Modernizing Aging Data Centers: A Phased Approach to Electrical Infrastructure

──── FEATURED ARTICLES

Modernizing Aging Data Centers: A Phased Approach to Electrical Infrastructure

A meaningful share of the world’s enterprise data center capacity is now fifteen to twenty‑five years old. The buildings still operate, the workloads still run, and the customers behind them still expect five‑nines availability. But the electrical infrastructure that was sized for early‑2000s rack densities, early‑2010s power factors, and a much smaller fault duty profile is reaching the end of its useful life — often quietly, and usually faster than the capital plan anticipated.

Modernizing this infrastructure without taking the facility offline is one of the hardest jobs in mission‑critical work. It is also one of the most consequential. This article looks at how operations leaders should think about phased electrical modernization, why the conversation needs to include both switchgear and switchboards, and how to build a capital plan that protects uptime while keeping pace with new workload demands.

Why modernization is fundamentally different from a greenfield build

A new data center can be designed around a clean specification, energized in a controlled sequence, and commissioned to a documented baseline. A live facility cannot. Every modernization decision must account for tenants whose contracts assume continuous power, redundancy paths that may already be partially compromised, and operational teams whose calendars cannot absorb extended outage windows.

That changes the work in three ways. The schedule must be built around maintenance windows rather than the contractor’s preferred sequence. Every transfer scheme must be tested before it is executed under load. And the equipment selections often must bridge an older system to a newer one, which means understanding the existing protective device coordination, breaker settings, and arc flash labeling well enough to keep the legacy plant safe while the new equipment is being phased in.

This is why successful modernization programs are run as multi‑year campaigns, not single projects. Each phase establishes the conditions for the next phase, and the partner team needs to remain consistent across the program so that institutional knowledge does not have to be rebuilt every fiscal year.

Modernization as a capacity decision, not just a replacement decision

The best modernization programs do not just replace what is failing. They use the program as an opportunity to right‑size the facility for the next ten years of workload demand. That includes assessing the entire MEP System and asking if the overall system capacity, architecture, and location is still pertinent for another 10-20 years.

If modernization means increasing capacity, confirming whether the existing service entrance can support the densities the operator expects to host, whether the distribution architecture can accommodate possible higher‑voltage delivery to the rack, and whether the redundancy strategy still aligns with the value of the workloads on the floor. In many cases, modernization is the only realistic moment to make those changes — once the new equipment is energized, the next opportunity may be a decade away.

Obviously, if major configuration and/or capacity changes are made, it is often necessary to shutdown at least portions of the facility for longer periods of time than a more straight forward moderation plan as described below.

Switchgear and switchboards — why both belong in the plan

Most modernization conversations start with switchgear, but the planning problem includes both medium voltage switchgear and the low voltage switchboards that distribute power inside the building. The two have different lifecycles, different failure modes, and different operational consequences.

Medium voltage switchgear typically lives outside the building or in a dedicated electrical yard. It handles utility‑side service, primary protection, and transfer between sources. Failures here are rare but high‑impact: an internal arc fault inside a switchgear cubicle can cascade through coordination zones and take entire portions of the facility offline.

Switchboards live inside the building, typically in the electrical rooms feeding the IT halls. They handle distribution to PDUs, mechanical loads, and ancillary systems. Their failure profile is different — more frequent component‑level issues, more thermal signatures, more wear from repeated switching — and their replacement cycle often runs slightly faster than the upstream switchgear.

A modernization plan that addresses only the medium voltage equipment and assumes the switchboards “still have life left in them” is an incomplete plan. So is the inverse. Both need to be assessed, both need a documented condition baseline, and both need to be part of the multi‑year capital sequence.

The condition assessment that anchors the plan

Before any equipment is specified or ordered, a serious modernization program starts with a condition assessment of the existing electrical plant. This is the work that separates a real capital plan from a wish list.

A complete assessment typically includes infrared thermography on energized equipment, insulation resistance testing on cables and bus, contact resistance measurements at terminations and breakers, mechanical operation testing on breakers and switches, and a refreshed arc flash hazard analysis tied to the current protective device settings and utility infeed. Where appropriate, partial discharge testing on medium voltage equipment often reveals issues that infrared scanning alone can miss.

The assessment is also where the operator should confirm parts availability for the existing equipment. Many switchgear and switchboard lines that were standard fifteen years ago are now obsolete or available only through limited service channels. Confirming that certified replacement parts and trained service technicians are still available — for both the primary equipment and the embedded controls, relays, and trip units — is often the single most important input to the replacement timeline.

NETA Maintenance Testing Specifications (MTS) provide a useful baseline for what the assessment should cover. So does coordination with the OEMs whose equipment is in place. The deliverable is a condition‑rated inventory, prioritized by risk, that the capital plan can sequence against.

Building a phased capital plan

Once the assessment is in hand, the program becomes a sequencing problem. A typical phased approach over four to five years looks something like this:

Year 1 focuses on the highest‑risk findings: thermal anomalies on critical paths, breakers showing mechanical wear, arc flash labels that no longer match the system as installed, and any equipment where parts availability is already compromised. Year 1 is also when arc flash analysis should be refreshed, LOTO procedures should be revalidated, and any urgent capacity bottlenecks should be addressed.

Years 2–3 typically address medium voltage modernization. Long lead times — often sixteen to twenty‑four weeks for custom switchgear — make this work hard to accelerate, and the planning effort to coordinate utility outages, transfer schemes, and tenant communications usually starts six to twelve months before any equipment lands on site.

Years 4–5 address the inside‑the‑building distribution: switchboard replacements, busway upgrades, PDU and tap‑off refreshes, and any panelboard work that has been deferred. This phase often coincides with monitoring and EPMS upgrades that take advantage of the new equipment’s instrumentation.

Throughout the program, the capital plan needs a contingency reserve. Modernization work uncovers issues that were not visible in the original assessment, particularly when equipment is opened up for the first time in years. A plan with no reserve is a plan that will be revised under pressure.

Coordinating with live load

The single biggest constraint in a modernization program is the live load the facility is already serving. Every transfer, energization, and de‑energization must be planned against the existing redundancy paths, the current protective device coordination, and the operational calendar of the customers depending on the facility.

This is where start‑up and commissioning support takes on outsized importance. The team that supports the energization of new equipment needs to understand both the new equipment specifications and the existing system it is being integrated into. Factory acceptance testing, site acceptance testing, and integrated systems testing all have to be sequenced so that no transfer is executed for the first time on a fully loaded system.

OEM relationships matter here too. The OEMs whose equipment is being installed should be engaged in the commissioning sequence, particularly for primary medium voltage gear and any equipment with embedded protective relays or digital controls. So should the partners providing component, section, and system‑level coordination support across the broader project — which typically spans multiple trades, multiple OEMs, and multiple subcontractor scopes.

What this means for operations leaders

Aging electrical infrastructure rarely sends a clear warning before it fails. By the time it does, the choice between proactive modernization and emergency replacement has already been made — and not in the operator’s favor. A disciplined, phased, multi‑year program, anchored in a real condition assessment and sequenced around live load, is the most reliable way to protect uptime and to position the facility for what comes next.

The work is harder than a greenfield build. It is also where the most experienced critical power partners earn their keep.