading.
How many vCPUs do I need?
One is enough for a small website, bot or VPN; two to four for busy sites, Docker stacks and game servers; more for builds and databases.