Carbon Robotics Unveils Large Plant Model to Help LaserWeeder Robots Identify Weeds Instantly

Carbon Robotics is rolling out a new AI model designed to make its LaserWeeder robots much faster at recognizing plants in the field, giving farmers a new way to target weeds without the lengthy retraining process that has traditionally been required. The Seattle-based company announced Monday that its new Large Plant Model (LPM) can identify plant species instantly, allowing operators to flag new weeds in real time rather than waiting for engineers to build fresh training datasets.

The model now powers Carbon AI, the system inside the company’s autonomous robots that use lasers to kill weeds. Until now, Carbon Robotics often had to retrain its machines whenever a new weed appeared on a farm — or even when familiar weeds looked different because of soil conditions or growth stages. CEO Paul Mikesell told TechCrunch that process typically took about 24 hours each time.

With LPM, the company says that step largely disappears. Farmers can simply select photos collected by the robot through its interface and specify what should be removed or protected, without additional labeling or retraining. “The farmer can live in real time and say, ‘Hey, this is a new weed. I want you to kill this,’” Mikesell said.

Carbon Robotics trained the model on more than 150 million photos and data points gathered from its machines operating across more than 100 farms in 15 countries. The company believes the scale of that dataset allows LPM to generalize across plant species and conditions, even when encountering weeds it hasn’t seen before.

The update will be delivered to existing LaserWeeder systems through a software rollout, and Carbon says it will continue refining the model as more data comes in from deployments. The company, founded in 2018, has raised more than $185 million from investors including Nvidia NVentures, Bond, and Anthos Capital.

For Carbon Robotics, the announcement highlights a shift toward more adaptable agricultural automation, where farmers can respond quickly to changing field conditions without waiting on manual model updates — potentially making laser-based weed control more practical at scale.

This analysis is based on reporting from TechCrunch.

Image courtesy of Unsplash.

This article was generated with AI assistance and reviewed for accuracy and quality.