
A QR-coded asset tag becomes the link between a data centre’s physical racks and its digital records
(Image courtesy of Brady Corporation)
A label peels off the back of a server rack. With it goes whatever confidence anyone had in the data centre’s asset records. It isn’t trivial, not for an engineer reconciling a physical audit against a DCIM system. Not for a technician running moves, adds and changes across thousands of ports either. Brady Corporation’s Linda Chochoy links the strain to rising computational demand pushing data centres toward higher asset density. She argues identification has become the thing everything else depends on.
That growth doesn’t happen in isolation. According to the International Energy Agency, data centre electricity demand is on course to roughly double by 2030, and AI is the leading driver behind it. The same expansion is hitting the same limit almost everywhere. Local grids can’t connect new capacity fast enough, and EV charging infrastructure and renewable projects are queuing for connections of their own.
Further down the supply chain, the strain shows up as hardware. Wood Mackenzie’s US analysis puts the shortfall at roughly 30% for power transformers and 10% for distribution transformers. Lead times on distribution padmount transformers have stretched to around 75 weeks, and Wood Mackenzie expects the squeeze to hold for another two to three years, as data centres, manufacturing and EV infrastructure all draw on the same limited manufacturing capacity. For an industry already fighting for grid connections and scarce hardware, knowing exactly what’s installed and where is part of what makes planning any of it possible. When a replacement transformer takes over a year to arrive, an asset record that’s wrong costs a lot more than it used to.
Why data centre asset density keeps climbing
In Chochoy’s account, more connected devices and heavier AI workloads mean more physical kit. That means more cables, more servers, more ports, and racks packed tighter than before. “How do you keep the data centre’s real world, tangible assets aligned with its DCIM or core asset databases?” is the question Chochoy says she hears constantly from data centre managers. The physical estate changes constantly. The digital record only keeps up if someone updates it by hand, or the identification does that work automatically.
From barcodes to QR and datamatrix codes
Barcodes have done this job since the 1960s, and they still work much as they always have. Scan one, and a digital record appears. The trouble, Chochoy says, is scale. Large labels, and the cable flags that keep them legible, don’t fit onto individual server ports. In a high density rack, they pile up.
QR codes and industrial datamatrix codes are Brady’s answer. Both print at millimetre scale while remaining scanner readable. They carry more information than a barcode of the same size. They also fit onto the small cable flags and ports found in dense servers, routers and switches. Reading them, Chochoy says, calls for scanners with high accuracy image capture, combining near-field and far-field technology. Label printers capable of 203 or 600 dots per inch produce the codes.
Why label durability matters for data centre asset identification
None of this matters if the label doesn’t survive the environment it’s stuck in. Server-generated heat degrades lower quality adhesives, Chochoy says. Curved cable surfaces need specialised materials to begin with. Brady’s response is a range of labels in nylon cloth and advanced polymers, tested under FINAT and ASTM methods. The company says these stay attached and legible for up to 10 years in data centre conditions, longer than most of the assets underneath them will last.
Chochoy says identification, done properly, covers every server, router and switch port, as well as the cables between them. That updates a DCIM record with two one-second scans, one for the cable and one for the port. The resulting benefits include faster moves, adds and changes, and quicker troubleshooting. Faster audits, more workable inventory checks, and smoother rack commissioning and decommissioning round out the list. Some higher tier facilities extend the same labelling to transformers and pipes, to speed up a wider range of interventions.
RFID adds speed to data centre asset identification
For data centres wanting to move faster still, Chochoy points to RFID. A reader can pick up multiple tagged assets within range in a single pass. Nobody has to scan a code by hand. Passive UHF RFID labels need no battery, and their thin, integrated antennas reach up to 15 metres in open space. Brady’s on-metal labels resist the signal distortion metal surfaces normally cause. Chochoy says they typically manage 3 to 4 metres inside a data centre. That’s enough to confirm which labelled assets sit in several racks within seconds.

Brady’s on-metal RFID labels are built to resist the signal distortion metal server racks normally cause
Chochoy also describes a portable reader that lets a technician pick a detected asset on screen. From there, they can write to its tag or record. It can home in on a specific item using signal strength changes, with on-screen or audible cues. Fixed readers at server room entrances can detect assets moving in or out and track the direction of travel. That data can update inventory automatically, flagging anything that moves without authorisation.
Picture an engineer standing in a server room with a scanner and a mismatched asset list. None of this is abstract to them, not when the equipment upstream is this scarce and slow to replace. Identification is moving from being a periodic clean-up job to something closer to infrastructure in its own right, one less variable to get wrong in an equipment supply chain that’s already stretched thin.
Brady’s data centre identification and RFID solutions are listed on Electrified Future’s supplier directory. The company’s webinar on data centre lifecycle identification, covering design, commissioning, handover and lifecycle operations, is available to watch on demand.
