What is a vCPU?
A vCPU (virtual CPU) is a share of a physical CPU core allocated to your virtual machine. Understanding vCPUs helps you make informed decisions about compute resources.
Shared vs Dedicated vCPUs
Shared vCPU
Multiple VMs share the same physical core. You get a time-slice of CPU capacity.
- Lower cost ($4-20/mo typical)
- Burstable performance
- Variable during peak times
- Best for: dev, staging, low-traffic
Dedicated vCPU
The entire physical core is reserved for your VM. Consistent, predictable performance.
- Premium pricing ($30+/mo typical)
- Guaranteed performance
- No resource contention
- Best for: production, databases
vCPU to Physical Core Mapping
Most providers map vCPUs like this:
1 vCPU = 1 thread (typically half a physical core on hyperthreaded CPUs)
2 vCPU = 1 full physical core (approximately)
4 vCPU = 2 physical cores (approximately)
Note: Exact mapping varies by provider and CPU model.
CPU Types by Provider
| Provider | Typical CPU Types | Notes |
|---|---|---|
| Hetzner | AMD EPYC | Documentation ↗ |
| Vultr | AMD EPYC / Intel Xeon | Documentation ↗ |
| DigitalOcean | Intel / AMD | Documentation ↗ |
| Linode | AMD EPYC | Documentation ↗ |
How Many vCPUs Do You Need?
1 vCPU: Single-threaded apps, static sites, light APIs
2 vCPU: Most web apps, Node.js/Python APIs, WordPress
4 vCPU: Multi-worker apps, background jobs, moderate databases
8+ vCPU: High concurrency, ML inference, video processing
When to Choose Dedicated CPUs
Consider dedicated vCPUs when:
- You need consistent latency (APIs, real-time apps)
- Running CPU-intensive workloads (ML, video encoding)
- Performance variance is unacceptable
- You're on a Performance or Enterprise tier
Sources
CPU information sourced from official provider documentation. Performance characteristics may vary based on specific instance types and regions.