Encord Raises $60M Series C: Fueling Physical AI Data Wave

Hey folks, big news from the AI world that’s got me excited. Encord, a startup building smart tools for handling AI data, just scored a massive $60 million in Series C funding. This cash injection brings their total raise to $110 million, and it’s all about powering the next wave of physical AI—like robots roaming warehouses or drones zipping through skies.​

The round was led by Wellington Management, with help from longtime backers Y Combinator, CRV, N47, and Crane Venture Partners. New players Bright Pixel and Isomer Capital jumped in too. It’s a vote of confidence as AI shifts from chatbots to real-world action. Imagine self-driving cars dodging potholes or humanoid bots grabbing your coffee—these need top-notch data to work right, and Encord makes that happen.​

At the helm are co-founders Eric Landau and Ulrik Stig Hansen. Eric, the CEO, brings brains from Harvard in applied physics and years crunching numbers at high-frequency trading firm DRW. Ulrik, president, has a computer science master’s from Imperial College London and started at J.P. Morgan. Together since 2021, they’ve grown Encord from a Y Combinator idea into a go-to for 300+ AI teams. Their platform jumped from 1 petabyte to over 5 petabytes of data managed in the last year, with physical AI revenue exploding 10x.

So, what’s the big deal with this funding? AI is exploding into the physical realm. Text-based large language models like GPT got huge on web data, but physical stuff—think autonomous vehicles from Woven by Toyota, drones from Skydio, or AI videos from Synthesia—demands multimodal data. That’s video, images, sensors all mixed, captured in real time, and cleaned up fast. Old-school cloud tools? They choke on this. Encord steps in with an AI-native data layer that handles curation, management, annotation, and alignment end-to-end.​

Let’s break it down simply. Curation means picking the gold nuggets from data mountains—spotting what’s useful for training models. Management indexes everything so you can search, filter, and track like a pro. Annotation slaps labels on images or videos, often with human-AI team-ups for accuracy. And alignment tweaks models based on real feedback, kicking out bad data. This isn’t just lab stuff; it’s for production where one bad data point can crash a robot.