Carbon Robotics LaserWeeder G2
Made by Carbon Robotics
Photo: Carbon Robotics
- compute
- multi-GPU NVIDIA onboard
- geography
- customers in 15 countries
- precision
- sub-millimeter
- widths ft
- 6.6 - 60 (G2 200 through G2 1800)
- plant model
- Large Plant Model (Feb 2026), trained on more than 150 million images from the fleet
- model lineup
- G2 200, G2 300, G2 400, G2 600, G2 1200, G2 1800
- herbicide use
- none: laser ablation, no chemical input
- kill rate pct
- 99
- weeds per hour g2 600
- 600000
- max weeds per minute g2 600
- 10000
Who's exposed
Jobs in the threat radius
- manual weeding crew member (hand-hoeing crew)
- field sprayer operator (herbicide spray runs)
- organic farm weed scout
- row crop hand laborer
Deployment status
[Re-verified 2026-10-25: figures updated; one earlier claim withdrawn.] Carbon Robotics' May 2026 release says revenue passed $100 million in the fiscal year ending January 31, 2026, with systems covering more than half a million acres and over 40 billion weeds eliminated, for customers in 15 countries. TechCrunch, in February 2026, put the fleet on more than 100 farms in 15 countries and the venture funding at more than $185 million. That conflicts with the 'hundreds of customers' wording an earlier version of this entry carried, which we cannot source, so we withdrew it. No source gives a machine count. The G2 family, introduced in 2025, replaced the original single-size LaserWeeder with six widths from 6.6 ft to 60 ft. In February 2026 the company launched a Large Plant Model, trained on more than 150 million images from its own fleet, that lets operators retarget the laser to new weed species without retraining.
When this hits the labor market
Already at commercial scale on specialty row crops (onions, carrots, leafy greens, sugar beets) where hand-hoeing crews have historically been the only viable weed control. 1-3 years for broader adoption as G2's smaller widths make the per-acre math close on mid-size diversified vegetable farms. 3-5 years for the technology to push into commodity row crops where herbicide is still the default, as labor cost and herbicide-resistance pressure compound. The displaced labor is concrete: seasonal hand-weeding crews on specialty vegetable farms, paid by the row.
Is this robot coming for your job?
Press releases speak in averages — LostJobs talks about your role. Find out which parts of your job hold up and which don't. Free.
Talk to LostJobs about my future
A laser instead of a hoe
LaserWeeder G2 is a tractor-pulled implement that drives down a crop row, scans each plant with onboard cameras and NVIDIA GPUs, classifies it as crop or weed in real time, and fires a laser at the weeds. Sub-millimeter accuracy, no herbicide. The G2 600 hits up to 10,000 weeds per minute, 600,000 an hour.
The 2025 G2 launch broke the single-size barrier. The first LaserWeeder suited only the largest specialty-crop operations. G2 ships in six widths from 6.6 ft to 60 ft, with the smaller ones aimed at mid-size and diversified vegetable farms.
What the numbers say now
The company’s May 2026 release is the strongest evidence we have. Revenue passed $100 million in the fiscal year ending January 31, covering more than half a million acres and, by the company’s count, over 40 billion weeds. The fleet sits in 15 countries.
The customer count is softer than we first wrote. TechCrunch’s February report says more than 100 farms, not hundreds, and neither it nor the company gives a machine count. A $100 million year from a few hundred machines is plausible at these prices. We just cannot show it.
In February the company added a Large Plant Model, trained on over 150 million images collected by its own machines. Operators can retarget the laser at a new weed with a tap. That cuts the main brake on the technology, which was teaching the machine each new field.
Why we care for LostJobs
The exposed labor is not subtle. Hand-weeding crews on onion, carrot, leafy green and sugar beet farms were the irreplaceable layer of those operations: contract workers, paid by the row, often seasonal. A G2 in the field takes the weeding work out of the crew. The supervisor’s job survives longer than the crew member’s. The first wave was the largest farms. The next, on the smaller widths and the plant model, is the farms that were too small for the argument to land, and after that the commodity row crops, where herbicide resistance and labor cost start to close the gap and field sprayer operators come under pressure.