London-based Encord has secured €50 million in Series C funding to expand its AI-native data infrastructure as physical AI systems move from pilot programs into large-scale deployment.
The round was led by Wellington Management and brings the company’s total funding to €93 million. Existing investors Y Combinator, CRV, N47, Crane Venture Partners and Harpoon Ventures participated, alongside new backers Bright Pixel Capital and Isomer Capital.
Building The Data Layer For Physical AI
Founded in 2021, Encord develops a universal data layer designed to support AI teams across the full model lifecycle. Its platform enables organizations to manage, curate, annotate and align complex datasets used to train and operate AI systems.
The company currently supports more than 300 AI teams, including Woven by Toyota, Zipline, AXA and Skydio.
While much of the AI industry has focused on large language models trained on internet-scale text, physical AI systems depend on structured, proprietary data such as sensor streams, field-captured edge cases, video, audio and 3D point clouds. These multimodal datasets are significantly more complex to process and manage.
Advertisement
From Pilot To Production
The funding comes as robotics, autonomous vehicles and drone systems transition from experimentation to production environments.
Industry projections suggest that hundreds of millions of AI-powered robotic systems could be deployed globally within the next few years, with the physical AI market expected to expand rapidly.
Unlike text-based models, physical AI systems must operate reliably in dynamic real-world conditions. That makes data quality, consistency and alignment critical.
Encord positions itself as the infrastructure layer that ensures training data is structured, validated and continuously optimized as models evolve in production.
Scaling Infrastructure And Global Reach
The new capital will be used to accelerate product development, expand into additional markets and further scale the company’s AI-native infrastructure platform.
According to the company’s leadership, model performance in physical AI is directly constrained by data readiness rather than model size. As systems interact with real-world environments, continuous data improvement becomes a core operational requirement.
Encord aims to provide the foundation that enables physical AI systems to learn, adapt and improve over time.
About Encord
Encord provides a universal AI data platform supporting the indexing, curation, annotation and evaluation of datasets across the full AI lifecycle. Its infrastructure is designed to help organizations build and operate reliable AI systems in production environments, particularly in robotics, autonomous mobility and other physical AI domains.
