10 Gbps guaranteed bandwidth

10Gbps Dedicated Servers: One Tenant, One Machine

Ten gigabits is the tier where our bare metal line and our VPS line overlap. What you are actually choosing between here is isolation, not speed.

X-ZoneServers 10 Gbps dedicated servers start at €659 per month for a guaranteed 10 Gbps port — 7 bare-metal configurations with up to 3.24 PB of monthly transfer across 12 EU and US datacenter locations. The port is effectively unmetered: hold the full committed rate around the clock, with no overage charges and no fair-use throttling inside your tier.

659
starting price per month
65.9
per Gbps per month
7
configurations
3.24 PB
transfer ceiling per month

Most of our tiers only have one answer. This one has two, because X-ZoneServers sells a 10Gbps VPS as well as a 10Gbps dedicated machine, and the port figure on the label is identical. What differs is who else is on the hardware. A VPS shares a physical host, a hypervisor scheduler and a network card with other customers, and it is sold on a different bandwidth model. A dedicated 10Gbps server has exactly one tenant, and that tenant sets every policy on the box: kernel version, core pinning, interrupt affinity, memory placement, and who is allowed to know the machine exists. If you can name what that separation is worth to you, this is the right product. If you cannot name it, the honest answer is that you may not need it yet.

The Capacity Is Committed to One Tenant

A guaranteed 10 Gbps port carries 1250 MB/s of payload. Held for a full 30-day month it moves 3.24 PB, serves 2000 simultaneous 1080p viewers at 5 Mbps or 400 4K viewers at 25 Mbps, and clears a 1 TB dataset in 13 minutes 20 seconds. That part is arithmetic, and arithmetic is the easy part. What single tenancy changes is that these figures behave as a floor instead of an average. Nothing else on the machine is queued behind the same network card, competing for the same interrupt lines or waking the same cores, so the throughput you measure at three in the morning is the throughput you can still plan against at peak. Capacity planning stops being a statistical exercise and becomes a division sum.

3.24 PB
Transfer per month
Sustained volume at full port utilisation across a 30-day month.
2,000
1080p viewers
Simultaneous streams at 5 Mbps each.
400
4K viewers
Simultaneous streams at 25 Mbps each.
13 min 20 s
Per terabyte
Wall-clock time to move 1 TB off the server at line rate.

Figures use decimal units (1 Gbps = 125 MB/s) across a 30-day month, at 5 Mbps per 1080p stream and 25 Mbps per 4K stream. They describe what the port can carry, not a guarantee of application throughput: disk, CPU and the far end all apply their own limits.

7 configurations at 10Gbps

Dual Xeon E5-2630v3, 16c/32t at 2.4GHz with 128GB DDR4, opens the tier at €659. Dual Xeon E5-2690v4 adds cores at €679 and is the most popular pick here, while the 44c/88t E5-2699v4 with 256GB sits at €749. The single socket EPYC 9254, €1169, gives you 24c/48t at 2.9GHz on DDR5-4800 inside one NUMA node, which is the simplest topology to reason about when you are pinning threads. Dual EPYC 9254 at €1399 doubles that into two nodes, Dual EPYC 9554 brings 128c/256t and 512GB for €1679, and Dual EPYC 9754 tops the tier at €2239 with 256c/512t and 1TB of DDR5. Each of them arrives as a whole machine, so the kernel, the interrupt map and the choice of which cores stay idle for the network path are all yours.

Dual Xeon E5-2630v3

Entry Level

659
/month
16c / 32t @ 2.4GHz128GB DDR4 @ 2133MHz
6x SSD Slots10Gbps Guaranteed
AI Optimized Routing Included
Order Now
POPULAR

Dual Xeon E5-2690v4

Popular Choice

679
/month
28c / 56t @ 2.6GHz128GB DDR4 @ 2133MHz
6x SSD Slots10Gbps Guaranteed
AI Optimized Routing Included
Order Now

Dual Xeon E5-2699v4

High Performance

749
/month
44c / 88t @ 2.2GHz256GB DDR4 @ 2133MHz
6x SSD Slots10Gbps Guaranteed
AI Optimized Routing Included
Order Now

Single EPYC 9254

Entry Level

1169
/month
24c / 48t @ 2.9GHz128GB DDR5 @ 4800MHz
10x NVMe Slots10Gbps Guaranteed
AI Optimized Routing Included
Order Now

Dual EPYC 9254

High Performance

1399
/month
48c / 96t @ 2.9GHz256GB DDR5 @ 4800MHz
10x NVMe Slots10Gbps Guaranteed
AI Optimized Routing Included
Order Now

Dual EPYC 9554

Ultimate Performance

1679
/month
128c / 256t @ 3.1GHz512GB DDR5 @ 4800MHz
24x NVMe Slots10Gbps Guaranteed
AI Optimized Routing Included
Order Now
ULTIMATE

Dual EPYC 9754

Maximum Performance

2239
/month
256c / 512t @ 2.3GHz1024GB DDR5 @ 4800MHz
24x NVMe Slots10Gbps Guaranteed
AI Optimized Routing Included
Order Now

What a 10Gbps server is for

Workloads with a separation clause

When a client contract or an internal control set says the data must sit on hardware no other tenant touches, a shared instance cannot satisfy it however fast the port is.

Real-time services that must not jitter

Voice, game and interactive session traffic care about the shape of latency rather than its average. Owning every core and every network queue removes the neighbour that would blur that shape.

Your own hypervisor, your own rules

Run your own virtualization layer on top and you set the NUMA layout, the pinning policy and how far memory is oversubscribed, instead of inheriting decisions made inside somebody else's host.

Steady egress delivery nodes

Cache and delivery endpoints that hold the port busy for hours at a stretch, where the committed tier keeps the invoice known in advance and the packet path is never shared with another tenant's traffic peaks.

What the Isolation Costs

Entry is €659 a month, which works out at €65.9 for each guaranteed gigabit, and seven configurations sit on this tier, running up to €2239. Read that entry figure as the price of exclusivity rather than the price of throughput, because the port speed on its own is also available from the VPS line. The dedicated machine earns the difference when you can point at the specific reason: an auditor who requires separate physical hardware, a latency budget that a co-resident tenant would break, or a tuning profile that only root on real hardware can apply. If none of those describe your workload, and the way the VPS line is sold suits how you buy transfer, we would rather you started there and moved across when the reason actually arrives. Within a tier there are no overage charges, so the monthly figure is the whole figure.

All guaranteed bandwidth tiers

The same hardware range is sold at nine committed port speeds. Compare the entry price of each.

10Gbps dedicated server questions

What actually changes when I move from the 10Gbps VPS to the 10Gbps dedicated server?
The port figure stays the same and everything underneath it changes hands. On the VPS a hypervisor decides when your vCPUs run and other customers share the physical network card. On the dedicated machine there is no hypervisor unless you install one, no co-resident tenant, and the card, its queues and every core belong to your kernel. The VPS line is also sold on a different bandwidth model, so compare the two on tenancy first and on paperwork second.
Our compliance rules require dedicated hardware. Does this tier satisfy that?
A dedicated 10Gbps server is single tenant: the physical machine is allocated to your account alone and no other customer runs workloads on it. That is the property most separation clauses are written around, and it is why reviewers reject shared instances. We are not your legal advisor, so read your own control set, but where the requirement is physical separation rather than logical separation, this product meets it and a VPS cannot. Our administrators can confirm the allocation in writing for your file.
Can a single TCP connection really reach the full 10Gbps?
It can, but not with default settings. One flow only reaches line rate when the send and receive windows are large enough to hold a full round trip of data in flight, which means window scaling, socket buffer limits, the congestion control algorithm, offload settings and interrupt placement all have to suit the path. On a dedicated machine you hold root on the real kernel and the real card, so every one of those is yours to change and nothing resets it underneath you. If you would rather not tune the stack, several parallel streams fill the port instead.
Does latency stay predictable while the port is saturated?
This is the most practical reason people pay for single tenancy. Your committed 10 Gbps cannot be consumed by a neighbour, and more to the point nobody else is raising interrupts, evicting your cache lines or contending for the cores that service the network queues. You can isolate cores, pin the receive queues to them and keep the tail of your latency distribution flat while the link is busy. Our AI-optimized BGP routing keeps the path stable from the outside, and the 99.9% uptime SLA applies to every machine on this tier.
What do NUMA and IRQ control give me in practice?
On a dual socket machine, memory belongs to one socket or the other, and a packet handled by a core on the far socket pays for a cross-socket hop before your application ever sees it. Owning the whole machine lets you place the process, its memory and the interrupt for the queue it reads on the same node, then keep unrelated work off those cores entirely. A guest on a shared host cannot set any of that, because the layout it is shown belongs to the hypervisor rather than to the hardware. It is the difference between hitting a latency target deliberately and hitting it by luck.
Is a 10Gbps guaranteed port the same as an unmetered 10Gbps server?
In practice, yes — with a stronger guarantee. A typical 'unmetered' offer shares the port and throttles under a fair-use policy; our 10 Gbps is a committed rate you can hold 24/7, with no overage charges inside the tier. If you can saturate it, it is yours to saturate.

10 Gbps dedicated servers by location

Same hardware and guaranteed port, deployed from the datacenter closest to your users.

Not Sure Which Side of the Line You Are On

Tell our own system administrators what your isolation requirement actually is and they will say plainly whether the dedicated 10Gbps machine or the VPS line fits. CPU model, memory to 1TB DDR5, storage and operating system are quoted on request, and deployment takes about an hour.

Up to 1Tbps DDoS protectionDeployed in about 1 hour99.9% uptime SLA