Encord Closes $60 Million Series C for Data Infrastructure Powering Robotics AI

Encord, a data infrastructure startup building annotation and curation tools for robotics and autonomous systems, has closed a $60 million Series C round led by Wellington Management. The funding will be used to expand Encord's platform for 3D point cloud, video, and multi-sensor fusion data — the types of training data that power self-driving cars, warehouse robots, and surgical systems.

The Data Challenge in Robotics

Training AI models for physical-world applications requires fundamentally different data infrastructure than training language models. Robotics data includes 3D point clouds from LiDAR sensors, stereo camera feeds, IMU (inertial measurement unit) data, and force-torque sensor readings — all of which must be precisely synchronized, annotated, and curated. Encord's platform handles this multi-modal data pipeline, providing tools for labeling 3D objects in point clouds, tracking objects across video frames, and managing the quality and consistency of annotations at scale.

Market Drivers

The funding comes as the robotics industry enters a data-intensive phase. Companies developing autonomous vehicles, warehouse robots, and industrial automation systems are discovering that model performance is increasingly limited by data quality rather than model architecture. Encord's customers include several autonomous vehicle companies and robotics startups that are spending millions annually on data annotation — a cost that Encord claims its platform can reduce by 50-70% through AI-assisted annotation and automated quality control.

Competitive Position

Encord competes with Scale AI, Labelbox, and V7 in the data annotation market, but differentiates through its focus on 3D and multi-sensor data types that are underserved by platforms designed primarily for 2D image and text annotation. The Series C funding, which brings Encord's total raise to approximately $95 million, positions the company to expand its engineering team and invest in automated annotation capabilities that could further reduce the cost and time required to prepare training data for robotics applications.