Battery Runtime in Cleaning Robots: How to Read the Figures

Start with the fact that trips up most buyers: runtime describes a mode, not a machine. Pudu states four separate values for the CC1 — 5 hours scrubbing, 5 hours for the combined sweeping, vacuuming and mopping programme, 4 hours carpet vacuuming, 9 hours silent mopping. The frequently quoted "up to 5 hours" is therefore the scrubbing figure, not a maximum; the maximum is 9 hours in the mode that uses the least energy. The NR 1700 is rated at 250 minutes in Eco, 110 in Balanced and 105 in Performance — the fastest mode runs for roughly 40 per cent of the longest one. Gausium publishes 4.5 hours for scrubbing and vacuuming, 6 hours for sweeping and 10 hours for dust mopping on the same Phantas.

Next: the measuring conditions. Pudu is unusually candid here — the CC1 operation guide carries a footnote stating that the runtimes were tested in the various cleaning modes under laboratory conditions, at the lowest gear setting, starting from a 100 % charge and continuing until 10 % was reached. Lowest gear means slowest travel speed, which is also the least demanding case. The official CC1 product page prints exactly the same four numbers and omits the footnote entirely. Nexaro does something comparable and worth knowing: its datasheet footnotes the three runtimes as measured on hard floor, while the noise values on the same page are footnoted as measured on carpet. Two rows, two reference surfaces — for a vacuum robot whose energy draw depends heavily on which of the two it is driving over.

Third: the battery fit-out being quoted. The Adlatus SR1300 is the clearest case in this market. Its official datasheet describes a scalable pack of three to eight battery modules, 180 Ah in the standard version and expandable to 480 Ah, and then states that this makes a runtime of up to eight hours possible. Eight hours is the top of that range, not the machine as delivered — the standard version carries 180 Ah, and the figure of up to 2.5 hours per charge for it comes from a Swiss dealer, not from Adlatus. The SR1300 is available through us but newly onboarded; everything we write here about it, the Kemaro K900 and the Gausium Phantas comes from manufacturer documentation, which we cover separately in our market overviews of the SR1300, the K900 Gen II and the Phantas.

And finally the version history that nobody updates. Kemaro rates the K900 Gen II at up to 4 hours on its current factsheet and in its November 2025 press release. The manufacturer’s own knowledge base still lists "~5 hours runtime" — the first-generation value, which the 2023 site carried and which several dealer pages continue to quote. Gausium has the same problem in the other direction: its official sweeping runtime for the Phantas is 6 hours, while several distributor listings circulate far higher figures. When two sources from the same manufacturer disagree, the current official document wins, and the older figure is history, not a spec.

What manufacturers actually publish varies more than you would expect. Pudu documents 50 Ah for the CC1; the lithium iron phosphate (LiFePO₄) chemistry appears in the CC1 Pro documentation for the shared platform. Gausium publishes lithium iron phosphate, 24 V, 40 Ah for the Phantas. Nexaro states two lithium-ion packs at 14.68 V and roughly 75 Wh each for the NR 1700 — a different chemistry family, and a much smaller machine. Kemaro gives "Lithium Ionen / 36 V" for the K900 Gen II without a capacity figure. Adlatus publishes capacity in detail, 180 to 480 Ah, and no chemistry at all. Two of these five manufacturers therefore leave out exactly the half of the specification you would need to compare them.

The qualitative background is uncontroversial and worth having. LiFePO₄ is a lithium-ion variant whose selling points in commercial machines are cycle life and thermal robustness; the trade-off is lower energy density, so an LFP pack tends to be heavier and bulkier than another lithium-ion chemistry of the same energy content. That is background, not a spec. Where a manufacturer publishes no chemistry, we do not infer one, and where a manufacturer publishes no cycle count, we do not borrow a typical value from the chemistry — which is the single most common way inflated battery claims enter a quotation.

A comparability warning, because the two best-documented figures in this article invite exactly the wrong conclusion. Gausium states 2,000 recharge cycles down to 80 % of capacity for the Phantas. Nexaro states approximately 70 % remaining capacity after approximately 800 full cycles for the NR 1700. Those are not the same measurement: different end-of-life thresholds (80 % versus 70 %), different machines, different duty cycles, different pack sizes. They tell you that both manufacturers publish a degradation figure at all — which is rare and good. They do not tell you that one battery is 2.5 times better than the other, and anyone who presents them that way is doing arithmetic on incompatible units.

Area per charge is the figure runtime is standing in for, and it is a product of two footnoted numbers: the cleaning rate in the mode you will actually run, multiplied by the runtime in that same mode. Both factors carry conditions. Pudu rates the CC1 at 700–1,000 m²/h and, in scrubbing, at 5 hours measured in the laboratory at the lowest gear. Nexaro rates the NR 1700 at up to 100 m²/h with the explicit caveat that the figure depends on floor plan, furnishing, floor covering and soiling. Multiply two optimistic numbers and you get a very optimistic third one — which is why we treat area per charge as something to measure on site, not to calculate in a quotation.

Charge time against the cleaning window is the constraint people discover after signing. Pudu charges the CC1 in under three hours. Kemaro charges the K900 Gen II in 90 minutes: with four hours of runtime, an eight-hour night window yields roughly six and a half hours of cleaning across two runs, not eight. Adlatus is the instructive one again — the standard charging equipment takes five hours, and only the double charging configuration halves that to 2.5 hours. A standard SR1300 that runs 2.5 hours and then charges for five has used 7.5 hours of an eight-hour window for 2.5 hours of work. And Nexaro is a reminder to check which document you are reading: the current datasheet gives roughly 120 minutes from 15 to 95 %, the operating manual gives roughly 195 minutes for the same span. Both are official, they contradict each other, and the difference decides whether a second run fits into the night.

Opportunity charging is what makes long windows work at all. The CC1 returns to its docking station when the battery runs low and resumes the interrupted job afterwards, which is the only honest way to read claims like "cleans all night on one charge": a night shift is six to eight hours, the scrubbing runtime is five, and the gap is closed by a recharge that costs window time rather than by a bigger battery. The Whiz answers the same problem differently, with swappable batteries that let an operator continue across shifts without a charging break. Which approach suits you depends on whether anyone is on site at all — the trade-offs are laid out in our scrubber-dryer guide and, on the commercial side, in what actually drives the cost of a cleaning robot.

Degradation is the number almost nobody asks about and almost nobody publishes. Gausium does: 2,000 recharge cycles to 80 % capacity, printed in the official Phantas manual. Nexaro does too, with approximately 70 % after approximately 800 full cycles plus a specified service life of 3,000 operating hours. Kemaro answers the question commercially rather than technically, with a warranty of three years or 5,000 operating hours across the Gen II series. Those three are the transparent minority. For every machine where you find nothing, the honest planning assumption is that capacity will fall over the contract term and that the runtime in year three is not the runtime in the brochure.

The calculation is short. Take the area that has to be cleaned in one pass, divide it by a realistic cleaning rate for the mode you will actually run, and you get the hours of cleaning required. Then subtract from your window the charge time that will be needed inside it. If the remainder is negative, the answer is not a longer runtime — it is a second machine, a second window, or a mode that costs less energy per square metre. Doing it in that order also stops the usual mistake of comparing a nominal runtime against a real cleaning window.

A worked example with verified figures: a 1,200 m² sales floor scrubbed by a CC1 at the official 700–1,000 m²/h takes between 1.2 and 1.7 hours. That sits comfortably inside the 5-hour scrubbing runtime, and charge time never becomes a constraint. Scale the same store up to 3,500 m² and the same arithmetic gives 3.5 to 5 hours — at the pessimistic end of the manufacturer’s own range, the job now consumes the entire scrubbing runtime, and any interruption pushes it into a recharge. The machine did not change. The margin did.

The last correction is the one only field data can supply. Datasheet rates are measured under conditions your building does not reproduce, and the gap is not a rounding error: for the NR 1700 we publish both the manufacturer figure and our own deployment figures side by side in the NR 1700 in operation, and the difference is large enough to change how many machines a site needs. Plan with the lower number, in supermarkets as in warehouses and offices. If you are still choosing a class of machine rather than sizing one, which cleaning robot suits which operation works through that decision first.

Every claim in this article came out of a document that was freely available and that almost nobody reads to the end. You can get the same information in a sales conversation by asking five questions, and the quality of the answers tells you as much as the answers themselves.

First: what is the runtime in each individual cleaning mode, and which mode will this site actually run? A single headline figure is an incomplete answer. Second: under what conditions was that runtime measured — which gear or speed setting, on which floor covering, and from which state of charge down to which? Pudu writes it into the manual; if a supplier cannot say, the figure is unqualified. Third: which battery fit-out does the quoted runtime refer to, and is that the fit-out in this offer? On a scalable machine such as the SR1300 this single question separates eight hours from 2.5. Fourth: what does the machine charge in, and does the offer include the faster charging configuration where one exists? Fifth: what is published on battery ageing — cycles, remaining capacity at end of life, specified operating hours — and what do the warranty terms say about the battery specifically, as opposed to the machine?

Then ask for the sixth thing, which is not a question but a measurement: let the machine run in your building, in your mode, on your floor, and record the area it covers on one charge. We do exactly that in a demo before quoting anything, because it is the only number in this whole subject that is measured under the conditions you will actually pay for.