Encord raises $60M Series C for multimodal robot-training data

Date: 2026-02-26
Type: ARTICLE
Desk: Robotics
Topics: Funding, Data Annotation, Multimodal Learning, Robot Training Data
Deck: Wellington Management leads the round as Encord reports growth in physical-AI customers and managed data.
URL: https://korthosrobotics.com/research/encord-60m-series-c-physical-ai-data-february-2026
Referenced profiles: Encord (https://korthosrobotics.com/ecosystem/company/encord)

Image: Encord
Encord has raised a $60 million Series C led by Wellington Management, bringing total funding to $110 million. Announced on 26 February 2026, the round includes existing investors Y Combinator, CRV, N47, Crane Venture Partners and Harpoon Ventures, alongside Bright Pixel Capital and Isomer Capital. The financing supports product development and expansion of its data infrastructure for robots, autonomous vehicles, drones and other AI systems.
The platform manages, curates, annotates and aligns data across images, video, audio, sensor streams and 3D point clouds. These inputs require teams to select useful examples, attach consistent labels and maintain the context needed to train and evaluate models. Encord’s software covers the data workflow from pre-training preparation through human feedback and subsequent model improvement.
Encord reports working with more than 300 AI teams, naming Woven by Toyota, Skydio and AXA among its customers. In its corrected financing release, the company says data on the platform increased from 1 petabyte to more than 5 petabytes over the preceding twelve months, while revenue from physical-AI customers grew tenfold.
Customer Vantor describes using Encord for geospatial workflows spanning curation, annotation and evaluation in critical infrastructure and national security. That workload adds another form of spatial data to Encord’s robotics and autonomy applications. The new capital will expand the software used to organise those datasets and support teams repeatedly incorporating new operational examples into their models.