Data Center Switchgear Planning: Busbar Design, Capacity, and Procurement

Data center switchgear should be procured from an electrical architecture, not from a cabinet count. AI load growth, phased expansion, maintenance states, utility constraints, and alternate-source arrangements can change the current path long after a preliminary one-line is drawn. Those changes eventually reach the busbar package as new section ratings, tie configurations, hole patterns, joints, and delivery priorities.

The purpose of the procurement plan is to stabilize those inputs early enough for design verification and manufacturing, while preserving controlled options for expansion. A supplier’s nameplate catalogue and a promised ship date are not enough.

Freeze the Architecture Before Requesting Comparable Quotes

At RFQ release, the buyer should define the facility boundary and the operating states the equipment must support. Include utility and on-site sources, transformers, generators or other backup, UPS topology, maintenance bypasses, tie breakers, load blocks, and future modules. Mark which portions are fixed and which are alternatives.

The Congressional Research Service report on electricity demand from data centers, updated in May 2026, describes the policy and grid-planning context behind rapidly changing US data center loads. It does not specify a switchgear design. Project teams still need site-specific load-flow, short-circuit, coordination, grounding, and reliability studies.

A useful basis-of-design package includes:

• normal, emergency, maintenance, and expansion one-lines;
• voltage, frequency, phases, wires, and grounding;
• present load, committed expansion, and design reserve;
• available fault current and protective-device strategy;
• redundancy objective and allowed maintenance outages;
• room conditions, altitude, ventilation, access, and seismic requirements;
• equipment classification and destination-market standards;
• interfaces to transformers, UPS, busway, cables, and controls.

Without that package, vendors price different systems under the same label.

Translate Facility Priorities Into Busbar Inputs

Facility requirement	Busbar implication	Evidence to request
Continuous load and growth	conductor arrangement and thermal design	calculations or verification allocated by the assembly design
Available fault current	support spacing, joint and conductor-force design	short-circuit verification basis
Redundant sources and ties	multiple credible current paths	operating-state study and section ratings
Maintainability	accessible joints, isolation boundaries, replaceable sections	layout and maintenance procedure
Monitoring	sensor locations and safe access	interface drawing and calibration plan
Phased deployment	repeatable modules and controlled revisions	configuration matrix and change log

The conductor cross-section cannot be selected independently from enclosure ventilation, supports, adjacent heat sources, joints, and permitted temperature rise. Likewise, short-circuit withstand is an assembly property; using thick copper does not prove that supports and joints can withstand electromagnetic forces.

Separate Switchgear, Switchboard, and Busway Work Packages

These terms describe different equipment and responsibilities. The RFQ should show where switchgear or switchboard terminals end, where busway begins, who supplies transition pieces, and who owns interface tolerances and site measurements.

For a North American dead-front switchboard, the UL 891 fabrication guide explains why equipment classification precedes copper release. For IEC assemblies, IEC 61439-1:2020 provides the general-rules context. Neither route should be selected by casually changing the terminology in a purchase order.

The existing 1500 VDC data center busway guide addresses the facility-side conductor system. This article keeps the switchgear work package separate so both pages answer a distinct procurement task.

Evaluate Fabrication Capacity, Not Only Nameplate Ratings

A switchgear design can be valid while its production plan is fragile. Buyers should map the planned copper parts into families: straight punched bars, complex bends, joint plates, neutral and ground conductors, transition pieces, and repeated module sets. Then ask how the supplier will schedule the bottleneck operations.

Capacity evidence should address material size range, punching and shearing, bend types, deburring, finish, first-piece inspection, program control, and handling between operations. It should also show what happens when a drawing changes after the first batch.

For repeat parts, a controlled CNC busbar punching and shearing workstation can reduce manual layout variation. Buyers still need a trial using their material and drawing; catalogue speed is not the same as finished-part throughput.

Make Verification and Traceability Deliverables Part of the RFQ

Do not wait until shipment to ask for quality records. Define the deliverable matrix with the bid:

• approved one-line, assembly drawings, busbar drawings, and revision list;
• material specifications and required certificates;
• first-piece and dimensional reports for critical parts;
• joint assembly, finish, insulation, and torque procedures as allocated;
• design-verification and routine-verification documents required by the project;
• measuring-equipment and test-report identification;
• nonconformance, concession, and engineering-change records;
• packing, preservation, installation, and commissioning documents.

The matrix should state who creates, reviews, and approves each record. A generic certificate bundle does not replace project traceability.

Plan Lead-Time and Design-Change Risk by Dependency

Break the schedule into approved drawings, long-lead protective devices, transformer interfaces, copper and steel, busbar fabrication, assembly, verification, shipment, room readiness, and commissioning. Each dependency needs a required-by date and an owner.

Modularization can help when it reduces unique parts and stabilizes interfaces. It hurts when a “standard” section is repeatedly modified after validation. Establish a design-freeze process with defined classes of change: documentation-only, manufacturing change, interface change, and change requiring engineering or certification review.

The transformer-shortage capacity guide addresses one upstream risk. The broader hyperscale fabrication article covers capacity strategy beyond the switchgear package.

Prepare Interfaces for Emerging 800 VDC Architectures

NVIDIA’s published 800 VDC architecture and the Open Compute Project’s LVDC power-distribution white paper have increased attention on higher-voltage DC distribution closer to high-density AI racks. This does not mean every current data center switchgear RFQ should be converted to 800 VDC. It means the team should define the AC/DC conversion boundary, row-power equipment, protection ownership, busway interfaces, isolation, monitoring, and maintenance model before reserving space or terminals.

The dedicated 800 VDC data center architecture guide separates published roadmap statements from current project requirements.

Use a Procurement Scorecard That Exposes Risk

Scorecard area	What a credible response contains
Technical fit	deviations against a frozen basis of design
Verification	named standard route and evidence matrix
Fabrication	part-family capacity, trial plan, inspection method
Schedule	dependency-level dates and buyer inputs
Change control	authority, notification, re-verification triggers
Installation	interface checks, site utilities, commissioning plan
Service	training, spares, backups, response path

Weight the categories for the project rather than awarding the order to the lowest cabinet price. For switchgear panel busbar production, the strongest proposal is the one that makes design boundaries, capacity evidence, and change responsibility visible before manufacturing starts.