Laser-Weeding Robots That Learn on Their Own: U.S. Company Creates Superbrain Trained with 150 Million Plants to Revolutionize Agriculture in the Field

Written by Fabio Lucas Carvalho

Published on 03/02/2026 at 12:11

Updated on 03/02/2026 at 12:12

Modelo de IA treinado com 150 milhões de plantas passa a operar robôs agrícolas com capina a laser e adaptação contínua em campo.

3 people reacted to this.

Trained With A Database Of 150 Million Plants, The Large-Scale Plant Model Presented By An American Startup Begins To Operate Agricultural Robots Capable Of Identifying Crops And Weeds, Learning Continuously In The Field And Executing Laser Weeding In Different Productive Environments

The American startup Carbon Robotics presented the first Large-Scale Plant Model, trained with 150 million plants, to operate LaserWeeder robots, identify weeds, and perform laser weeding across different crops, with continuous adaptation in the field.

Large-Scale Plant Model Trained With 150 Million Plants

Carbon Robotics claims to have developed the world’s first Large-Scale Plant Model, based on artificial intelligence and trained with a dataset of 150 million labeled plants. The system was created to transform the way agricultural crops are managed on a large scale.

The model feeds the LaserWeeder robots, enabling them to identify and remove weeds using lasers in virtually any crop or field. According to the company, recognition occurs within minutes, regardless of the type of crop or environment.

ARTICLE CONTINUES BELOW

System Learns From Global Fleet And Shares Improvements

The Large-Scale Plant Model was designed to continuously improve from real-world data.

Each deployment of the LaserWeeder sends new images of plants to the central system, expanding detection and classification capabilities across multiple scenarios.

This feedback loop creates a cumulative data effect, in which performance improvements are shared across the fleet. Thus, advances do not remain restricted to individual machines, as described by the company to Quantum News.

Carbon Robotics emphasizes that the model does not function as a static system. It acts as the core of Carbon AI, a framework that supports both the LaserWeeder and the Autonomous Tractor Kit, integrating real-time operational decisions in the field.

Cost Reduction And Less Dependence On Herbicides

According to the Seattle-based company, using the Large-Scale Plant Model can help farmers reduce labor costs, decrease the application of chemical herbicides, and maintain or increase crop productivity.

The combination of precise identification and laser removal aims to reduce the need for manual weeding and chemical inputs. The company claims that automation allows for operational gains without compromising the quality of agricultural management across different crops.

“When our robots can immediately recognize any plant in any field and adapt their behavior in real time, farmers get the most value from the machines,” said Paul Mikesell, founder and CEO of the company, in a statement.

Plant Profiles Allow Quick Personalization In The Field

In addition to the Large-Scale Plant Model, Carbon Robotics introduced Plant Profiles, a feature that allows for the rapid customization of the LaserWeeder’s operation according to specific crops, weeds, and field conditions.

Available across the entire LaserWeeder line, the tool uses a tablet interface. By selecting two or three representative images via an iPad app, operators can immediately adjust the system’s behavior.

The company states that this real-time personalization reduces the need for prolonged setups or complex reprogramming. Compared to other AI-based agricultural systems, the feature was designed for speed and operational simplicity.

Adjustments That Once Took Months Now Take Minutes

According to Carbon Robotics, adaptations that would typically take weeks or months can now be completed in just a few minutes. This reduces machine downtime and simplifies field operation while maintaining precision in different environments.

By facilitating personalization, the system aims to allow farmers to derive value from autonomous weeding technology more quickly.

The Large-Scale Plant Model remains the foundation for these rapid and continuous adaptations.

“We use plant profiles in seed beds, transplants, and directly sown onions. This has revolutionized our work,” stated a manager from Bland Farms, highlighting the platform’s simplicity and real-time performance in the field.

Laser-Weeding Robots That Learn on Their Own: U.S. Company Creates Superbrain Trained with 150 Million Plants to Revolutionize Agriculture in the Field

Trained With A Database Of 150 Million Plants, The Large-Scale Plant Model Presented By An American Startup Begins To Operate Agricultural Robots Capable Of Identifying Crops And Weeds, Learning Continuously In The Field And Executing Laser Weeding In Different Productive Environments

Large-Scale Plant Model Trained With 150 Million Plants

System Learns From Global Fleet And Shares Improvements

Cost Reduction And Less Dependence On Herbicides

Plant Profiles Allow Quick Personalization In The Field

Adjustments That Once Took Months Now Take Minutes

Forget petroleum-based asphalt: a Scottish company transforms 684,000 plastic bottles into one kilometer of road that is 60% more resistant and up to 10 times more durable, and the technology has already started replacing traditional asphalt in more than 30 countries.

With billions in loans, China is rebuilding Nigeria’s railway network from scratch, abandoning the British colonial tracks that have turned to scrap and creating a corridor of over 1,300 km to take the load off the roads of Africa’s largest economy.

Austrian engineers have developed an inflatable concrete technology that transforms flat slabs into domes using only air pressure, eliminating scaffolding and temporary forms, and potentially changing the way bridges, tunnels, and curved roofs are constructed worldwide.

Australia prepares to dredge 96.5 million m³ in Moreton Bay to deepen, widen, and realign channels of the Port of Brisbane, in a project of up to 25 years designed to accommodate larger ships and enhance maritime safety and regional port efficiency.

Study solves great ancient mystery: remains of Philip II, father of Alexander the Great, are identified in Vergina and resolves the enigma of the Macedonian royal tombs that had lasted half a century.

Larger than the Toyota Hilux and with a towing capacity of 3.5 tons, the new MG U9 2026 features a 2.5 turbodiesel engine, 520 Nm, 4×4 traction with three locks, and an almost full-size build.

For the first time in history, a submarine cable will descend to four thousand meters deep under the ice of the North Pole to ensure that the internet between Europe and Asia no longer depends on conflict zones in the Middle East.

While the world discusses geopolitics, China is quietly closing a deal with its Kazakh neighbor to secure grain supplies during any global food crisis, with more than 3,600 Kazakh companies already able to export directly to the Chinese market.

A British company has installed in the middle of the ocean the world’s first floating platform that generates electricity 24 hours a day from the temperature difference between the surface and the depths of the Atlantic, without relying on wind or sun.

Amid the United States’ economic blockade, China delivers 15,000 tons of rice to Cuba as part of a 60,000-ton donation personally approved by Xi Jinping. The package includes 80 million dollars and solar panels for hospitals.

The James Webb telescope spotted a planet 700 light-years from Earth with mornings full of sand clouds and nights with clear skies, the temperature difference between the two hemispheres reaches an impressive 170 degrees.

Laser-Weeding Robots That Learn on Their Own: U.S. Company Creates Superbrain Trained with 150 Million Plants to Revolutionize Agriculture in the Field

Trained With A Database Of 150 Million Plants, The Large-Scale Plant Model Presented By An American Startup Begins To Operate Agricultural Robots Capable Of Identifying Crops And Weeds, Learning Continuously In The Field And Executing Laser Weeding In Different Productive Environments

Large-Scale Plant Model Trained With 150 Million Plants

System Learns From Global Fleet And Shares Improvements

Cost Reduction And Less Dependence On Herbicides

Plant Profiles Allow Quick Personalization In The Field

Adjustments That Once Took Months Now Take Minutes

Forget petroleum-based asphalt: a Scottish company transforms 684,000 plastic bottles into one kilometer of road that is 60% more resistant and up to 10 times more durable, and the technology has already started replacing traditional asphalt in more than 30 countries.

With billions in loans, China is rebuilding Nigeria’s railway network from scratch, abandoning the British colonial tracks that have turned to scrap and creating a corridor of over 1,300 km to take the load off the roads of Africa’s largest economy.

Austrian engineers have developed an inflatable concrete technology that transforms flat slabs into domes using only air pressure, eliminating scaffolding and temporary forms, and potentially changing the way bridges, tunnels, and curved roofs are constructed worldwide.

Australia prepares to dredge 96.5 million m³ in Moreton Bay to deepen, widen, and realign channels of the Port of Brisbane, in a project of up to 25 years designed to accommodate larger ships and enhance maritime safety and regional port efficiency.

Study solves great ancient mystery: remains of Philip II, father of Alexander the Great, are identified in Vergina and resolves the enigma of the Macedonian royal tombs that had lasted half a century.

Larger than the Toyota Hilux and with a towing capacity of 3.5 tons, the new MG U9 2026 features a 2.5 turbodiesel engine, 520 Nm, 4×4 traction with three locks, and an almost full-size build.

While the world discusses geopolitics, China is quietly closing a deal with its Kazakh neighbor to secure grain supplies during any global food crisis, with more than 3,600 Kazakh companies already able to export directly to the Chinese market.

A farm in the Mexican desert has become so green that today it has rivers flowing year-round and even rain in the dry season, all thanks to a technique that transforms dead sand into fertile soil with manure, mountain microorganisms, and what farmers call the secret soil formula.

Stable fly infestation frightens producers in Costa Rica, disrupts cattle routine, reduces milk, and puts vinasse at the center of rural controversy.

China suspended beef purchases from three Brazilian meatpacking plants of JBS, Prima Foods, and Frialto after finding a banned synthetic hormone in tests conducted on the meat, two days after reauthorizing three other plants that had been embargoed since last year.

For the first time in history, a submarine cable will descend to four thousand meters deep under the ice of the North Pole to ensure that the internet between Europe and Asia no longer depends on conflict zones in the Middle East.

While the world discusses geopolitics, China is quietly closing a deal with its Kazakh neighbor to secure grain supplies during any global food crisis, with more than 3,600 Kazakh companies already able to export directly to the Chinese market.

A British company has installed in the middle of the ocean the world’s first floating platform that generates electricity 24 hours a day from the temperature difference between the surface and the depths of the Atlantic, without relying on wind or sun.

Amid the United States’ economic blockade, China delivers 15,000 tons of rice to Cuba as part of a 60,000-ton donation personally approved by Xi Jinping. The package includes 80 million dollars and solar panels for hospitals.

The James Webb telescope spotted a planet 700 light-years from Earth with mornings full of sand clouds and nights with clear skies, the temperature difference between the two hemispheres reaches an impressive 170 degrees.