How long will a power station run a laptop?
Power basics Updated 2026-08-08
Longer than almost anyone expects, because a laptop is a far smaller load than the wattage printed on its charger suggests. The US certified computer dataset puts short-idle power for a 13-to-16-inch notebook at 5.23 W across 1,016 models, and one manufacturer’s own measurement of a current 14-inch machine, display on, is 3.91 W.
At that size the power station’s own housekeeping becomes the larger half of the drain. The biggest machine here states a self-consumption of 10 W against the laptop’s 5.23 W, so 66% of what leaves the battery is the box keeping itself switched on — and it still runs the laptop for 104 h.
The charger’s wattage is not the laptop’s wattage
The number most people reach for is the one moulded into the power brick. That figure describes how fast the charger can refill a flat battery — a peak capability, not a steady draw — and the record’s own distribution note makes the same point. A notebook doing ordinary work with its own battery already full sits in the single digits of watts, and the certified dataset below is measured rather than rated.
What a notebook actually draws
Two independent measurements of the same class of machine: a federal certification dataset covering a thousand models, and one manufacturer’s published environmental report for a specific one.
The mean across 1,016 certified models. Individual machines in the same set run from 0.2 W to 12.6 W.
See the record →Apple’s own product environmental report, taken with display brightness set by the certification programme’s method and auto-brightness off.
Roughly one part in twenty of the running figure, and far under any power station’s own idle draw.
Official specverified 2026-08-01laptop record · certified dataset + manufacturer report
Below about ten watts, the power station is the load
Runtime on this site is usable capacity divided by the load plus the machine’s own self-consumption, because that is the form BLUETTI’s manual states its runtime formula in. For a kettle or a heater the second term is a rounding correction. For a laptop it is most of the denominator: 10 W of self-consumption against 5.23 W of laptop is 66% of the drain going nowhere near the computer. It is also why a runtime worked out as capacity divided by the load alone is not merely optimistic at this size — it is roughly three times too long.
A notebook on two machines, computed on this build
Both are trivially inside continuous output; the load-ratio column gives that away. What the two rows are really showing is how little of a battery a laptop consumes once nothing else is plugged in.
17 h from the smallest machine in the library — an overnight and most of the next day.
Show the working →104 h. Four days of laptop, with about two-thirds of what leaves the battery going to the machine’s own idle draw rather than to the computer.
Show the working →The 5,750 VA on the record, and why no verdict turns on it
Cold-start inrush, from an adapter data sheet
A switching power supply charges its input capacitors the instant it is plugged in. It is a real event with a real published figure, and it decides nothing here.
From a third-party 65 W adapter’s own specification rather than a notebook maker’s. The counter-intuitive part: peak inrush barely scales with the size of the supply, so the smallest devices carry the highest multiples.
See the record →Peak current times nominal voltage, over about 50 ms.
Official specverified 2026-08-01laptop record · 65 W adapter data sheet
A power station’s surge rating describes what it will hold for seconds. This is a capacitive spike two orders of magnitude shorter, so the two figures share a unit and describe different things — comparing them would refuse pairings that work perfectly well. Our engine therefore treats this one as advisory: it is on the record, it is dated, and it never decides a verdict. Where it does become practical advice is a strip of chargers switched on together, because the spikes add and a machine can trip on nothing but being plugged in.