Skip to content
T

Cleaning review

Shark AI Ultra Review: It Cannot See a Sock

Two sources say the same thing in different words: the mapping is excellent and the obstacle detection is not. Clear the floor before you press start.

By Paulo GerchonPublished 2 min read

Placeholder graphic, not a product photo

Illustration, not a photo — we don’t have a licensed image of this product yet.

7.8/10

Shark

Shark AI Ultra

Good · researched from 3 sources

The cleaning numbers are among the best we have seen and the self-cleaning brushroll genuinely solves hair wrapping. What it cannot do is see what is on the floor, so this suits a tidy house and frustrates a cluttered one.

Buy if
Homes with pets, on the brushroll and pet hair results · Tidy floors where obstacles are not an issue · Anyone who wants vacuuming automated without mopping
Skip if
Floors with cables, toys or laundry on them · Quiet flats, given the emptying noise · Anyone who wants a combined vacuum and mop

Excellent at cleaning, blind to clutter

Sparkling Picks ran it 30 times and it returned to its base 30 times. The lidar identified all six rooms correctly without manual correction. The Home Picker's 60-day test reports the same mapping accuracy.

And all three sources agree it cannot see a cable, a toy or a sock.

Gadget Review makes the comparison explicit: unlike the Roomba j7+, this will not identify and avoid small objects, so cluttered floors mean tangles and stoppages. The Home Picker got stuck once in 60 days, under a low ottoman, and says plainly that a quick floor pickup is needed before each run.

That is the whole decision. On a clear floor this is excellent. On a floor with things on it, you will be rescuing it.

The cleaning numbers

Gadget Review: 99.8 percent on hard floors, 95.75 percent on carpet.

Sparkling Picks: 98 percent on hard floors in one pass, 92 on low pile, 89 on medium carpet with pet hair.

The Home Picker: roughly 95 percent of fine particles on the first pass on hardwood, with suction ramping automatically at carpet transitions.

Three independent tests, all in the same range, all high. Whatever else is contested about this robot, the cleaning is not.

The brushroll is the pet-owner feature

The Home Picker lived with this for 60 days alongside a heavily shedding golden retriever.

Their finding: minimal hair wrapping on the brushroll across the whole period, with no scissor maintenance required at all, which they contrast against previous vacuums needing weekly manual clearing. They estimate 30 minutes a month saved.

Anyone who has cut hair off a brushroll with a seam ripper will understand why that is worth more than a few percentage points of pickup.

Two numbers nobody agrees on

Suction: Gadget Review says 2,100 pascals and calls that low for the class. Sparkling Picks says 2,500.

Runtime: Gadget Review says about 172 minutes. Sparkling Picks and The Home Picker both say about 120.

The runtime gap is probably power mode, since eco runs far longer than maximum, and The Home Picker covered 1,800 square feet in 55 minutes which fits either figure. The suction figures simply differ, and 400 pascals is not a rounding error.

Noise, and the twelve seconds that matter

The Home Picker, from six feet: 54 dB in eco, 62 in normal, 70 at maximum.

Gadget Review: 79.8 dB, one of the loudest on their tested list.

Sparkling Picks: 60 dB running, 75 during emptying.

The consistent point is the self-empty cycle. The Home Picker measured 78 dB for about twelve seconds and compares it to a leaf blower running indoors. It is brief, it is loud, and it will wake a sleeping child.

The running cost

A bagged base rather than a bagless one, which The Home Picker confirms lasted the claimed 60 days in an 1,800 square foot home with moderate pet shedding.

Bags cost around $15 for three, or roughly $30 a year.

The Home Picker's other note is worth weighing at this price: there is no mopping at all, where several rivals at the same money offer it.

How it scored

The overall 7.8 is the weighted average of these five criteria — we never adjust it by hand.

Cleaning performance

25% of score9.0

Gadget Review recorded 99.8 percent on hard floors and 95.75 percent on carpet. Sparkling Picks got 98 percent on hard floors, 92 on low pile and 89 on carpet with pet hair.

Value

25% of score8.5

Sparkling Picks scored it 8.8 out of 10 and made it their best value pick. Gadget Review scored it 83 out of 100.

Build quality

17% of score8.0

A self-cleaning brushroll that The Home Picker says virtually eliminated hair tangles across 60 days with a golden retriever in the house.

Ease of use

17% of score6.5

Lidar mapping identified six rooms correctly for both The Home Picker and Sparkling Picks. Both criticise the app, and neither can get the robot to see a cable or a sock.

Noise

17% of score6.0

Gadget Review measured 79.8 dB and calls it one of the loudest they have tested. The Home Picker measured 54 to 70 dB by mode with 78 dB during emptying.

Pros and cons

What we liked

  • Gadget Review recorded 99.8 percent debris removal on hard floors
  • A self-cleaning brushroll that eliminated hair wrapping over 60 days per The Home Picker
  • Lidar mapping that identified all six rooms correctly in two separate tests
  • Sparkling Picks recorded a 100 percent return-to-base rate across 30 sessions
  • A bagged self-empty base that The Home Picker confirms lasted the claimed 60 days
  • The Home Picker measured 54 dB in eco mode, which is quiet for a robot vacuum

What we didn’t

  • It cannot identify cables, toys or socks, which both Sparkling Picks and Gadget Review confirm
  • Gadget Review measured 79.8 dB and calls it one of the loudest they have tested
  • The two sources give suction figures 400 pascals apart
  • The Home Picker notes there is no mopping at a price where rivals offer it
  • Runtime figures differ by nearly an hour between sources
  • Replacement bags are a recurring cost of about $30 a year per The Home Picker

Specifications

Specifications
Hard floor pickup99.8 percent per Gadget Review; 98 percent per Sparkling Picks
Carpet pickup95.75 percent per Gadget Review; 92 percent low pile and 89 percent with pet hair per Sparkling Picks
Suction2,100 Pa per Gadget Review; 2,500 Pa per Sparkling Picks
RuntimeAbout 120 minutes per Sparkling Picks and The Home Picker; 172 minutes per Gadget Review
Navigation360 degree lidar mapping
Obstacle detectionNone for small objects such as cables and socks
BaseBagged self-empty, about 60 days of capacity
Robot bin350 ml per Sparkling Picks
Noise79.8 dB per Gadget Review; 54 to 70 dB by mode and 78 during emptying per The Home Picker
BagsAbout $15 for three, roughly $30 a year per The Home Picker

How we researched this

We have not had this unit on the bench. This review synthesises the manufacturer’s published specifications, independent measurements and owner reports listed below. Every figure on this page comes from one of those sources rather than our own measurement.

3

sources reviewed

Gadget Review supplies the highest pickup percentages, one suction figure, the bin capacity, the loudest noise measurement and the runtime figure that differs from the others, and scored it 83 out of 100. Sparkling Picks supplies a second set of pickup percentages by surface, the return-to-base success rate, the mapping result and the app assessment, and scored it 8.8 out of 10. The Home Picker ran it for 60 days and supplies the pet hair and brushroll findings, the noise readings by mode at a stated distance, the base capacity confirmation and the consumables cost. The sources disagree on suction by 400 pascals and on runtime by nearly an hour, and both spreads are reported.

Our full testing policy

Common questions

Does the Shark AI Ultra avoid obstacles?
It avoids furniture and nothing smaller. Sparkling Picks is precise: the 360 degree lidar identified all six rooms in their test and returned to base successfully 30 times out of 30, but obstacle detection cannot identify cables, toys or socks. Gadget Review reaches the same conclusion, noting that unlike the Roomba j7+ it will not identify and avoid small objects, so cluttered floors mean tangles and stoppages. The Home Picker, over 60 days, got stuck once under a low ottoman and confirms needing a quick floor pickup before each run. If your floor has things on it, this is not the robot.
How well does it actually clean?
Among the best figures in this catalogue. Gadget Review recorded 99.8 percent debris removal on hard floors and 95.75 percent on carpet, placing it in the top third of the robots they have tested. Sparkling Picks measured 98 percent on hard floors in one pass, 92 percent on low pile and 89 percent on medium carpet with pet hair. The Home Picker's 60-day account describes roughly 95 percent of fine particles captured on the first pass on hardwood, with suction ramping automatically on carpet transitions.
How is it with pet hair?
This is where The Home Picker's long test is most useful. With a heavily shedding golden retriever in the house, they report minimal hair wrapping accumulating on the brushroll across 60 days, requiring no scissor maintenance at all, which they describe as a dramatic improvement over previous vacuums needing weekly manual clearing. Sparkling Picks' 89 percent figure on medium carpet with pet hair is the measured version of the same thing. The self-cleaning brushroll is the feature doing the work.
How loud is it?
Loud, and the sources disagree about how loud. Gadget Review measured 79.8 dB at maximum and calls it one of the loudest on their tested list. The Home Picker measured from six feet and got 54 dB in eco, 62 in normal and 70 at maximum, with the self-empty cycle hitting 78 dB for about twelve seconds, which they compare to a leaf blower running indoors. Sparkling Picks reports 60 dB in operation and 75 during emptying. All three agree the emptying cycle is the loud part and that it is brief.
How long does it run, and what is the suction?
Both figures are disputed. Gadget Review puts runtime at about 172 minutes and suction at 2,100 pascals, describing that as low for a mid-range robot. Sparkling Picks says 2,500 pascals and about 120 minutes. The Home Picker also reports 120 minutes, and covered a full downstairs of roughly 1,800 square feet in about 55 minutes. The runtime gap is probably about which power mode was used, since eco mode runs far longer than maximum. On suction the two published figures simply differ.
7.8/10
Good

The bottom line on the Shark AI Ultra

The cleaning numbers are among the best we have seen and the self-cleaning brushroll genuinely solves hair wrapping. What it cannot do is see what is on the floor, so this suits a tidy house and frustrates a cluttered one.

More home & kitchen reviews

Placeholder graphic, not a product photo
6.5/10

Cleaning

Bissell CrossWave Cordless Max

Gadget Review measured suction at 0.34 kPa, which is very low. On sealed hard floors that barely matters, and on carpet it is the whole story.

Placeholder graphic, not a product photo
7.8/10

Cleaning

Bissell Little Green

Blood, cat sick and coffee all came out in one session. A set-in chain oil stain lightened by about half and stayed there, which is the honest boundary of what this machine does.

Bissell SpotClean Pro portable carpet cleaner in grey and blue with its hose, cleaning tool and two tanks
7.3/10

Cleaning

Bissell SpotClean Pro

A mud stain came out in under a minute. Set-in coffee needed a pre-treatment and still left a mark. That boundary is the most useful thing to know about it.