Dream Labs

Robotics · AI

Funding

Dream Labs raises $90M seed to build a universal brain for robots

Industrial robots number in the millions but each one is locked to a single task, a bottleneck Dream Labs wants to break with video-trained AI.

• Dream Labs raised $90 million in seed funding led by B Capital and Kindred Ventures, alongside the launch of Dream API, a public interface that turns a robot's camera feed and plain-English instructions into motor commands.
• The San Francisco startup was founded in 2026 by former Nvidia robotics researchers who worked on GR00T, DreamZero and World Action Models; CEO is Joel Jang.
• Its Dream-1 model is trained on video of people doing everyday tasks rather than on any single machine, letting it control robots with grippers from different manufacturers without extra training.
• Factories worldwide run a record 5 million industrial robots, but nearly all are programmed for one task with no ability to transfer skills to another machine.
• Dream-1 already handles tabletop tasks like picking, placing and folding laundry, and works with more than 20 hardware partners; most of the new capital will go toward compute to expand supported tasks and robots.
• The company plans to extend support beyond grippers to wheeled mobile robots and more dexterous manipulation as it broadens Dream-1's capabilities.
• A general-purpose robot 'brain' that transfers across hardware could lower the cost of automating new tasks from months of specialist work to minutes, reshaping how cheap robots get deployed.