A data center move forces a question most teams avoid for years: of everything in these racks, what actually deserves to make the trip?
The answer is rarely “all of it.” Some gear is too old to trust on a truck. Some is surplus. Some is worth more sold today than it will be tomorrow. And some should move, because replacing it costs more than shipping it.
The timing is right for this conversation. According to market research, the data center decommissioning market is projected to grow from about $12.95 billion in 2026 to roughly $19.94 billion by 2032. AI workloads are part of the reason. AI racks can draw 20 to 100 kilowatts, compared with the 5 to 10 kilowatts older facilities were built for, and industry sources report AI environments refreshing hardware every 18 to 36 months. Many organizations can’t upgrade in place, so they relocate.
Here’s a 90-day plan to do it without losing money or data.
Day 90 to 75: Decide Whether, Where, and Why
Start with the reason. Common ones include:
- A lease ending or a landlord raising rates
- Power or cooling limits, especially for GPU and AI workloads
- Consolidating after a merger or closing a site
- Moving to colocation or the cloud
- Better disaster recovery or security
- Lower operating costs
Then pick the path. You have three: relocate the equipment to a new facility, migrate workloads to colocation or the cloud, or consolidate into less space. Often it’s a mix. Our partner’s cloud migration services can help if part of the plan moves off-premises.
Before you commit, confirm the new site can actually take your load. Check available power per rack, cooling capacity, floor space, network connectivity, and access rules. Then set your downtime tolerance, because that decides how the cutover works.
Day 75 to 60: Inventory and Triage
This is the most valuable step, and the one teams rush. Build a serial-numbered inventory of everything in scope: servers, storage, networking gear, tape, memory, and drives. Record make, model, configuration, age, condition, and support status. Then map dependencies, so you know which systems rely on which.
Next, sort every asset into one of four buckets:
Move. The gear is healthy, still supported, fairly recent, and expensive to replace. It’s worth the trip.
Replace. The gear works, but moving it isn’t worth the risk, or you want a refresh. You can buy refurbished or new replacements for the new site.
Sell. The gear is surplus, redundant, or valuable on the resale market right now.
Retire. The gear has no resale value and needs secure data destruction and recycling.
Support status often decides the bucket. For example, Microsoft’s final paid security updates for Windows Server 2012 R2 ended on October 13, 2026. Hardware lifecycles matter too. See how they play out in our guides to the Cisco UCS B200 M5 and the HPE ProLiant ML350. Equipment without vendor support is a poor candidate for a move.
Also consider the move itself. Drives that have run for years can fail when powered down and transported. Take current backups, and test that you can restore them, before anything leaves.
Day 60 to 45: Sell What You Won’t Move
Here’s the step most relocation guides skip. Selling surplus equipment before the move does three things:
- A lighter move. Fewer racks to pack, ship, and install.
- Lower cost. Less freight, less labor, and fewer hours in the new site.
- Proceeds that offset the project. That money can cover part of the move.
Timing matters. Industry benchmarks show server resale values running well above their long-term averages, and recent market reports show server memory prices roughly doubled from a year ago. Some reports say high-end AI accelerators can lose a large share of their value within weeks of a retirement decision. Equipment you’ve already decided to retire is worth more sooner.
Sell in Two Waves
Don’t sell everything at once. Production gear has to keep running until cutover.
- Wave 1: Sell now. Surplus, spare, redundant, and already-decommissioned equipment.
- Wave 2: Sell after cutover. The old site’s production gear, once the new environment is validated.
For a quote, send us the make, model, and part numbers, plus quantity, condition, and configuration (for servers, the CPUs, memory, and drives). Photos help. We buy servers, storage, networking gear, memory, CPUs, tape, and full data center lots. Our server buyback guide explains the process.
Day 45 to 30: Protect Your Data
Every device that holds data needs a plan, whether it’s moving, selling, or retiring.
Equipment you’re selling or retiring. Sanitize it following current NIST SP 800-88 Revision 2 guidance, using certified data erasure or physical data destruction. Keep chain of custody records and get a Certificate of Destruction for each device. Our Phoenix Certified process handles this for equipment we buy.
Equipment you’re moving. Data travels with it. Use encryption where you can, track every asset by serial number, and use tamper-evident seals and secure transport. Some teams ship drives separately from chassis, or pull them entirely. See our secure shipping details.
Day 30 to 7: Build the Move Plan
Now write the runbook. It should cover:
- Sequence. Which systems move first, and what must stay up.
- Cutover and rollback. Exactly how you switch over, and how you go back if something fails.
- Labeling. Label every cable, port, and rack, and document rack layouts.
- Network changes. IP ranges, DNS, firewall rules, and circuits at the new site.
- Licensing. Some software licenses are tied to a site or hardware.
- Staffing and communication. Who does what, and who gets updates.
- A change freeze. Stop unrelated changes before the move.
If you can, run a trial with non-critical systems first.
Day 0: Move Day
On the day itself, follow the runbook, and don’t improvise. Photograph racks before you disassemble them. Check equipment off a manifest at pickup and at delivery. Install and cable in order, then test power, network, storage, and applications before you declare success.
Day +1 to +30: Validate, Reconcile, and Close Out
The project isn’t done when the trucks leave.
- Monitor everything for issues, and fix problems quickly.
- Reconcile the inventory. Compare serial numbers expected against serial numbers received.
- Run Wave 2 of liquidation. Sell and retire the old site’s equipment once the new environment is stable.
- Collect documentation. File certificates, chain-of-custody records, and resale statements.
- Close the old site. Handle lease and facility obligations, and recycle anything left through R2v3-certified recycling.
- Update your records, and hold a short lessons-learned review.
If you need hands-on help clearing the old site, our partner DES Technologies offers data center decommissioning, hardware removal, and on-site services. For upgrades along the way, see their IT equipment refresh services and AI infrastructure ITAD.
A Simple Way to Think About Cost
Skip fancy models. For the whole project, ask:
Net cost = move costs + new-site costs + downtime risk − liquidation proceeds − avoided costs
Liquidation proceeds and avoided costs, such as old maintenance contracts and lease payments, are the two terms teams forget. Capturing them can change whether the move pays off. Our partner’s asset recovery and value return programs explain how recovery works.
Frequently Asked Questions
What is data center relocation?
It’s moving servers, storage, networking, and other infrastructure from one facility to another with as little disruption as possible. Many projects also involve selling or retiring equipment you won’t move.
What should I move, and what should I sell?
Move gear that’s healthy, still supported, fairly recent, and costly to replace. Sell equipment that’s surplus, redundant, past vendor support, or valuable on the resale market right now. Retire anything without resale value through secure destruction and recycling.
Can I sell equipment before the move?
Yes, and it’s often smart. Selling surplus gear early reduces what you have to move and offsets project costs. Keep production gear running until cutover, and sell it in a second wave.
How do I protect data during a relocation?
Sanitize anything you’re selling or retiring following NIST SP 800-88 Revision 2, and keep a Certificate of Destruction. For equipment you’re moving, use encryption, track serial numbers, and use secure, sealed transport.
How long does a relocation take?
It varies widely. A small server room might take weeks. A large environment can take many months. A 90-day plan is a reasonable framework for planning, though bigger projects need longer.
What happens to equipment I don’t move?
It can be resold, or securely wiped and recycled. We buy a wide range of used IT equipment, and we process every device with documented data destruction.
Ready to Lighten the Load?
Before your move, find out what your surplus equipment is worth. We keep the process simple:
- Send us your list. Include equipment type, quantity, and condition.
- Get a quote. We review your list and send an offer.
- We cover shipping. We handle labels and logistics.
- We erase the data. Every device is processed securely.
- You get paid, with documentation. A Certificate of Data Destruction comes with your payment.
Get a free quote today, or see our FAQ. Data center operators can also explore our data centers industry page, and smaller teams may like our ITAD best practices for small and medium-sized businesses.
