Partial-rack vs whole-cage decommissioning: which one is your data centre actually doing?
Publicly-listed ITAD companies have minimum-project thresholds. A 500-unit whole-cage pull makes economic sense; a 50-unit partial-rack pull doesn't. This gap leaves mid-market data-centre tenants under-served. This explainer walks through the economics of partial-rack vs whole-cage, what each one actually involves, typical timelines, and when to choose which.
Why volume players skip partial-rack work
A large ITAD prime contractor has overhead: facility costs, insurance, staff, equipment, Customs filings. That overhead scales. A 1,000-unit job amortizes it easily. A 50-unit job doesn't.
Typical economics: 1,000 units = 2–3 weeks work, S$5,000–7,000 logistics cost, S$1,000–2,000 destruction cost. Gross margin: S$8,000–9,000. That's worth a project manager and a two-person team.
A 50-unit partial-rack job: 2–3 days work, S$1,500–2,500 logistics, S$300–500 destruction. Gross margin: S$1,800–3,000. Not worth deploying a team. Minimum project sizes are 200–500 units at most volume vendors.
Mid-market data-centre tenants often have 50–200 unit retirements. They're too big for local recyclers (who handle 5–20 units), too small for volume vendors. That gap is where Maxicom operates.
What a partial-rack pull actually involves
Here's the work scope for a 50–200 unit partial-rack decommission in a co-lo facility.
- Pre-work site survey · Walk the cage, identify which racks have equipment to retire. Confirm power, cabling, and any network dependencies. Coordinate with facility ops and neighbouring tenants.
- Pre-pickup inventory · Physical count and serial verification. Match CMDB records to reality. Flag any discrepancies before uplift.
- Scheduled downtime · Unlike a whole-cage pull (you turn off everything), a partial-rack pull requires careful timing. You're pulling servers from active cage(s) while neighbours' kit stays live. Coordinate downtime window with your tenant operations team.
- De-cabling and de-racking · Disconnect power, network, SAN (if applicable). Remove mounting hardware. Wrap cables. Palletise or box equipment per weight / fragility.
- Photo documentation · Pre-pull, in-progress, post-pull (empty cage / open tile space). Facility loves this — clear evidence the space was returned clean.
- Locked transit · Equipment is sealed in a locked, GPS-tracked vehicle. Transit time: 2–4 hours within Singapore (depending on facility location and traffic).
Partial-rack vs whole-cage at a glance
Partial-rack (50–200 units)
Multi-day work, selective equipment removal, coordinated with live cage operations, suited for tenant refresh or equipment rotation, flexibility on timing, higher per-unit touch.
Whole-cage (500+ units)
1–2 day event, full facility power-down, single scheduled window, suited for end-of-lease or complete refresh, fixed timeline, economies of scale on per-unit cost.
Logistics cost per unit
Partial-rack: S$30–60/unit. Whole-cage: S$5–15/unit. Partial pays higher per-unit logistics; whole-cage spreads cost.
Destruction timeline
Partial-rack: 5–10 business days (batched with other small jobs). Whole-cage: 2–5 business days (dedicated run). Whole-cage is faster.
When to choose partial-rack vs whole-cage
Simple rule: depends on your volume and timeline.
- Choose partial-rack if: · You have 50–500 units to retire. You're in the middle of a lease or operating the facility. You can't afford downtime for a whole-cage pull. You need flexibility on timing (not a fixed 'power-off' date).
- Choose whole-cage if: · You have 500+ units. You're exiting a lease entirely (decommissioning the entire facility's IT). You can schedule a fixed power-down window. You want economies of scale (lower per-unit cost).
- Hybrid approach: · Pull 80% whole-cage on a scheduled date (saves cost). Handle the remaining 20% as partial-rack over the following weeks (flexibility for stragglers, isolated systems, or phased retirement).
Typical partial-rack decommission timeline
- ♦ Week 1 – Monday: Site survey, CMDB reconciliation, downtime window agreed with facility ops.
- ♦ Week 1 – Wednesday: Pre-pull inventory photo, asset tags verified, staging area prepped.
- ♦ Week 1 – Friday: Uplift window opens. De-cabling, de-racking, photo-documented. Sealed crate ready for pickup.
- ♦ Week 2 – Monday: Transit to Maxicom facility. Arrival inspection, inventory recount, asset condition assessment.
- ♦ Week 2–3 – Tue–Thu: Data destruction per device type (wipe, shred, etc.). Per-device logging. Witness sign-off.
- ♦ Week 3 – Friday: Destruction complete. Certificate of Destruction issued, covering all devices and methods.
- ♦ Week 3 – Friday: Settlement invoiced. Buyback proceeds (if any) applied to master invoice within 5 business days.
A 120-unit partial-rack refresh
Scenario: A SaaS company in Equinix SG3 is refreshing 120 of 300 servers in their cage. The refresh is a phased 3-month plan: retirement, procurement, staging, deployment. The company can't power down the entire cage.
Approach: partial-rack pull
- Week 1: Site survey. We identify 120 servers in Racks A, B, C (left side). Racks D, E, F (right side) stay live. Facility okays a Wednesday 2am–6am downtime for Racks A–C.
- Week 1: Pre-pull inventory. Serials matched to CMDB. Three units don't match (likely decommissioned months ago, still sitting); facility confirms they can be removed.
- Week 2: Wednesday pull. 2am: we start. Power racks A–C. Disconnect SAN (cross-check: no active LUNS on this subset). Disconnect network. De-cable. De-rack. 6am: done. Crate sealed. Facility confirms tile space is clear.
- Week 2: Thursday morning. Equipment loaded on truck, GPS tracking on. Equipment arrives Maxicom at 11am.
- Week 2 – 3: Destruction. 110 servers get NIST Clear (wipe). 10 old servers (legacy RAID controllers, encrypted) get Destroy (shred). Logging per device. Witness sign-off.
- Week 3 Friday: Certificate issued. 120 serials, 110 wipe entries, 10 shred batch ID entries. Residual material routed to recycler. Signed.
- Invoice: logistics S$3,000 + destruction S$2,000 – no residual value (old hardware). Net cost: S$5,000. Company's 3-month refresh plan stays on track.
Maxicom Singapore — frequently asked
Is a partial-rack pull more expensive per unit than a whole-cage pull?
Yes, typically. Whole-cage: S$5–15/unit logistics. Partial-rack: S$30–60/unit. The difference is that a whole-cage pull amortizes fixed costs (project manager, coordination overhead) across 500+ units. A partial-rack amortizes across 50–200. Lower volume = higher per-unit cost. However, our per-unit pricing is still competitive with general recyclers because we're optimized for mid-market work.
We have 80 servers to retire, but we're uncertain about exact timing. Can you hold a price quote while we finalize the date?
We'll quote valid for 4 weeks if you commit to a rough window (e.g., 'anytime in June'). If the actual pull date slips beyond 4 weeks, we re-quote (market prices may have changed). Once you confirm the exact pull date, the price locks and you're scheduled.
For a partial-rack pull, how do you avoid disrupting our live systems in adjacent racks?
We work closely with facility ops and your team on the power layout. We only power down the specific racks you're retiring — not the entire cage. We pre-verify there are no SAN/network dependencies between your retiring racks and live racks. In the rare case where there's a cross-rack dependency, we coordinate the pull to avoid live-system downtime. Facility ops are always present during the pull.