Definition
A dedicated server is a physical machine rented to a single customer, with no other tenants sharing its CPU, RAM, storage or network. The provider owns the hardware, rack space, power and cooling; you get root or administrator access to configure the operating system, software stack and network settings as if the machine sat in your own office. Typical builds use Xeon or EPYC processors, NVMe or SSD storage, and dedicated uplinks rather than shared bandwidth pools. This suits high-traffic sites, large databases, compliance-sensitive workloads, game servers and any project needing predictable performance or specific hardware that virtualisation cannot guarantee, since there is no noisy-neighbour risk from other tenants competing for resources. Prices start around USD 80 a month for entry-level builds and scale into the thousands for high-end specifications. Provisioning takes longer than a VPS, often hours to a day, and scaling means swapping hardware rather than resizing an instance. Leading providers include OVHcloud, Hetzner, Liquid Web and Hivelocity.
How it works
The provider supplies the physical machine, the rack space, the power, the cooling and the network. You pick the CPU (often Xeon or EPYC), the RAM, the storage (NVMe, SSD, HDD), the network speed, and the operating system. The hardware is yours alone for the duration of the contract.
Why it matters
For workloads that need consistent performance with no noisy-neighbour risk, specific hardware (NVMe at a certain spec, a particular CPU generation), or compliance that benefits from physical isolation, dedicated is the right step up from VPS. The trade is higher cost and longer provisioning times than virtual machines.