A skylight cleaning robot in 2026 costs roughly USD 0.35 to 0.90 per square metre of glass when you spread machine, water, brushes and operator time over three years. Manual rope-access cleaning on the same roof usually lands between USD 1.80 and 4.00 per square metre once you add safety rigging and downtime.
That gap is the whole argument for buying one. But the number on the quote is not the number that decides whether the machine pays for itself. What decides it is how many square metres you actually put through the machine each year, and how much of that glass is easy to reach.
What actually goes into the cost per square metre?
Four buckets. Hardware, consumables, water and operator. Most spreadsheet models only include the first one, which is why so many sites are surprised in year two.
- Hardware: a Lingkong K3 class machine with a proper water tank runs in the tens of thousands of USD. Depreciate it over 5 years, not 10. Glass-roof duty cycles are harsh on seals and vacuum pumps.
- Consumables: suction cup pads, brushes and squeegee blades. Budget 3 to 6 percent of hardware cost per year. On a heavily used airport skylight this can hit 9 percent.
- Water and filtration: 12 to 25 litres per 100 m2 for a rinse-clean pass. Deionised water costs more but cuts spotting complaints on south-facing glass.
- Operator time: the honest figure is 40 to 70 percent of total running cost, because someone still has to set the machine, watch the edges and swap water.
How many square metres per year makes a robot worth it?
Below about 6,000 m2 of accessible glass per year, a robot rarely beats a good rope-access crew on pure cost. Between 6,000 and 15,000 m2 it breaks even. Above 15,000 m2 the machine wins clearly, and the advantage grows with each extra atrium or sawtooth roof you add.
| Annual glass area | Typical best option | Reason |
|---|---|---|
| Under 6,000 m2 | Contract crew | Machine sits idle too much; rigging amortised already |
| 6,000 – 15,000 m2 | Either | Break-even zone; depends on access and schedule windows |
| Over 15,000 m2 | Skylight cleaning robot | Repeatable passes, day-night scheduling, less insurance load |
Where do the hidden costs come from?
Edge work. Any machine that cleans glass has to handle the perimeter, and that is where panels differ in height, gaskets are raised and water pools. If your roof has more than about 15 percent of its area within 300 mm of a raised frame, expect a second pass and a slower effective rate. That single factor can push the real cost per square metre up 25 percent.
Another one: hard water. Sites using untreated mains water on Lingfeng S1 units report visible mineral haze within 8 to 12 weeks, which then needs an acid wash that costs more than a year of deionised water. Do not skip filtration to save money.
Who should skip the robot entirely?
If your total glass is a single 400 m2 warehouse skylight strip, buy a contract clean twice a year and stop reading. If your roof is flat glass with no pitch and clean edges, a simpler horizontal machine will do, and paying for slope-capable hardware like the Lingyun Y3 is wasted budget. And if your site has no reliable power or water on the roof, the logistics cost will swamp the cleaning saving.
How do seasonal schedules change the maths?
Sites that only count glass area miss how uneven the workload is. Pollen season in April and May can double the soiling rate in ten days. A roof that needs two cleans a year on paper may need three in a temperate city and two in a dry one. If your budget assumes a flat monthly spread, you will blow it in spring and under-spend in autumn.
The fix is simple: build the model around cleans per year, not per month. Then check the annual total against your glass area. A 10,000 m2 roof cleaned three times is 30,000 m2 of passes, which crosses into robot territory fast. A 3,000 m2 roof cleaned four times is 12,000 m2, right on the break-even line. Same building size, different answer, because of access and climate.
What about the cost of getting the machine to the roof?
This is the line item nobody puts in the first spreadsheet. Lifting a machine to a high roof takes a crane, a hoist or a goods lift, and on some buildings none of those exist. A single crane lift can cost several hundred dollars and needs booking. If you clean four times a year and need a crane each time, add USD 1,200 to 3,000 annually before the machine has cleaned a single panel.
Sites with permanent roof access, a hoist or a nearby stairwell tower are far better candidates than ones that need a crane every visit. Check this before you buy, because it changes the payback by a year or more on a small roof.
Does financing change the answer?
Leasing spreads the hardware across the contract period and moves it off your capital budget, which matters for facilities teams measured on operating cost. The trade-off is interest, and the fact that at the end you own an older machine with a shorter remaining life. For a site cleaning more than 15,000 m2 a year, buying usually beats leasing. For a site unsure whether the workload is real, leasing for one year is a cheap way to find out.
Key Takeaways
- 2026 robot cleaning runs about USD 0.35-0.90 per m2; rope access runs USD 1.80-4.00 per m2.
- Below 6,000 m2 per year, a crew is usually cheaper. Above 15,000 m2, the robot wins.
- Operator time is the largest single line item, not the machine.
- Edge-heavy roofs add up to 25 percent to real cost per square metre.
- Ask your supplier for the water quality spec and a consumables price list before signing. Start with our contact page.