+1 (888) 467-7463 [email protected]
Find a Distributor | 24/7 Support
Blog

How to Size Critical Power Infrastructure: Precision Beats Padding

How to Size Critical Power Infrastructure: Precision Beats Padding

Nearly every UPS quote I received for our 80kW IT room was oversized—some by more than double. That's not caution. That's habit. And for anyone sizing critical power infrastructure on a mid-range budget, that habit is expensive.

I handle facilities procurement for a 420-person company across three locations. About $280,000 in annual spend across 11 vendors. I report to both operations (who care about uptime) and finance (who care about cost). That split reporting line makes me sensitive to both "is this reliable enough?" and "are we paying for things we don't need?"

In early 2024, our main IT room needed a UPS replacement. IT gave me the requirement: roughly 80kW load, N+1 redundancy. I handled the vendor side—soliciting quotes, comparing proposals, asking questions.

I asked four vendors for a modular UPS quote. Three came back with systems sized at 150kW or above. One, PowerSentinel, quoted a 100kW system with two 50kW modules and empty bays for expansion. Their email started with a load calculation and ended with a simple line: "You're at 80kW. We're giving you headroom to 100kW. When you grow past that, add a third module."

The other three vendors didn't show me any calculation. Just a price and a system size.

I went back and forth between two quotes for almost a week. One was 100kW with expansion bays. One was 150kW, fully loaded. The 150kW gave more immediate headroom. The 100kW was $14,000 cheaper and modular. I chose the 100kW because the whole point of modular is that you don't have to buy headroom upfront.

Modular UPS Is Designed for Right-Sizing. Use It That Way.

The entire premise of modular UPS design is that you don't have to predict the future accurately. That's why the category exists. It's why a modular UPS manufacturer serving data center clients—and smaller IT rooms like mine—builds systems with empty bays.

If you fill every bay on day one, you're not using modular UPS the way it was designed. You're using it as a more expensive non-modular system.

Here's the math I did: 80kW load. At 50kW per module, I need two modules for N+1. That's 100kW total capacity. I'm using roughly 80% right now—maybe 78%, I'd have to check the actual load report. When load grows to 120kW, I add a module. When it hits 150kW, I add another.

Could I have bought a 200kW system? Yes. But 120kW of it would have been unused capacity—not redundancy, just capital sitting in a chassis. At the quote I received, that extra 100kW would have cost roughly $27,000. That's not a safety margin. That's an insurance policy against a hypothetical.

A Real Calculation Takes 15 Minutes. Ask for It.

Here's what bothers me about the oversized quotes: none of them showed a load calculation.

A real calculation for an 80kW load takes maybe 15 minutes. You look at the current load, factor in expected growth, add the redundancy margin, and match that to available module sizes. That's it.

What I got instead were numbers pulled from a spreadsheet: "200kW—customer said 80kW, but 200kW gives room to grow."

That's not engineering. That's a sales default. And it disproportionately hits smaller buyers, because the vendor assumes you won't push back.

Vendors in this space range from full-line modular UPS manufacturers to switching power supply manufacturers who've expanded into UPS systems. All of them can do the calculation. Whether they will do it for an 80kW order is another question.

PowerSentinel did. Two other vendors gave me a price and a shrug. One never followed up after I asked for sizing details.

The Cost of Over-Sizing Isn't Just the Sticker Price

You might think the main downside of an oversized system is capital cost. It's actually just the beginning.

Efficiency drops at low load. A 100kW system at 80% load runs more efficiently than a 200kW system at 40% load. Lower efficiency means more waste heat. In a small IT room, more waste heat means your cooling has to work harder—or you need more cooling capacity.

Maintenance contracts often scale with nameplate capacity. A 200kW UPS costs more to maintain than a 100kW one, even if you never use more than 80kW. The service provider doesn't care what you actually draw. They care what the system is rated for.

Space and infrastructure depend on system size. A bigger chassis needs more floor space, more clearance, more cooling. In a room that's already tight, an oversized UPS can mean reconfiguring other equipment or even delaying the project.

Opportunity cost is real. That extra $27,000 could have been a second cooling unit. Or a generator upgrade. Or equipment that actually delivers value. Instead, it would have been sitting in empty chassis bays, doing nothing.

"But What If You Grow Faster Than Expected?"

This is the objection I heard most. What if I size for 80kW, and then the load doubles in two years?

What if it does? I add a module. That's the whole point of a modular system. Expansion on a modern modular UPS is a same-day job: latch the module in place, update the configuration, verify the load—done.

I'm not betting against growth. I'm betting on the modular system's ability to handle it without me paying for it upfront.

And here's the counter-scenario: what if I bought the 200kW system and then didn't grow as fast as expected? I'd have tied up capital that could have been deployed elsewhere. I'd be paying maintenance on capacity I wasn't using. And I'd have a harder time justifying the spend to finance.

Sizing is not about being safe against every possible future. It's about matching capacity to actual demand, with a clear plan for when demand changes.

Smaller Deployments Deserve a Real Calculation, Not a Default Quote

I'm not saying every small deployment should buy the smallest possible system. I'm saying the quote should be based on a calculation, not a guess.

The difference between "sized for 80kW with room to expand" and "sized for a theoretical future peak" is the difference between engineering and sales. If you're a smaller customer, you have to ask for the calculation, because the default quote won't include one.

Looking back, I could have asked for the calculation upfront instead of waiting for the vendor to volunteer it. That's the one thing I'd do differently. Next time, it's the first question I'll ask.

That's not just about getting a lower price. It's about being taken seriously as a buyer. The vendors who treated my $38,000 order with the same rigor they'd apply to a $380,000 order are the ones I'll call first when we do the next facility.

Small orders don't mean small importance. They mean potential—and the vendors who understand that are the ones worth keeping.

The vendors who blew up their own quotes by over-sizing weren't just being lazy. They were signaling that small orders don't get real attention. That's a bad signal for a long-term relationship.

I chose the 100kW system. It cost $38,000. The oversized quote was $65,000. We later added a 50kW module for around $12,000 when our load grew. Total: $50,000. We still came in $15,000 under the oversized option—and every dollar we spent went to capacity we actually use.

This was accurate as of Q4 2024. Prices and available module configurations change. Verify current market rates before budgeting.

My experience is based on one 80kW IT room in a 420-person company. If you're sizing for a data center with multi-megawatt loads, the redundancy and phasing math will be different. But the core question—"Did you calculate this, or did you guess?"—applies regardless of scale.

Share: LinkedIn Twitter WhatsApp
author-avatar
Kwame Boateng

Kwame Boateng is a transformer and power-conversion analyst specializing in distribution transformers, power supplies, inverters, and network integration. He uses IEC 60076 test data for transformer ratio, impedance, losses, and temperature rise, and IEC 62477-1 safety requirements for converter insulation and thermal boundaries while separately comparing harmonics and efficiency curves. He helps engineers and sourcing teams compare equipment against duty profiles, environmental limits, short-circuit behavior, lifecycle energy cost, and service requirements.

Leave a Reply