Encord Raises $60M for Physical AI Data Infrastructure
Encord, a data infrastructure startup for physical AI, has raised $60 million to scale its platform for managing and annotating complex sensor data. The company focuses on transforming raw data from robots and drones into actionable training data, aiming to bridge the gap between simulation and real-world performance. The funding highlights investor focus on the data operations and infrastructure layers required to scale embodied AI systems.
The latest funding round brings Encord's total capital raised to $110 million, with the Series C led by Wellington Management. This investment underscores a growing venture capital focus on the AI infrastructure layer, particularly for what's termed "physical AI" – systems that interact with and perceive the real world. The company was co-founded by Ulrik Stig Hansen and Eric Landau, who identified that AI teams were spending over 80% of their time on data preparation rather than model development. Encord's platform is engineered to handle complex, multimodal data streams essential for robotics and autonomous systems, including LiDAR, radar, video, and other sensor data. This focus on high-fidelity, real-world sensor data is a key differentiator from more generalized data labeling platforms. The company's architecture is designed for the entire data lifecycle, from curation and annotation to model evaluation, creating a "data flywheel" where models are used to improve the data that trains them. Key customers in the autonomous sector include Woven by Toyota, and drone manufacturers Zipline and Skydio, highlighting the platform's relevance for training perception and navigation systems in real-world environments. Encord's technology is also being applied in warehouse automation with customers like Pickle Robot, demonstrating its utility in industrial robotics. The company reports a tenfold increase in revenue from its physical AI customers and a growth in data managed on its platform from one to over five petabytes in the last year. The focus on robust data infrastructure for sensor fusion is highly relevant to defense modernization. The U.S. Department of Defense's AI strategy emphasizes the need for high-quality data to achieve "decision superiority" and improve capabilities like battlespace awareness and resilient kill chains. Challenges in military AI, such as processing data from a multitude of sensors in real-time at the tactical edge, directly mirror the problems Encord aims to solve. For defense applications like counter-UAS (counter-unmanned aerial systems), the ability to accurately and rapidly annotate multi-sensor data (e.g., fusing radar, electro-optical, and infrared inputs) is critical for training reliable detection and tracking algorithms. Modern counter-drone systems increasingly rely on AI to identify and mitigate threats, and the performance of these AI models is fundamentally dependent on the quality of the training data. The competitive landscape for AI data infrastructure includes major players like Scale AI. However, Encord is positioning itself as a specialist for physical AI, emphasizing its platform's capabilities for handling the complex, continuous streams of sensor data that are characteristic of robotics, autonomous vehicles, and drones, as opposed to the more generalist approach of competitors. Investor interest in "embodied AI" and the robotics sector is on the rise, with a recognition that real-world applications are the next frontier after large language models. The market for data annotation tools is projected to grow significantly, with some forecasts predicting it will reach over $12 billion by 2031, driven by the demand for high-quality training data for autonomous systems. This funding round is poised to help Encord expand its product development and move into new markets, capitalizing on the shift of AI from experimental lab settings to production deployments in the physical world. The growth of warehouse automation, autonomous vehicles, and the increasing use of drones in various sectors suggest a large and growing addressable market for specialized data infrastructure.
