How to Size a UPS for Your Home Lab

Buying a UPS without sizing it first is like buying a generator without knowing what your house draws. You will either pay for capacity you do not need or end up with a unit that overloads, or that gives you a couple of minutes of runtime instead of enough for a clean shutdown.

This guide walks through the math: how to read VA and watt ratings, how to measure your lab’s load, how to read manufacturer runtime data, and which capacity tier fits.

If you just want product recommendations, head to best UPS for home lab. If you are comparing the two most common brands, see CyberPower vs APC.

Understanding VA vs Watts

Every UPS has two capacity ratings: VA (volt-amps) and watts. They are related but not interchangeable, and confusing them is the most common sizing mistake.

VA = volts × amps. This is apparent power: everything the UPS must supply, including reactive power that does no useful work.

Watts = real power, the portion that actually runs your devices.

The ratio between watts and VA is the power factor. A 1500VA UPS with a 0.6 power factor delivers 1500 × 0.6 = 900W, which is why 1500VA/900W is a common pairing. CyberPower’s CP1500PFCLCD is rated 1500VA/1000W, and some rack units, such as CyberPower’s PR1500RTXL2UC, are rated 1500VA/1500W.

Modern computer power supplies with active PFC have a power factor close to 1, so for them the watt rating is the limit you hit first. The practical rules:

• A “1500VA” UPS does not give you 1500W. It gives you whatever the watt rating says, usually 900W to 1000W for a tower.
• Size against the watt rating. Check the VA rating too if you protect older or unusual loads with a low power factor.

If a UPS lists only VA without a watt rating, treat that as a red flag.

Measuring Your Home Lab’s Power Draw

Spec sheets describe maximums and typical cases. Your NAS might draw a fraction of its listed consumption with its disks spun down at 3 AM and far more during a parity rebuild.

The only way to size a UPS accurately is to measure the load yourself.

Use a Plug-In Power Meter

A plug-in power meter (a Kill-A-Watt or similar) goes between your gear and the wall. Run the whole lab through a power strip into the meter and note two numbers:

1. Idle draw: everything running but not busy. This is your baseline, and most home labs spend most of their time near it.
2. Peak draw: the highest number you see during heavy work, such as a NAS rebuild, a compile job, or all drives spinning up at once. The UPS must handle this without overloading.

Watch it for at least a day if you can. A 3.5-inch hard drive can draw roughly 20 to 30W for a few seconds while it spins up, so several drives starting together create a noticeable spike.

Spec Sheet vs Reality

Rough planning ranges for common home-lab hardware (measure your own, since these vary by model and workload):

• Mini PC (Intel N100/N150 class): roughly 8 to 15W idle and 20 to 30W under load. The N100 itself is a 6W chip, but the whole box includes memory, storage, and the power adapter.
• Mini PC (mobile Core i5/i7 or Ryzen 7 class): roughly 15 to 25W idle and 45 to 80W under load.
• 2-bay NAS: use the manufacturer’s power specification as a planning boundary, then measure with your drives installed.
• 4-bay NAS with spinning drives: often 30 to 60W, and higher during rebuilds; drive count and drive model matter most.
• 8-port managed switch: roughly 10 to 20W, plus whatever PoE devices draw.

Idle power is usually well below the manufacturer’s maximum, but peaks can reach or exceed it. Size for your peak, not the idle number.

Common Home Lab Power Budgets

These are rough planning ranges for common setups. Use them as a starting point, then check them with your own meter.

Most home labs fall in the light-to-medium range, 50 to 130W at idle. A single mini PC home server with a NAS typically lands around 60 to 90W.

Runtime Data: How to Read It

Manufacturers publish runtime at specific loads, as charts, graphs, or at minimum half-load and full-load figures. Here is how to read them.

Load Percentage Is Everything

Runtime is not linear. Halving the load more than doubles the runtime, because lead-acid batteries deliver more of their stored energy when discharged slowly. The published figures show the pattern:

For the CP1500PFCLCD, half load gives four times the full-load runtime, and a typical 80 to 150W home-lab load runs much longer again. Read the manufacturer’s chart at your load rather than extrapolating; CyberPower and APC both publish runtime charts or graphs on their product pages.

Check Runtime Against Battery Capacity

If you cannot find a chart for your load, the battery gives you an upper bound. The CP1500PFCLCD has two 12V/9Ah batteries, about 216Wh nominal. A lead-acid battery delivers only part of that, and less at high loads, and the inverter loses more. Treat any estimate from battery watt-hours as a rough ceiling and confirm it with the maker’s chart or a real test.

Important Caveats

Battery age matters. Published runtimes assume new, fully charged batteries. Capacity falls as batteries age, so a three-year-old battery gives you less than the chart says.

Temperature matters. Lead-acid batteries are rated at 25°C (77°F), and their service life shortens sharply in warm rooms. The UPS models in our picks, such as the CP1500PFCLCD and BR1500MS2, are rated to operate between 0 and 40°C (32 and 104°F); do not install one anywhere hotter.

Leave margin. Treat published runtimes as best-case figures, and verify with a real power-loss test once the lab is set up.

The 1.5x Rule and Headroom

The single most important sizing principle: keep your peak load at or below about two-thirds of the UPS watt rating.

This is not arbitrary:

1. Runtime collapses at high loads. As the table above shows, full load gives a fraction of the half-load runtime. A UPS near its limit leaves barely enough time for a graceful shutdown.

2. Startup surges can trip an overload. When power returns and everything restarts at once, inrush current can briefly exceed the steady draw by a wide margin. A UPS already near its limit may trip.

3. You will add gear. Another NAS, a Raspberry Pi, or a PoE camera eats into the margin.

The practical rule: multiply your measured peak by 1.5. That is your minimum UPS watt rating.

• Peak draw of 100W → 150W minimum → an 850VA/510W unit works
• Peak draw of 200W → 300W minimum → a 1000VA/600W unit works; 1500VA adds runtime and room to grow
• Peak draw of 400W → 600W minimum → a 1000VA/600W unit is exactly at the limit, so choose a 1500VA unit rated 900W or more

Choosing Between 850VA, 1000VA, and 1500VA

These are the three most common tower UPS tiers for home labs.

850VA (about 500W): The Minimal Setup

A unit like the CyberPower CP850PFCLCD (850VA/510W) handles a single mini PC plus a switch, or a small NAS on its own. It suits one or two low-power devices when you want basic outage protection without overspending.

Best for: single mini PC setups, network-only protection (router, switch, access point), and builds with a peak under about 340W.

Runtime: CyberPower lists about 9 to 11 minutes at half load (255W), and a 60W load runs considerably longer.

1000VA (about 600W): The Tweener

There are fewer options here, and prices are often close to 1500VA units. Unless the budget is tight, stepping up to 1500VA usually buys more headroom and runtime for a modest premium.

Best for: a good deal on a unit like the CyberPower CP1000PFCLCD when your peak stays under about 400W. Check 1500VA pricing first.

1500VA (about 900 to 1000W): The Home Lab Standard

The CyberPower CP1500PFCLCD and APC BR1500MS2 are our default tower picks. They handle everything from a single NAS to a multi-node mini PC cluster with room to spare, and both support USB shutdown signaling.

Best for: most home labs with a NAS, mini PC, and networking gear, with room to grow. See our best UPS for home lab roundup and CyberPower vs APC for the trade-offs.

Runtime: CyberPower lists 10 minutes at half load (500W) for the CP1500PFCLCD; a typical 100 to 150W lab runs much longer.

If your peak draw exceeds about 650W, look at 1500W-class rack units or split the lab across two UPS units. For most home labs, 1500VA is the ceiling you will need.

Common Mistakes to Avoid

Sizing by VA instead of watts. A 1500VA tower is usually a 900W to 1000W UPS. Always use the watt number.

Trusting spec sheets for power draw. Idle draw is usually far below the listed maximum. Measure first and buy second; an inexpensive power meter can prevent a bad UPS purchase.

Buying simulated sine wave for active-PFC gear. Some active-PFC power supplies shut down or refuse to transfer on stepped output. For desktop PCs, servers, and NAS units with internal power supplies, buy sine wave. A stepped unit can be fine for a modem, router, and brick-powered devices (see our under-$150 guide).

Forgetting about battery replacement. Every UPS battery wears out. Plan for a replacement every few years and factor it into the cost of ownership. The CyberPower and APC towers in our guides use user-replaceable batteries.

Putting everything on one overloaded UPS. If one 1500VA/1000W unit would sit at 80% (800W), splitting that load across two such units puts each near 40%, which gives more runtime and some redundancy.

Skipping the USB monitoring cable. A UPS without monitoring only buys time. Connect it to your NAS or hypervisor by USB, set up NUT or the manufacturer’s software, and test automatic shutdowns.

Installing it somewhere hot. Keep the UPS in a conditioned space within its rated 0 to 40°C (32 to 104°F) range. Attics and hot garages shorten battery life and can exceed that limit.

Wrap-Up

The sizing formula is simple: measure your peak draw, multiply by 1.5, and buy the tier whose watt rating covers it. For most home labs, that means a 1500VA sine-wave tower such as the CyberPower CP1500PFCLCD or APC BR1500MS2. If the UPS goes in a rack, budget for a properly rated shelf or a rack-mount UPS.

If your lab is a single mini PC and a switch, an 850VA unit saves money without giving up meaningful runtime. If you run a full rack, you may need a 1500W-class unit or two separate UPS units.

For product picks, read our best UPS for home lab roundup. Protecting only network gear on a budget? See best UPS under $150. And if you are torn between the two big brands, CyberPower vs APC breaks it down.