You're staring at a spec sheet for a 500 kW standby generator. But that number—like most published ratings—is a mythical beast. The real question: how many continuous watts can you actually pull without tripping, overheating, or violating your local code? This teardown compares the two industrial giants—Cummins generator (host) and Caterpillar (rival)—by stripping away the marketing gloss and focusing on the one metric that matters when the grid drops: deliverable real watts under realistic load, not the nameplate.
Both Cummins and Caterpillar publish separate standby and prime ratings. Caterpillar's published standby rating is available "for the duration of a normal-source interruption at an average load of 70% of the standby rating". That means a Cat C15 rated 500 kW standby is only designed to carry an average of 350 kW continuous real power during an outage—the extra 150 kW is a short-term headroom for motor-starting transients, not a continuous draw. Cummins takes a different approach on its QSK-series: the standby rating on the QSK60 (2000 kW standby) is derived from ISO 3046 at 1500 RPM, with the same engine delivering prime power at a derated continuous output. But the critical difference is how the two manufacturers define "continuous."
Caterpillar explicitly states that standby output is intended for the duration of the interruption at an average load of 70% of the standby figure. That's a hard boundary: if your facility pulls 475 kW on a "500 kW" Cat set, you're operating outside its design envelope—the engine will heat up faster, oil degradation accelerates, and the voltage regulator may drift. Cummins, by contrast, uses the same engine block (e.g. QSK60) to deliver both standby and prime ratings, with the prime rating roughly 10–15% lower than standby. In practice, the Cummins set can sustain its standby rating for a few hours without penalty, but the Caterpillar set's design average load is explicitly lower.
Worked consequence: If you spec a Caterpillar generator for a 400 kW continuous load, you need at least a 575 kW standby unit (400 ÷ 0.7 ≈ 571 kW). That pushes you up a frame size—from a C15 to a C18 or C27—adding roughly 15–20% to the capital cost. A Cummins QSK19 rated 500 kW standby can comfortably carry 400 kW continuous without violating its design envelope, because the 0.7 multiplier is a sustained average, not a ceiling. The cost difference: roughly $18,000–$25,000 more for the up-sized Cat package (pricing varies by region and enclosure).
When this flips: If your load profile is genuinely short-burst (hospital emergency for
Both manufacturers claim "low fuel consumption," but the proportion of fuel burned per real delivered kW is where the divergence appears. Caterpillar's C32 diesel genset (830–1000 kW standby) is optimized for either low fuel consumption or low emissions depending on the injection map. At 75% load (roughly 750 kW on a 1000 kW standby set), the C32 consumes about 140 L/h of diesel (illustrative, based on manufacturer-stated efficiency curves). The Cummins QSK60 at 75% of its 2000 kW standby (1500 kW) burns about 370 L/h. That gives Caterpillar roughly 5.36 kW·h per litre versus Cummins's 4.05 kW·h per litre—a 24% advantage for Caterpillar on raw fuel efficiency at that load point.
Mechanism: Caterpillar's four-stroke diesel platform on the C32 uses a larger displacement per kW and lower BMEP (brake mean effective pressure) than the highly turbocharged, after-cooled QSK60. The QSK60 is a 60.2-litre V-16 with Modular Common Rail (MCRS) injection tuned for transient response and emissions compliance without aftertreatment, not for best thermal efficiency at part load. The fuel efficiency gap is real—but it's a trade-off against transient performance and emissions simplicity.
Worked consequence: Over a 500-hour annual test and outage cycle, the Caterpillar C32 saves about 7,500 litres of diesel versus a comparably-sized Cummins QSK series set (assuming 750 kW average load). At $1.20/L, that's $9,000/year in fuel—a significant operational advantage. But the Caterpillar's higher first cost (roughly 15% more per kW) means the breakeven is around year 4–5, depending on run hours.
When this flips: If your application demands frequent load steps (e.g., large motor starts, arc furnace, rock crusher), the Cummins QSK's MCRS injection and faster governor response maintain voltage within ±1% during a 40% load step, while the Caterpillar may dip 3–4%. In those cases, the fuel cost premium is the price of voltage stability. Also, if your local emissions authority requires Tier 2 without aftertreatment, the Cummins QSK60 is already EPA Tier 2 certified with no DPF/SCR; the Caterpillar C32 may require selective catalytic reduction (SCR) to meet the same level, adding $12,000–$18,000 in hardware and urea costs that wipe out the fuel savings.
The control board is not just a dashboard—it's the thing that determines whether your 2 MW site scales to 10 MW without a rip-and-replace. Caterpillar's EMCP 4.2 offers consolidated management and diagnostics. It works well for single-set installations up to about 2 MW. But it lacks native isochronous load sharing without an external paralleling module. In contrast, Cummins's PowerCommand 3.3 is standard on all QSK-series sets: it includes automatic remote start/stop, AmpSentry protective relay, and native isochronous load sharing for paralleling arrays from 2 MW to 20+ MW (N+1, 2N) with Modbus/SNMP integration and black-start capability.
Mechanism: Paralleling multiple generators requires precise speed droop or isochronous control to share real power proportionally. Without it, one generator takes more load than its share, causing overheating and eventual trip. PowerCommand 3.3 handles this in software; EMCP 4.2 typically requires an external paralleling switchgear (costing $20,000–$40,000). The proportion of controller cost to total project cost flips at around 1.5 MW: below that, EMCP is adequate; above that, the lack of native sync adds significant complexity.
Worked consequence: For a 3 MW data center using 3×1 MW Caterpillar C32 sets in N+1, you'll need to add about $35,000 in paralleling switchgear and a separate synchronizer. Equivalent Cummins QSK60 sets (3×1 MW) with PowerCommand 3.3 can parallel via a simple bus tie—no extra switchgear—saving roughly $30,000 in hardware plus two weeks of commissioning labor. That's about 4% of total project cost, but it shifts the reliability profile: fewer external components mean fewer failure points.
When this flips: If you only need a single generator (
Both manufacturers sell generator sets that meet EPA Tier 2 for stationary emergency standby. But how they meet it changes the available real watts. The Cummins QSK60 is certified Tier 2 without any aftertreatment (no DPF, no SCR). That means zero backpressure from exhaust filters, no regeneration cycles, no urea refills. The full 2000 kW standby is available at any time. Caterpillar's larger engines (C32 and up) often require an SCR system to meet Tier 2, depending on the injection calibration. An SCR system adds roughly 3–5 kPa of backpressure, which derates the engine by about 2–3% (roughly 20–30 kW on a 1000 kW set). That derating is rarely listed on the front-page datasheet—you find it in the installation manual.
Mechanism: Exhaust backpressure forces the turbocharger to work harder, increasing exhaust manifold pressure and reducing volumetric efficiency. The engine produces less torque at the same rack position. The control system compensates by reducing fuel, which lowers output. On a Caterpillar C32 with SCR, you might spec a 1000 kW standby set but only get 970 kW net at the terminals, after backpressure. The loss is proportionally small, but in a tightly-loaded N+1 configuration, 30 kW can be the difference between staying online and tripping.
Worked consequence: If you design a 2 MW site with two 1 MW Caterpillar C32 sets in parallel, your net capacity after SCR backpressure derating is about 1.94 MW. If your critical load is 1.9 MW, you have no margin—any transient load step (e.g., chiller start) could overload the system. With two Cummins QSK60 sets (each delivering full 2000 kW standby with no aftertreatment), you have 2.0 MW net, giving 100 kW margin. That margin eliminates the risk of nuisance trips during load switching.
When this flips: If your site is in a non-attainment area (e.g., California's South Coast AQMD) requiring Tier 4 final, both manufacturers will need DPF+SCR, and the Cummins advantage narrows to zero—both will have similar backpressure and similar derating. In that case, the choice comes down to dealer support and fuel consumption, where Caterpillar may win on fuel cost per kW·h.
| Dimension | Cummins (Host) | Caterpillar (Rival) | Decision Threshold |
|---|---|---|---|
| Standby rating usable continuous | ~85–90% of standby rating for hours | 70% average of standby rating per design | If continuous > 70%, Cummins saves 1–2 frame sizes |
| Fuel efficiency at 75% load | ~4.05 kW·h/L (illustrative, QSK60) | ~5.36 kW·h/L (illustrative, C32) | If run > 500 hrs/yr, Caterpillar saves ~$9k/yr |
| Native paralleling ≤ 20 MW | PowerCommand 3.3, no extra switchgear | EMCP 4.2 requires external sync | If total > 1.5 MW, Cummins saves $30k+ |
| Emissions backpressure derating | None (Tier 2 no aftertreatment) | ~2–3% derating with SCR | If tight margin, Cummins adds 20–30 kW net |
Values derived from manufacturer datasheets; illustrative fuel consumption based on published curves. Actual performance varies with load, temperature, and fuel quality.
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. Cummins is a brand affiliated with this site; competitor names are used for identification only.
