Cloud server, VPS or dedicated: what’s the difference?

What a cloud server is

A cloud server is a virtual machine running on a large pool of physical hardware. You pick the size, the operating system and the region, and it is ready in a minute or two. When you no longer need it, you delete it and stop paying.

What makes it “cloud” is less the machine than the platform: a control panel and API to create servers automatically, images and snapshots, private networks, load balancers, object storage and managed services. Providers such as DigitalOcean, Vultr and Akamai Cloud (Linode) are built around this model.

Cloud server vs VPS vs dedicated server

The three overlap, but they differ in how you pay and how much of the hardware you get:

• VPS: a virtual server with fixed resources, usually billed monthly or prepaid for a longer term. Often the most RAM per dollar.
• Cloud server: also virtual, but billed by the hour or second with a monthly cap, easy to resize and automate, with extra services around it.
• Dedicated server: a whole physical machine for you alone, with every core and the full disk speed. Best for steady, heavy workloads such as busy databases.

What you get with a real cloud platform

Hourly billing is the headline. A month of 24/7 running is about 730 hours, and most providers cap the bill at the monthly price, so an always-on server costs the same as a monthly plan while a test server that runs for an afternoon costs cents.

Snapshots let you save the whole server before an upgrade and roll back if it goes wrong. Scaling means moving to a bigger size with a reboot, or adding servers behind a load balancer. Managed databases take backups, updates and failover off your hands. And an API means you can create and destroy servers from scripts, which is how CI pipelines and test environments work.

Some clouds also rent GPUs by the hour. According to the providers’ own pages, DigitalOcean offers GPU Droplets and Vultr offers cloud GPUs, which is useful for AI work you only need occasionally.

When “cloud” is just marketing

Many hosts now call every VPS a “cloud VPS” or “cloud server”. That is not wrong, since the server runs on virtualised hardware, but it doesn’t mean you get hourly billing, snapshots or an API. Before paying extra for the word cloud, check the billing model, whether snapshots and resizing are included, and whether there are services beyond the server itself.

The reverse is also true: if all you need is one server running all month for a website or n8n, a classic VPS with a monthly price may give you more RAM for the same money. Our rules of thumb apply either way: 1–2 GB for a small site or VPN, 4 GB for WordPress or a small Docker stack, and 8 GB or more for shops and databases.

Which one to choose

Choose a cloud server when your usage changes: short tests, projects that grow, automated deployments, or when you want managed databases and many regions. Akamai Cloud (Linode), for example, lists many regions including Jakarta and São Paulo, and Vultr offers regular plans in around 30 locations, according to their pages. Choose a VPS when you want the most resources for a fixed monthly price. Choose a dedicated server when the load is heavy and constant.

The table below shows current plans from cloud providers, including their hourly prices where published. Our server finder can suggest a size for your project.