Data center decommissioning in Tijuana, Mexico

What decommissioning actually means (and why it is not "hauling the gear out")

Data center decommissioning is the orderly removal of all the infrastructure in a compute site that is going out of service: racks, servers, storage arrays, network gear, cabling, PDUs and battery banks. On paper it reads like heavy lifting. In practice it is a control project: what separates a good decommission from a bad one is not how fast the room empties, but the quality of the paper trail left behind once it is empty.

The reason is simple. A data center concentrates, in a few square metres, the highest density of confidential information in the entire company. When that equipment moves, the information moves with it. And unlike a laptop — which has an owner, an asset tag and someone who will notice — a drive that comes out of a chassis in the middle of a night teardown is missed by no one until it turns up somewhere it should not.

The five triggers behind a decommission

Almost no decommission starts as a decision of its own: it arrives pushed by another project. Recognising which one is yours tells you what will squeeze — the calendar, the budget or compliance:

• Cloud migration. The on-prem site empties in stages. Usually the messiest case: the hardware powers down when the migration finishes, months after someone declared it dead on a diagram.
• Site consolidation. Two or three computer rooms become one. What is left over is working equipment with nowhere to go — not scrap.
• Colocation contract ending. You have a hard date and a penalty if the cage is not handed back empty and clean. Here logistics outranks everything else.
• Hardware refresh. The new generation lands and the old one has to leave without the two sharing floor space for more weeks than necessary.
• Plant closure or relocation. The most delicate case, and the one most often handled badly. It gets its own section below.

In all five the pattern repeats: the teardown shows up at the bottom of someone else's project plan, with the date already on top of it and no budget of its own. That is why so many end up resolved with the first available truck — which is precisely the most expensive way to do it.

The risk nobody quotes: the data is still alive inside the rack

When a decommission is quoted, what gets quoted is labour: how many people, how many days, how many trucks. What rarely appears in the quote is the handling of the storage media — and that is the part that can cost the most.

In Mexico, the personal data your company holds is protected by the Federal Law on Protection of Personal Data Held by Private Parties (LFPDPPP), whose article 63 penalises negligent handling. Fines are calculated in UMAs and escalate when sensitive data is involved. The law does not distinguish between a leak caused by a sophisticated attack and one caused by a server travelling in the back of an unlogged truck with its drive still in it: in both cases the responsibility sits with the data controller — you.

That is why in our process drives and SSDs are pulled from the chassis before the equipment leaves the site, counted against the inventory and kept under control. A server without storage is a metal box. A server with a drive is a rolling record. The method, the sanitization levels and what goes on the certificate are covered in our certified data destruction guide.

Plant closure or relocation: what to do with the IT equipment

This is the scenario handled badly most often, and on the Baja California border it happens frequently. A plant closes, relocates or scales down, and the IT equipment becomes the last item on a list where everything else — staff, customers, customs, the landlord — is more urgent.

There is a layer to it that few see coming: in a closure, the organisation that was supposed to document the process is dissolving at the same time. The IT manager who knew the inventory has already left, the passwords left with them, the person who signs is off payroll, and the landlord wants the building empty on a date that will not move. The equipment ends up leaving with no record, and months later someone at head office asks for the file to close the books or answer an audit, and it does not exist.

The practical advice is a single line: start the IT removal before the people leave, not after. While there is still someone who knows what each thing is and has signing authority, the inventory takes hours. Once there is no one, it is done blind and against the clock.

Our process, step by step

Order matters more than speed. This is how we run it:

1. 1

   On-site walkthrough

   We visit your site and count racks, occupied rack units, servers, storage arrays, switches, PDUs, UPS units and battery banks. We note access, freight elevators, loading docks, weights and how each item physically leaves the building. That is where the real scope comes from — not a hallway estimate.

2. 2

   Teardown plan and maintenance window

   We define the shutdown order, what comes out first and in which window. If the site keeps running in parallel — the norm in a consolidation — we work by zone and by row, not all at once.

3. 3

   Disconnection and asset-level labeling

   We disconnect power and network, pull structured and power cabling, and label every asset with its serial number before it moves. That labeling is what later allows the certificate to state exactly what was destroyed.

4. 4

   Storage media pulled and controlled

   Before a single rack leaves, drives and SSDs are pulled from the chassis, counted against the inventory and placed under control. A server without its drive is no longer a confidentiality problem; a server with its drive riding in a truck is.

5. 5

   Physical destruction of the media, with certificate

   We destroy the storage medium and issue a per-serial-number certificate stating method, date and operator. If your policy requires witnessed destruction, your witness is named on the document.

6. 6

   Removal, disposal and file

   Racks, chassis, cabling and metal leave with documented chain of custody, are recycled under the NOM-161-SEMARNAT framework, and you receive the complete file: inventory, certificates, disposal record and the metrics for your ESG report.

What we decommission

Scope covers the full infrastructure of a compute site and office communications rooms:

• Full racks and cabinets
• Rack, tower and blade servers
• Blade chassis and modules
• SAN / NAS storage arrays
• Switches, routers, firewalls, load balancers
• Structured and fiber cabling
• PDUs and power distribution
• UPS units and backup battery banks
• KVM consoles and monitoring gear
• Tape libraries and backup units
• Telecom equipment, MDF/IDF
• Complete office communications rooms

What we do not do, so there are no surprises mid-project: integrated industrial waste management — solvents, sludge, biohazardous waste, tires — and paper record destruction. Those are different lines of business, with different equipment and different authorizations.

Maintenance windows: running it without taking the operation down

Except in a full closure, a site is almost never decommissioned powered off. The normal case is that part of it keeps serving while another part leaves — and that is exactly where a badly planned teardown causes the outage nobody wanted.

We work by zone and by row, with a shutdown order agreed in writing before anything is touched: first what is already out of production, then what was migrated and is idling, and last what depends on the migration finishing. Heavy manoeuvres — battery banks, loaded cabinets — are scheduled in the widest window, not the most convenient one. And if the window shrinks on the day, what shrinks is that shift's scope, not the procedure.

The file you walk away with

At project close we hand over the full document package — the part that makes a decommission defensible in an audit:

• Outbound inventory with asset type and serial number.
• Destruction certificate for each storage medium, with method, date and operator — plus the witness, if destruction was witnessed.
• Collection record with weights, chain of custody and signatures from both sides.
• Final disposal record for the recycled material.
• ESG metrics: tonnage handled and CO₂ avoided, for your sustainability report.

Who keeps the value of the decommissioned equipment

Worth being explicit, because the industry tends to be vague here. We destroy the storage medium and the hardware stays with you: if you redeploy the servers to another site, donate them or sell them, that income is yours in full and never passes through us.

From what you do not want to keep, we do buy the parts with real material value — motherboards, processors and memory — valued on composition rather than a flat "price per kilo" that cannot tell a pallet of boards from a pallet of cabinets. The rest — chassis, rails, cabling, metal — is recycled under the NOM-161 framework. If the volume comes by the pallet or by the ton, the route that applies is production scrap buyback.

What to demand from whoever decommissions your data center (even if it is not us)

If you are comparing providers, these are the questions that separate the ones doing labour from the ones doing decommissioning:

• When exactly do the drives come out of the chassis? If the answer is "at our facility", the risk rides in the truck. They should come out at your site.
• Is the certificate per serial number or per lot? A certificate that says "150 drives" is no help to an auditor asking about one specific drive.
• Who signs the chain of custody, and against what inventory? There should be a document with weights and two signatures, not a confirmation email.
• Can you witness the destruction? And if you do, is your witness named on the certificate?
• What happens to equipment that still works? Ask directly whether they resell it. If they do, that margin came out of your asset.
• Which Mexican regulation do they document against? Be careful with a provider that only cites U.S. standards or the European WEEE directive: in Mexico what applies is NOM-161-SEMARNAT and the LFPDPPP.
• Which authorizations do they hold — and which do they not? A provider who answers precisely, including what they lack, is more trustworthy than one displaying five logos.

Decommissioning on the Tijuana–San Diego border

Our base in Playas de Rosarito and our operation in Tijuana put us inside the corridor where the most sites are currently moving: maquiladoras consolidating computer rooms, nearshoring operations arriving and standing up new infrastructure, and U.S. companies that need to dispose of their Mexican plant's equipment without shipping it back across the border.

For that last case the advantage is direct: the equipment is disposed of locally and legally, with bilingual documentation your head office can present in its own audit, and without the cost and customs exposure of taking it back. We serve Tijuana, Playas de Rosarito, Mexicali, Ensenada and Tecate. If your operation is cross-border, see our cross-border ITAD page.

Checklist: prepare your decommission

You do not need a perfect inventory to ask for a quote. This is enough for us to put together a proposal with a schedule and a method:

• How many racks and how many U occupied, even approximately.
• How many storage media in total — the figure that weighs most on method and time.
• Your window: deadlines, permitted hours, and whether the site keeps running in parallel.
• Access: floor, freight elevator, loading dock, building hour restrictions.
• Your data policy: destruction on site, witnessed, or media leaving under custody.
• Which documents you need at the end, and who will receive them (internal audit, head office, finance).

We come back with scope, method, schedule and a quote. Message us on WhatsApp or through the contact page. If your project is broader than the site itself, start with the IT asset disposition guide.