CarbonSix raises $40 million to bring intelligent learning machines to the factory floor

CarbonSix Inc., a developer of ready-to-deploy robotic artificial intelligence solutions, announced Wednesday that it has raised $40 million in seed funding to transform the way manufacturing works.

The company focuses on providing robotic intelligence hardware and software that manufacturing lines can quickly deploy into production.

Successfully operating AI robotics on the manufacturing floor requires large amounts of data and learning power. To remain competitive in this industry, CarbonSix uses what it calls a “data flywheel” rather than generalized datasets for its robot AI underlying models.

When robotics and intelligence work together in the field, high-quality, task-specific data is captured into the system. The company says this creates a powerful self-reinforcing loop that drives data generation and refinement, providing smarter automation tools on the factory floor.

“From the beginning, our goal was not to build technology for technology’s sake; it was to create practical, field-ready physical AI that delivers tangible revenue for manufacturers,” said CEO Tae-yeon Terry Moon.

Physical AI, or integrating AI into hardware that allows machines to perceive, reason with, and physically interact with the real world, is changing the way intelligent software is deployed in the real world. Common examples include self-driving cars, humanoid and industrial robots, drones, and medical equipment.

This technological shift is rapidly permeating factory floors and warehouses.

In line with that transformation, CarbonSix has brought together deep expertise to build a platform. These include Moon, formerly of SuaLab, an industrial AI vision company acquired by US machine vision system maker Cognex. The Massachusetts Institute of Technology graduate leads the development of the company’s robotic intelligence framework, and former Yale postdoc and chief hardware officer Je-hyeok Kim provides expertise in cutting-edge hand and manipulator hardware design.

Together, they developed SigmaKit, a collection of tools that includes AI software, robotic hands, educational devices, and sensor modules. It can be easily implemented in factories and production lines without deep expertise in robotics or AI. It is designed to be simple and does not require complex coding. By allowing the robot to incorporate operator know-how with minimal effort, users can quickly and intuitively teach the robot tasks.

The user repeatedly demonstrates the target task to the robot using remotely controlled motions captured as images and motion data. This builds the training dataset. Depending on your task, you can generate and deploy a model within a day. SigmaKit makes training efficient and allows users to quickly filter, review, and edit collected data.

Once training is complete, skills can be directly implemented into unstructured tasks. The company said it requires minimal skill coordination as it scales across production lines or additional sites, enabling mass adoption.

The Series A funding round was co-led by DSC Investment and LB Investment. New investors participating in this round included IMM Investment, Korea Industrial Bank, SV Investment, Cortentia, and A Squared US, and existing seed round investors including Foothill Ventures, Storm Ventures, Zeitgeist Capital, Xquared, and CarbonBlack Fund all participated fully in follow-on investments.

The company said it plans to use the new funding to hire new talent, expand infrastructure and expand globally as more factory floors are transformed with smart machines.

Image: Carbon Six