Gausium Phantas: What the 4-in-1 Cleaning Robot Really Does

The Phantas packs four floor-cleaning modes into one compact chassis: vacuuming, sweeping, scrubbing and dust mopping. Gausium gives the current generation as 585 × 445 × 619 mm, with a 330 mm working width for scrubbing and dust mopping, a 465 mm width for vacuuming and sweeping, and a 490 mm squeegee. The official minimum passing width is 650 mm — small enough to work under tables and through narrow aisles that larger scrubbers simply cannot enter. The TASKI OEM datasheet describes the machine as "designed to excel in areas up to 1,000 m²", which is the honest size bracket to plan with.

It is a hybrid, not a pure robot. Pull the handlebar up and the machine switches to manual mode: the operator pushes and steers it like a walk-behind machine, with five modes selectable from the handle. Lower the handlebar and it goes autonomous again. The manual is explicit about the staffing model — nobody has to walk alongside the robot, but a designated, specially trained operator must be reachable to handle its alerts.

Now the part that causes most of the confusion in German offers: there are two hardware generations in circulation. Diversey/Solenis sells the previous build under its own badge as "TASKI GS Phantas 1.2" (product code 9293373) with 54 × 44 × 62 cm, 48 kg, a maximum of 700 m²/h and a realistic 200–400 m²/h. Gausium's own current page sells the S1 Pro generation with different dimensions and higher figures. German distributor pages mix the two freely — a 410 mm width, for instance, belongs to the older build and to the manual's own differing width nomenclature, not to the 465 mm the current spec page states. Ask which generation is in the quote before you compare anything.

There is also a "Phantas Extra" edition on the official product page, advertised with AI debris and stain detection, residual-water detection and adaptive cleaning, headlined as "400% Efficiency Improvement" and described in the body text as "up to 4×". Treat that as a manufacturer claim: Gausium publishes no separate specification sheet for the Extra, so there is no verifiable performance delta against the standard machine — only a feature list.

Start with what is solid. Gausium states a practical cleaning efficiency of 350–700 m²/h; the operating manual is more granular and gives 400 m²/h for scrubbing, vacuuming and dust mopping and 700 m²/h for sweeping. Tanks are 11.5 litres fresh and 10.5 litres waste, plus a 6-litre dust bag. The battery is lithium iron phosphate, 24 V and 40 Ah, rated for 2,000 recharge cycles to 80 % capacity, with a charging time of about two hours. Official runtimes: 4.5 hours scrubbing, 4.5 hours vacuuming, 6 hours sweeping, 10 hours dust mopping — the manual's average operating times run slightly lower at 3.5–4 hours vacuuming and 4–4.5 hours scrubbing. Top speed is 0.8 m/s, noise ≤65 dB(A) with a quiet mode at ≤60 dB, protection rating IPX3, operating range 0–45 °C.

Now the contradictions, and there are several. Weight: the official page lists "Net Weight 65 kg", the manual gives 65 kg as gross vehicle weight and 53 kg as transportation weight, and the TASKI 1.2 build is 48 kg. Three numbers, no reconciliation — which is why we publish none of them as the weight and recommend you get a figure in writing for the exact build you are buying. Gradeability: 5° officially, 8° in the manual. Minimum passing width: 650 mm officially, 600 mm in the manual's parameter table, 52 cm in the manual's own body text — plan with 650 mm. Obstacle height: 20 mm officially, 17 mm in the manual. Trash box: 1 litre officially, 0.7 litres in the manual.

Two things are missing entirely, and both matter commercially. First, there is no suction figure in pascal or kilopascal for the Phantas anywhere in the official material — the only suction-related number that exists is an airflow rating of 1.5 m³/min in the TASKI datasheet. If a dealer quotes you a pascal figure for this machine, ask which official document it comes from. Second, sweeping runtime inflates badly downstream: Gausium states 6 hours sweeping and 10 hours dust mopping, while German distributor pages advertise 12, 14 and even 18 hours. Use the manufacturer's numbers.

Finally the operational truth, because specifications hide it. Scrubbing is the only wet mode, with a rated spray rate of 80 ml/min and a 490 mm squeegee behind it — like every scrubber it leaves a thin film that dries quickly, and the wet edge distance is 60 mm, so skirting boards and wall edges need a hand finish. The 0 mm edge-cleaning figure applies to the dry modes with the side brush. Sweeping accepts pieces of at most 20 × 7 × 7 mm and 8 g into a 0.7–1 litre box: dust, crumbs and light litter, not coarse debris. And the manual names a hard limitation you should test before signing — environments surrounded by glass walls or similarly transparent materials degrade the sensors.

Our benchmark in this class is the Pudu CC1, the 4-in-1 compact machine we run daily — so the comparison is between the Phantas as documented and a machine we know from operation. The CC1 is officially rated at 700–1,000 m²/h in full-coverage cleaning, 5 hours runtime, 15 + 15 litre tanks, 500 mm cleaning width, below 70 dB(A), a published maximum suction of 17,000 Pa, 75 kg, and a sweet spot of 500–3,000 m² at 629 × 552 × 695 mm. The Phantas is the smaller machine on every axis: 350–700 m²/h practical, 4.5 hours wet runtime, 11.5 + 10.5 litre tanks, 330 mm scrubbing width, ≤65 dB(A) with a ≤60 dB quiet mode, no published suction figure, and a size bracket up to roughly 1,000 m².

Two things genuinely differentiate it rather than just making it smaller. The first is the WS-03-S workstation: according to the official product page it charges the machine, refills fresh water, drains the waste water and self-cleans the squeegee blade, the suction pipe and the waste water tank to prevent clogging and secondary contamination. That automates the daily manual work a charging dock leaves behind — but it demands infrastructure in return, because the workstation needs a water supply, a floor drain, power, and two metres of clearance in front of it. The second is certification: TÜV Rheinland tested the build designated "Phantas S1 Pro M" against EN IEC 63327, confirmed PL=d functional safety and issued a CE Machinery Directive certificate in March 2024. In German B2B tendering that is a real trust signal — but it attaches to that specific build, so have the certificate confirmed for the machine you are actually being delivered.

Where the CC1 stays ahead is equally clear: more area per hour, larger tanks, a published suction figure, and in our case an established service and parts operation behind it. The honest conclusion is that the Phantas does not replace the CC1, it sits below and beside it — for furniture-dense floors under roughly 1,000 m² where the CC1 is oversized, and for accounts that require certified machinery safety. If you are still deciding within the CC1 family, our comparison of CC1 and CC1 Pro covers that split; if you are earlier in the process and unsure whether a scrubber is the right category at all, start with the scrubber robot guide.

Start with what is documented, not with what is suspected. The official sensor suite includes an RGB camera and 3D depth cameras alongside the 2D LiDAR — the machine sees, in the literal sense. Connectivity, per the TASKI datasheet, is a SIM card, meaning the robot carries its own mobile link and does not depend on your network. And on storage the same OEM datasheet is unusually direct: "Both, mostly on-device. Key data, such as cleaning performance and map coverage, is sent to the cloud." So there is a cloud, and operational data goes into it.

What the manufacturer says about that cloud is thinner than you would want. Gausium's privacy policy states that collected personal information is stored on Amazon cloud servers "in your country or region" and that the cloud server region can be selected based on the user's location; it also contains a clause about transferring domestically collected personal information to overseas institutions. It names no concrete EU location and does not mention the GDPR. Separately, a February 2026 statement attributed to Gausium's CTO says the company is "fully GDPR-compliant", collects training data exclusively from the China market, and deploys validated software updates to overseas fleets without collecting customer data from international markets. That claim does not stand unopposed. The Clingendael Institute, the Netherlands Institute of International Relations, reported that data from Gausium robots is stored in Singapore and questioned whether it enjoys GDPR protection there. Manufacturer and think tank contradict each other, nothing in the public record resolves it, and we are not going to resolve it here — which is precisely why this point belongs in a contract rather than in a brochure.

None of that is a contract, and that is the whole point. For any cloud-connected cleaning robot carrying cameras — regardless of manufacturer or country of origin — a German buyer should have five things in writing before deployment: a data processing agreement under Art. 28 GDPR, the named storage region for operational and map data, the list of sub-processors, what happens to camera imagery and how long it is retained, and the technical and organisational measures. For public-sector tenders, hospitals and works-council-sensitive sites, that is not paperwork, it is a knock-out criterion. Our position is simply that the burden of proof is documentary: a vendor who can answer these in a signed annex is a safe vendor, and one who answers them in a blog post is not yet.

Shortlist the Phantas if your floor is hard, furniture-dense and under roughly 1,000 m² per run: offices with desk clusters and narrow circulation, smaller retail formats, hotel lobbies and mezzanine levels. Gausium's own reference base points the same way — the InterContinental Wien runs a Phantas on its mezzanine alongside a larger Vacuum 40 on guest floors and in the lobby. Two further arguments carry real weight: daytime operation, where ≤65 dB(A) and the ≤60 dB quiet mode are quiet figures as manufacturer figures go — and manufacturer figures are all that exist here, and the WS-03-S workstation, if daily water filling and draining is genuinely eating your operator's time.

Do not shortlist it if you have coarse debris — an 8 g, 20 × 7 × 7 mm intake limit is a dust and light-litter specification, not a sweeper. Same for areas well above 1,000 m² per run, anything outdoors (IPX3, indoor only), and situations where you need a documented suction figure for a specification sheet, because none exists. If you are unsure whether this class is right at all, our guide to choosing the right cleaning robot sorts the categories before you sort the brands.

And when you get to the demo, test the things the datasheet cannot settle. Have the generation and build confirmed in writing, with the EN IEC 63327 certificate for that exact build. Get a weight figure — and ask explicitly whether it is empty or with full tanks, because you need the right one for lifts and floor loads. Measure the 650 mm minimum passing width against your narrowest real aisle, not the average one. Run a scrubbing pass along your worst wall and look at the 60 mm edge strip that stays untouched. Drive it through your glass-walled areas. Measure the quiet mode at the actual operating point rather than trusting the number. And ask for the data processing agreement before, not after, the pilot. If you want that same list worked through with us in the room, that is what our demos are for.