Skip to content
Buying basics

New vs Refurbished vs Used IT Hardware

What new, refurbished and used mean for servers, drives, memory and switches, and how to weigh price, warranty, lead time and risk for each workload.

Updated 4 min readBy the NorthernITShop team

New hardware is unused, refurbished hardware has been tested, cleaned and graded, and used hardware is pulled from working systems and function-tested. The right choice depends on the workload, how long you need the part to last, and how much risk the system can take.

What each condition means

Condition What it is Typical source
New Unused, in original or bulk factory packaging Distributors, manufacturer channels, surplus stock
Refurbished Tested, cleaned and graded; fully functional Decommissioned systems and returns, processed before sale
Used Pulled from working systems and function-tested; may show cosmetic wear Data center pulls and upgrades

Terms vary between sellers. "New" can mean factory sealed, new in open box, or new bulk packed without retail boxes. "Refurbished" can mean anything from a basic test to a full rebuild. Always check what a seller's definition covers, and what warranty comes with it. Our definitions and warranty terms are on the warranty page.

Price, warranty and lead time

Each condition is its own offer, with its own price, warranty and lead time.

  • New usually costs the most. Lead time depends on distribution stock and can be long for constrained parts.
  • Refurbished usually costs less than new and often ships faster for parts the manufacturer no longer makes.
  • Used usually costs the least. It suits spares and non-critical systems.

Your quote states the condition, warranty and lead time for every line, so you can compare options side by side.

When new makes sense

  • New deployments that will run for many years.
  • Systems under OEM support, where the manufacturer expects genuine new or authorized parts.
  • Write-heavy SSDs, where remaining endurance matters.
  • Regulated environments with procurement rules that require new equipment.
  • Current-generation parts with good distribution availability.

When refurbished makes sense

  • Expanding or matching an existing estate, where you need identical models.
  • Systems past their OEM warranty that you plan to keep running.
  • Lab, test, development and disaster recovery environments.
  • End-of-life parts that are no longer available new.
  • Budget-limited projects, where the saving pays for more capacity.

When used makes sense

  • Spares kits, where you keep parts on the shelf for fast swaps.
  • Non-critical systems, test benches and training labs.
  • Passive or low-wear parts such as rails, bezels, fan modules and cables.
  • Very old platforms where used pulls are the only source.

Risk by part type

Not every part wears the same way. Some parts are low risk in any condition; others need more care.

Part type Wear factors What to check
Hard drives Moving parts, power-on hours SMART health, reallocated sectors, power-on hours
SSDs Flash endurance Remaining life or percentage used, total bytes written
Memory Very little wear Correct type, rank and speed; memory test result
Processors Very little wear Bent socket pins (for LGA sockets), correct stepping
Power supplies and fans Capacitors and bearings age Function test under load, fan noise
Switches Fans and power supplies Port tests, fan and PSU status, software and license state
Servers All of the above Full power-on test, firmware level, error logs

For SSDs and hard drives, ask for health data if you need it. Drive condition matters more than any other part type. See our guide to SAS, SATA and NVMe drives.

Questions to ask any seller

  1. What exactly does each condition mean?
  2. What testing is done before shipping?
  3. How long is the warranty, and when does it start?
  4. Are data-bearing parts sanitized, and to what standard?
  5. Does the part include its carrier, bracket or heatsink?
  6. Is firmware updated, or left as found?
  7. What happens if the part arrives dead?

Our answers are on the warranty, returns and shipping pages.

A practical approach

Many buyers mix conditions in one order. For example, a new SSD for a database server, refurbished memory to expand an existing host, and used drive carriers or rails. Ask for more than one condition in the same RFQ when you are not sure. Your quote shows the options for each part so you can decide line by line.

Summary

  • Choose new for long life, write-heavy storage and systems under OEM support.
  • Choose refurbished to extend, match or maintain existing systems at lower cost.
  • Choose used for spares, labs and low-wear parts.
  • Whatever the condition, confirm the warranty and the test process before you buy.

Browse servers, storage drives, server memory and networking switches, then send an RFQ with the conditions you want to compare.

All guides →
  • What EOL and end-of-support notices mean, how to find replacement part numbers, and how to source spares for legacy servers, storage and switches.

    4 min readRead guide

  • How to write an RFQ or BOM for servers, drives, memory and networking parts - part numbers, quantities, conditions, dates, country and Incoterms.

    5 min readRead guide

  • How HPE option and spare numbers, Lenovo option and FRU numbers, Dell part numbers and Cisco PIDs work, and where to find them on a part label.

    5 min readRead guide