Key facts
- Funding: On 26 February 2026 Encord raised a US$60m Series C led by Wellington Management, taking total funding to US$110m.
- Growth: Data on Encord's platform grew from 1 petabyte to more than 5 petabytes in the year before the Series C, and revenue from physical AI customers rose tenfold.
- Offices: Encord is headquartered in London with a presence in San Francisco and New York.
- Entity: Encord was founded in 2021 by Ulrik Stig Hansen and Eric Landau; its legal entities are Cord Technologies, Inc. and Cord Technologies Limited.
- Events: Encord ran its first physical AI summit, Long Horizon, at Fort Mason in San Francisco on 20 and 21 October 2026, with more than 60 speakers from companies including Waymo, Google DeepMind and Boston Dynamics.
- Research: At Long Horizon Encord introduced Encord Data Labs, a research division for physical-world data. Its first release, EMG-1, is an open dataset of 307 hours of field wrist-muscle (EMG) recordings synchronised with egocentric video, which Encord reports cut median fingertip tracking error by 25% when combined with vision.
- Partnerships: Working with NVIDIA's Cosmos synthetic-data tools, Encord generated 1,500 synthetic clips for ANYbotics' inspection robots, which it reports raised detection accuracy for liquid spills by 10%.
Focus areas
About
Encord sells data infrastructure to teams that build AI models, with a focus on physical AI such as robots, drones and autonomous vehicles. Its platform covers data indexing and curation, annotation and labelling, and post-training alignment for video, images, LiDAR, radar, audio, text and DICOM medical images, and it also offers managed annotation and data collection. More than 300 AI teams use it, including Woven by Toyota, Skydio and Synthesia. A US$60m Series C led by Wellington Management in February 2026 took total funding to US$110m.
How they use AI
Encord uses models inside its platform to make data work faster: embedding-based search and model-in-the-loop curation find edge cases and distribution gaps in large multimodal datasets, and Data Agents combine AI and human review in automated pipelines. For post-training, teams run RLHF, rubric-based evaluation and pairwise comparison to find where production models fail. Merlin, a conversational agent usable in Encord, Claude, Codex, Slack or over MCP, builds labelling projects, spots coverage gaps and recommends fixes.
Products and services
- Annotation and labelling — Video, LiDAR, audio, text and sensor-fusion annotation in one workflow, with label history and quality controls
- Data indexing and curation — Embedding-based search and model-in-the-loop curation across multimodal data
- Post-training alignment — RLHF, rubric-based evaluation and pairwise comparison for production models
- Merlin — Conversational agent that builds labelling projects and recommends data fixes
- Data-as-a-Service — Expert annotators and data collection, including Encord's own teleoperation facilities
- Encord Data Labs — Research division launched in October 2026 that develops hand-tracking models, new sensor capture such as wrist EMG, and partner pilots for physical AI data
Notable customers / partners
- Woven by Toyota
- Skydio
- Synthesia
- Toyota Research Institute
- AXA
- UiPath
- Zipline
Key people
- Eric Landau — Co-founder
- Ulrik Stig Hansen — Co-founder
Investors
- Wellington Management
- Y Combinator
- CRV
- N47
- Crane Venture Partners
- Bright Pixel
- Isomer Capital