Colocation is renting space in a professional data centre for a server you own. The data centre provides the power, cooling, physical security, and internet connection. You provide the hardware and manage everything on it.
That’s the whole idea. The rest is just details about whether it makes sense for you. For most people reading this, the honest answer is probably not yet.
What the Data Centre Provides (and What It Doesn’t)
A colocation facility is, at its most basic level, a building purpose-built to keep servers running. It supplies:
- Redundant power with UPS systems and backup generators
- Precision cooling to keep hardware at stable operating temperatures
- Physical security: access controls, CCTV, biometric entry
- High-capacity network connectivity to the internet
- Fire suppression systems
What it does not provide is any management of your server. The hardware is yours. The operating system is yours. The software, configuration, and security patches are all your responsibility. The data centre staff keep the lights on and the temperature right. Everything beyond that is down to you or your team.
Some facilities offer remote hands as a paid add-on: a technician at the site who carries out physical tasks on your behalf, such as rebooting a server, swapping a cable, or checking indicator lights. If you can’t be on-site yourself and something needs a physical fix, remote hands is what covers it.
How Colocation Differs From Dedicated Hosting
This is the comparison that comes up most often, and for good reason. Both give you a physical server that isn’t shared with anyone else. The difference is ownership.
| Feature | Colocation | Dedicated Hosting |
|---|---|---|
| Who owns the server | You | The hosting provider |
| Hardware maintenance | Your responsibility | Provider handles it |
| What you pay for | Rack space, power, bandwidth | Server + managed infrastructure |
| Upfront hardware cost | Yes — you buy the server | None |
| Technical staff needed | Yes — in-house or contracted | Not required |
| Best suited for | 20+ servers, dedicated IT team | Most businesses, any scale |
With dedicated hosting, such as what InMotion Hosting offers, the provider owns the server hardware. You rent it. When a drive fails, they replace it. When the hardware ages out, they upgrade it. You pay a monthly fee that covers the machine, its maintenance, and typically some level of managed support.
With colocation, you bought the server. It lives in someone else’s data centre, but it’s yours. When hardware fails, you fix it or pay remote hands to do it. When the machine needs upgrading, you plan and fund that. The data centre charges you only for the facility services: the rack space, power, cooling, and connectivity.
The practical upshot:
- Dedicated hosting suits most businesses. No hardware investment, no maintenance burden, predictable costs, and someone else handles the physical infrastructure.
- Colocation suits organisations that already own enterprise hardware, have the technical staff to manage it remotely, and are running at a scale where owning the equipment makes financial sense.
A commonly cited benchmark: colocation typically starts making more financial sense than dedicated hosting when you’re running 20 or more servers and have a dedicated infrastructure team. Below that threshold, the economics rarely work in colo’s favour once you factor in hardware purchase, maintenance, and remote management overhead.
What You’re Actually Paying For
Colocation pricing has several components. Understanding each one before you compare quotes saves a lot of confusion.
Rack space is billed by the rack unit (U). One rack unit equals 44.45mm of vertical space in a standard 19-inch cabinet. A single 1U server is a slim, flat machine that occupies one slot. A full standard rack holds 42U. Pricing scales accordingly.
Market rates as of mid-2026:
| Configuration | US (monthly) | UK (monthly) | Europe (monthly) |
|---|---|---|---|
| 1U (budget provider) | From ~$75 | From ~£45 | From ~€50 |
| 1U (Tier III+ facility) | $150–$300 | £60–£300 | €80–€250 |
| Quarter rack | $300–$600 | From ~£225 | From ~€200 |
| Full rack (42U, standard) | $900–$2,500 | ~£1,200 (outside London) | ~€800–€1,500 |
| Full rack (42U, premium city) | $2,500–$5,000+ (NYC/LA) | ~£2,000+ (Central London) | €1,500–€3,000+ (Amsterdam/Frankfurt) |
All prices exclude VAT and vary significantly by facility tier, power allocation, bandwidth, and contract length. Treat these as market orientation, not quotes.
Power is billed either as a flat monthly rate or metered by actual consumption. A typical 1U server draws 200–400 watts, which translates to roughly $20–$80 per month in power costs depending on the facility, your server’s efficiency, and local electricity rates. European facilities tend to have lower electricity costs than US equivalents, though this varies by country.
Bandwidth may be included in a base allocation or billed separately. Check what’s covered before signing anything. Overage charges on bandwidth can be significant.
Setup fees cover the initial rack and stack work: physically installing your equipment, connecting cables, and getting your server onto the network. These are one-off costs, typically in the range of a few hundred dollars or euros at most providers, though larger enterprise deployments can run higher.
What Data Centre Tiers Actually Mean
When comparing colocation providers you’ll see references to Tier I, Tier II, Tier III, and Tier IV facilities. These classifications come from the Uptime Institute and describe how resilient the facility is to failure.
Tier I is a basic data centre with no redundancy. A single power or cooling failure can take the whole facility offline. Tier II adds some redundancy for power and cooling but still has single points of failure. Tier III is the enterprise standard: N+1 redundancy across all systems, meaning there’s a backup for every critical component. Planned maintenance can be carried out without taking the facility offline. Uptime Institute defines Tier III as supporting 99.982% availability, which works out to roughly 1.6 hours of potential downtime per year. Tier IV goes further with fully fault-tolerant systems, where any single failure or maintenance action cannot affect operations. Tier IV targets 99.995% availability, under 30 minutes of downtime per year.
For most colocation customers, Tier III is the right benchmark. Tier IV commands a significant price premium and is typically used for financial services, government, or mission-critical infrastructure where even brief outages carry serious consequences.
When comparing quotes, always ask which tier the facility holds and whether the certification is current. Self-declared tiers without independent certification are not the same thing.
Do You Need to Visit the Data Centre?
Not for most things. This is one of the more common misconceptions about colocation.
Most server management tasks, including reboots, configuration changes, OS access, and monitoring, can be done remotely via a technology called IPMI (Intelligent Platform Management Interface). This is an out-of-band management interface built into enterprise server hardware that lets you access and control the machine at a hardware level, even if the operating system has crashed or the server won’t boot normally. Think of it as a remote control that works independently of the software running on the server.
Physical presence is only needed when something requires hands on the hardware: replacing a failed drive, swapping a network card, or physically inspecting a problem that can’t be diagnosed remotely. That’s where remote hands services come in. For an hourly fee, a technician at the facility handles the task so you don’t have to travel.
If you don’t have a server management team and can’t easily dispatch someone to a data centre when needed, this is a genuine operational risk to weigh before committing to colocation.
When hardware fails
When hardware does fail, the process typically runs like this. You’ll usually detect the problem first through remote monitoring or an alert from your server management software. You then connect via IPMI to diagnose the issue at the hardware level. If it’s something physical, such as a failed drive or a loose cable, you raise a remote hands request with the facility.
A technician physically accesses your server, carries out the task, and confirms it’s resolved. You verify remotely and bring the system back online. For common failures like drive replacements, a well-run facility can turn this around in under an hour. The key is having spare components either stored on-site at the facility or shipped in advance, because waiting for parts to arrive adds hours or days to any outage.
And it’s worth remembering that facility-level incidents can affect your hardware too, regardless of how well managed your server is. A fire at a data centre in the Netherlands is a reminder that physical infrastructure carries physical risks. It’s one of the reasons geographic redundancy (spreading hardware across more than one facility) is standard practice for organisations where downtime has real consequences.
Who Colocation Is Actually For
Colocation makes sense in a fairly specific set of circumstances. You’re likely in the right position for it if:
- Your organisation already owns enterprise-grade server hardware and wants to move it to a better-connected, more resilient environment than an office server room
- You have technical staff who can manage the server remotely and travel to the site if needed
- You’re running enough hardware that the total cost of colocation is lower than renting equivalent dedicated capacity from a managed provider
- Your industry requires physical ownership of the hardware that processes sensitive data, for compliance reasons such as data sovereignty or specific regulatory frameworks
- Your workload is consistent and predictable rather than variable, making the fixed cost of owned hardware work in your favour
Hetzner, which operates its own data centres in Germany and Finland, offers colocation for organisations needing EU-based infrastructure under GDPR-aligned jurisdiction. That kind of geographic and legal specificity is difficult to guarantee with cloud or managed hosting.
When Colocation Is the Wrong Choice
For most businesses, colocation is the wrong product, at least at their current stage.
If you don’t already own servers, you’re looking at significant upfront hardware costs before you’ve paid for a single month of rack space. Enterprise server hardware is not cheap, and it needs maintenance, replacement parts, and technical expertise to run reliably.
If your team doesn’t include people who can manage a server remotely and respond to hardware incidents, the operational burden shifts from “someone else’s problem” to “your emergency at 2am on a Sunday.”
If your traffic or compute needs vary, owned hardware works against you. A dedicated server or cloud instance can be resized or cancelled. A server you’ve purchased and shipped to a data centre cannot.
For businesses earlier in their infrastructure journey, the decision between shared hosting and VPS is almost always the more relevant question. Colocation sits several rungs above that on the complexity and cost ladder.
Common Questions About Colocation
What is the difference between colocation and dedicated hosting?
The core difference is who owns the hardware. With dedicated hosting, the provider owns the server and handles maintenance. You rent exclusive access to it. With colocation, you own the server and manage it yourself. The data centre provides the facility: power, cooling, connectivity, and physical security. Dedicated hosting requires no hardware investment and less technical overhead. Colocation gives you more control and can be cheaper at scale, but requires both the upfront hardware cost and the in-house expertise to manage it.
What is a rack unit?
A rack unit (U) is the standard measurement for server rack space. One rack unit equals 44.45mm of vertical space in a standard 19-inch server rack. A 1U server is a thin, flat machine occupying a single slot. A full rack typically holds 42U. Colocation providers price their services partly based on how many rack units your hardware occupies.
How much does colocation cost per month in the UK?
A single 1U server in a reputable UK Tier III facility typically costs £60–£300/mo for rack space and basic connectivity, before power charges. Outside London is considerably cheaper than central London. Full racks run from around £1,200/mo outside London to £2,000/mo or more in Docklands or Central London. These figures are market indications and vary significantly by provider, power allocation, and contract terms.
Do I need to visit the data centre to manage my server?
Rarely. Most server management can be done remotely via IPMI, which is a hardware-level remote management interface built into enterprise servers. It lets you reboot, reconfigure, and access the machine even if the operating system has crashed. Physical visits are only needed for hardware replacements or problems that can’t be resolved remotely. Remote hands services at the data centre can handle many of these tasks for an hourly fee, so you don’t always need to travel yourself.
Is colocation suitable for small businesses?
In most cases, no. Small businesses typically don’t own enterprise hardware, don’t have the technical staff to manage a colocated server, and aren’t running at a scale where the economics of hardware ownership make sense. Managed dedicated hosting or a VPS is almost always a better fit until you’re operating at significant scale with consistent infrastructure needs and in-house technical capability.
