Dream Labs Raises $90 Million for Robotics AI

The San Francisco-based startup plans to scale computing power and expand support for new robotic hardware.

Updated on Oct. 19, 2026 in Robotics

Isometric editorial illustration of a robotic gripper and optical sensors, representing advancements in robotics AI integration. — Dream Labs secured $90 million in funding to scale its AI platform, which enables robotics hardware to translate visual inputs into physical actions. AI Illustration. Upload story photo >

Live Poll

Do you trust artificial intelligence to safely perform manual tasks in your home or workplace?

Dream Labs has secured $90 million in seed funding to advance its flagship Dream-1 AI model. The startup also launched an API and fine-tuning service to help robots convert camera feeds and instructions into physical motor commands.

Why it matters

The company aims to solve the challenge of robot generalization, creating systems that apply learned skills across diverse tasks and hardware platforms. This funding supports the expansion of the compute infrastructure needed to train models on human demonstration data.

Dream Labs utilizes a library of video-based human task demonstrations to train the Dream-1 model, which currently integrates with over 20 hardware partners. The API interprets English inputs alongside visual data to execute precise motor actions.

The players

Dream Labs

A San Francisco-based robotics artificial intelligence company focused on enabling robots to learn skills from human visual demonstrations.

B Capital

A global multi-stage investment firm that led the seed funding round for Dream Labs.

Kindred Ventures

A venture capital firm that participated as a lead investor in the Dream Labs seed round.

The details

Dream-1 learns to predict robot actions by observing people performing everyday tasks, allowing for greater adaptability in real-world environments. Developers can use the company's new Fine-tuning as a Service to create custom models for hardware that is not natively supported.

Dream Labs was founded in 2026.
The $90 million seed round was raised in October 2026.

The Tech Race

The development of universal motor-control models like Dream-1 represents a shift away from rigid, task-specific programming toward generalized AI robotics. This positions the company in direct competition with established automation firms attempting to standardize cross-platform robotic intelligence.

As these models are deployed, developers may gain faster access to high-functioning robotic automation for custom hardware without building software from scratch. Increased adoption could lead to more versatile robotic assistants capable of performing complex chores in homes and businesses.

The takeaway

The move toward vision-based motor learning highlights a transition toward AI that can interpret human instructions in real-time. Investors are increasingly betting that standardizing this interface is the key to scaling robotics across different physical platforms.