What Happens to the Hardware When a Data Center Moves?

A typical rack power consumption is less than 8kW, according to the Uptime Institute’s 2025 survey. Some operators, however, have cabinets exceeding 100.

Now picture unbolting a few hundred of either and driving them across town. Firms do it to consolidate, to jump into colocation, to chase the cloud, or because the old hardware has gotten embarrassing.

Whatever the reason, somebody ends up holding a screwdriver and a very long list of servers, drives, switches and GPUs. Every item gets moved, upgraded or retired.

Sounds simple. It really isn’t, and the reasons are mostly physical, which is the angle most move checklists skim past. How each call gets made is where it gets interesting.

Issues During a Move

Most people love the cutover time. It’s clean and can be put into a booking. Issues arise due to limited downtime, apps relying on each other, network changes during the move, and the heaviness and fragility of the hardware.

Beyond the hardware itself, the biggest data center migration risks usually come from poor planning and unmapped dependencies. For example, you move a storage array and everything looks perfect. However, on Monday, the backup is still pointing at the old IP address.

Physical issues can also cause problems. The rails don’t match up, the rack is a different depth, and your cable runs are short by 2m. These are all small issues but can take up a lot of time.

No one can really be blamed. Documentation will always degrade quicker than the hardware, and everyone is so busy they won’t be able to update it.

Transporting Fragile Hardware

Shock, vibration, static, and temperature will all cause damage to your kit. The main concern will be drives. According to enterprise spec sheets I found whilst doing my research, 3.5″ drives can take between 200–300G when powered off.

However, when powered on, they can only take around 40–70G. This means many teams will remove their drives and box them up. Usually, some high-density storage chassis will have specific requirements from the vendor, so be sure to check this first.

Take photos of the racks. Label both ends of all cables. Note which port goes to which switch. Boring? Yes. But it’s the cheapest form of insurance you can do. You don’t want to be trying to recreate a rack from memory!

Vibration can cause some issues. Your truck doesn’t need to crash to cause issues. After hours of vibrating, your cards and cables can become loose. This is why many choose an air-ride suspension.

High-value systems should be properly packaged and transported in climate-controlled trucks. One thing many forget is that cold kit can sweat when rolled into a warm and humid room. Let it settle.

Should I Move, Upgrade, or Dispose of?

Moving can change your refresh strategy. As you’re paying to deal with all boxes, it makes an old box look less like a good deal.

Older CPUs, inefficient power supplies (Bronze units will waste more heat than a Platinum unit), and slow storage can all incur more costs when shipping, racking, and cooling than it would to replace.

What decides it? Age, warranty, power efficiency, performance per watt, and whether the thing suits the new site’s power and cooling. A rough starting point:

Hardware Move It If Replace It If
Servers Recent, in warranty Old, power hungry, unsupported
Storage Arrays Healthy, supported Aging disks, costly support
Networking Fast ports, current firmware Slow, unpatched
GPUs Recent, and the new site can feed them The new site can’t power or cool them

Warranties can be a big factor. Ensure you know if your support will still be valid after changing location. Varies by vendor. Use the below as a talking point. Use your numbers.

Power & Cooling

As mentioned earlier, you may have that 100kW cabinet, usually due to GPU servers. Older halls won’t be designed for this. Ask yourself if the new location can provide the amps, remove the heat, and support the weight.

Air cooling can struggle with higher densities, so many operators will invest in Rear Door Heat Exchangers or Chip-Level Liquid Cooling. Ensure you know what circuits, connectors, and floor loading you have.

Liquid cooling isn’t a simple add-on. You’ll need plumbing, coolant, and people who know how to maintain it. Nothing can be done overnight.

The issue is you may have a perfect move, but will only be running at half speed.

Disposal

Drives will be sanitised. This is covered by the NIST SP 800-88, which has 3 levels: Clear, Purge & Destroy. This was updated in September 2025 (Revision 2). Flash drives shouldn’t be wiped as this can be unreliable.

The guidelines suggest using cryptographic erasure or physically destroying. Keep the serial numbers and certificates. If you wipe it, you can’t prove it’s been wiped. Clear & Purge will leave the drive usable; however, destroy will dispose of it. Decide if you’ll need it in future.

What happens to your kit? The cheapest option will be reuse. Have a look at this first. Then you can refurbish and resell, strip for spares, and certified recycling. Recycling will be the most important at the end of the line.

In 2022, there was 62 million tonnes of e-waste, according to the UN-backed Global E-waste Monitor. Of this, only 22.3% was documented to have been correctly recycled.

A lot of refurbished enterprise kit ends up in home labs! Used rack servers, SAS drives, and 10GbE cards. Perfect for learning. Just check the power consumption and fan noise and wipe your drives.

Final Thoughts

This is as much a hardware task as an IT task. Plan out your shipping, refresh, power, and disposal. If you do this correctly, your transport will be the least exciting part.

Have you ever taken any old enterprise kit for your home lab? What was the best thing you found? What would you have done differently?

Add a Comment

Leave a Reply

Your email address will not be published. Required fields are marked *