Dream Labs, which offers an API for its Dream-1 robot foundation model trained on human videos, raised a $90M seed led by B Capital and Kindred Ventures

You can now deploy robots with general intelligence in minutes, not months, using an API.

Dream Labs, a startup founded by former NVIDIA robotics researchers in 2026, has secured $90 million in seed funding. The investment round was led by B Capital and Kindred Ventures, with participation from several other venture capital firms and individual investors. The company is developing general-purpose AI foundation models for robots, aiming to provide transferable intelligence rather than task-specific programming. Their flagship model, Dream-1, is trained on human activity videos to predict and execute robot actions. Dream Labs also launched Dream API, enabling developers to connect compatible robots and issue natural-language commands, and Fine-tuning as a Service for customized models. The funds will primarily support compute infrastructure expansion to handle more tasks, hardware, and users, addressing the growing gap between affordable robot hardware and the necessary software intelligence. Dream-1 currently supports basic tabletop tasks and plans to expand to mobile robots and more complex manipulators.

The launch of Dream API and Fine-tuning as a Service signals a significant shift towards democratizing robot control and customization. By abstracting the complexities of AI model training and deployment behind an accessible API, Dream Labs is lowering the barrier to entry for developers and manufacturers. This approach could accelerate the adoption of robotics in a wider range of industries by allowing users to leverage pre-trained foundation models and fine-tune them for specific applications without requiring deep AI expertise.

This development addresses a critical bottleneck in the robotics industry: the gap between increasingly capable and affordable hardware and the software intelligence needed to make robots broadly useful. Dream Labs' focus on transferable intelligence through foundation models trained on diverse human activities aims to create a more plug-and-play robotic ecosystem. As more hardware partners integrate with the Dream API, the potential for rapid deployment of intelligent robots for everyday tasks becomes a tangible reality, potentially reshaping how businesses and consumers interact with automation.