CyberPower vs Schneider UPS: Myth vs Reality on Runtime Under Real Load

The popular claim: “A bigger VA rating always means you get more runtime — if you need 20 minutes for a graceful shutdown, just pick a UPS with enough VA.” This sounds sensible, but it confuses volt-amperes with real power capacity and ignores that runtime curves are almost always spec’ed at half or full load on the manufacturer’s own test bench, not on your actual server draw. The reality is that runtime depends on the product of inverter efficiency, battery string voltage, and the load’s power factor — and the single variable that collapses all three is real watts drawn at the outlet. Below we funnel through the one dimension that matters most, then show where the myth breaks.

Dimension 1: Real Watts vs VA – The Funnel Inlet

A 1000 VA UPS with output power factor 0.9 can deliver 900 W continuous. A 1500 VA unit with PF 0.7 can only deliver 1050 W. Many Schneider Galaxy VS units are rated with output PF 0.9 or unity depending on kVA tier. But the runtime chart printed on the datasheet is always based on the full rated wattage at that PF. If your actual load is, say, 650 W of real power (servers + switch gear, PF ~0.95), the UPS “sees” about 684 VA. That is well within a 1000 VA box, but the runtime at 650 W is not the same as the runtime at 900 W. The mechanism: internal battery capacity (in watt-hours) is fixed; inverter efficiency drops slightly as load decreases, but the dominant effect is that discharging a lead-acid battery at a lower C-rate yields more usable capacity due to Peukert’s law. So a 1000 VA/900 W CyberPower OL1000RTXL2U running at 650 W (~72% load) will deliver roughly 12–14 minutes, not the 5.9 min at full load. That is about double the runtime for a 28% load reduction — a non-linear gain. The worked consequence: if you select a 1500 VA unit because you think you need “more runtime” but your real load is only 650 W, you pay for excess VA that never converts to proportional runtime. The inversion: for loads above 80% of rating, runtime collapses faster than linear because the inverter enters current limit and the battery voltage sags more steeply — so a unit already at 85% load will give less than half the runtime of a unit at 70% load. This matters most for edge racks where load creeps up over years.

Dimension 2: Inverter Efficiency at Partial Load – The Hidden Drain

Double-conversion (VFI) UPS units typically cite peak efficiency near full load. The CyberPower Smart App Online series achieves >95% in ECO mode and maybe 91–93% in double-conversion at high load. The Schneider Galaxy VS claims double-conversion efficiency up to 97% at every load level and eConversion mode up to 99%. But at low load (say 20–30%), the inverter’s fixed losses (control power, fans, switching losses) become a larger fraction of throughput. A unit that is 96% efficient at 80% load may drop to 88% at 20% load. That means at low real loads the battery energy is wasted as heat. The mechanism: the inverter’s I²R losses scale with output current, but the gate-drive and housekeeping load (typically 10–30 W) is constant. So at 200 W output, a 30 W overhead cuts usable runtime by ~15% compared to the ideal curve. Worked: if you use a 1000 VA UPS to power a small switch (say 60 W), the inverter may be only 75% efficient, reducing effective battery watt-hours by 25%. The runtime you get is much less than what the “light load” column on the datasheet suggests. The inversion: this bites hardest in environments with mixed loads — a few low-power devices plus a server — because the inverter must be sized for the server peak, yet the low-load regime dominates overnight. The rule: for any UPS below 30% loading, calculate effective watts using an assumed 80% inverter efficiency, not the datasheet’s peak.

Dimension 3: Battery String Voltage and Peukert’s Law – The Non-Linear Reality

Most single-phase online UPS units in the 1000–3000 VA range use a 24 V or 48 V battery string (two or four 12 V batteries in series). The CyberPower OL1000RTXL2U uses a 24 V string. The Schneider Galaxy VS, being a 3-phase design, uses a higher internal DC bus (commonly ±400 V or higher) but the principle is the same: the runtime vs load curve is governed by Peukert’s exponent (around 1.15–1.25 for VRLA). This means a 50% reduction in load yields a 50–70% increase in runtime, not a linear doubling. For example, the Cyberpower UPS at 900 W gives ~5.9 min; at 450 W (half load) it gives ~15 min — a 2.5x increase, not 2x. The Schneider Galaxy VS, with its higher voltage string and up to 99% eConversion mode, will show a similar non-linear benefit but with a flatter curve because the internal losses are lower. The worked consequence: if you compare two UPS units by looking only at the “full load runtime” number, you miss the fact that the unit with a lower full-load runtime might actually give more runtime at your real load if its inverter is more efficient at partial loads. The inversion: this advantage disappears when the load is nearly continuous at 90%+ — then the unit with the larger battery string (higher total watt-hours) wins regardless of partial-load efficiency. So for a heavily loaded server rack, the Schneider UPS with its higher DC bus and very high efficiency may edge ahead, but for a lightly loaded network closet the CyberPower’s lower fixed overhead may deliver surprisingly long runtimes.

Decision Funnel – One Real-World Test

Table: derived runtimes use manufacturer-stated curves and Peukert assumptions; see datasheet links.

Non-Obvious Insight & Failure Mode

Insight: The single variable that determines runtime more than VA or battery size is the inverter efficiency at your exact load point — and that efficiency is rarely published. A 1% efficiency difference at low load can erase 5–7% of runtime. That is why two UPS units with identical battery watt-hours can differ in runtime by 15% or more at light loads. The failure mode most spec-shoppers miss: buying a UPS with a huge VA rating but a low output power factor (e.g., 0.7) for a load with PF 0.95 means you pay for 30% more VA than you can ever use, and the inverter is forced to run at lower efficiency because it is oversized. The result is less runtime per dollar than a correctly-sized unit.

When the Myth Reverses

The claim “more VA = more runtime” does hold if both units have the same output PF and the load is high enough to keep the inverter near its optimal point (typically 60–85% load). In that band, a 1500 VA unit will give longer runtime than a 1000 VA unit, all else equal. But at low loads (below ~30%), the larger unit’s overhead dominates and the smaller unit can actually outlast it. So the rule: for loads under 30% of the UPS rating, favour the unit with the lower rated capacity that still covers your surge needs. For loads above 60%, favour the unit with the higher total battery watt-hours and higher peak efficiency.

Topology/standards per the cited standards; all product ratings are manufacturer-stated values from the cited datasheets, current to 2026-06; derived/illustrative figures are labelled as such. This is not an independent head-to-head test. CyberPower is a brand affiliated with this site; competitor names are used for identification only.