Holiday Robotics raises $105M for its FRIDAY wheeled humanoid

Holiday Robotics said its robot FRIDAY reduces the burden of repetitive physical work. | Source: Holiday Robotics

Holiday Robotics this month raised $105 million in Series A funding. The Seoul, South Korea-based company is building a wheeled humanoid for industrial work.

Humanoid robots have made progress in mobility, interaction, and demonstrations, observed Holiday Robotics. However, for humanoids to work in manufacturing , they need the reliability to safely work alongside people and take on repetitive physical tasks, said the company .

A real factory imposes constraints that never show up in a controlled demo, it said. These include things like contact, timing, repeatability, safety, recovery, and operational fit. The robot has to work inside the environment as it already exists, not ask the environment to change around it.

Holiday’s FRIDAY semi-humanoid features a wheel-based mobile base, a upper body that for dexterous manipulation , and hot-swappable batteries to keep it going across a full shift.

The company designed safety from the hardware up. FRIDAY has lightweight hands with low inertia, fully back-drivable hand and arm joints, and 360° perception with multiple cameras and lidar sensors to eliminate blind spots. Holiday said it sees safe operation around people as a requirement, not a feature.

For a humanoid on a factory floor, walking is not enough. Value is created through their hands , according to Holiday Robotics. Hands grasp parts, place components, adjust positions, sense contact, recover from small variations, and repeat precise movements thousands of times.

Hands handle the physical details that make industrial work difficult to automate: friction, weight, alignment, compliance, surface changes, and uncertainty, said Holiday. The company claimed that it built FRIDAY around high-dexterity robotic hands.

The hands include tactile sensing and force-aware control that lets the robot feel contact and adapt to it. “FRIDAY has 64 degrees of freedom, 40 of which are dedicated to its hands — a direct measure of where we put our engineering focus,” Holiday asserted.

Holiday Robotics said it hopes to build on the progress of vision-language models VLMs) in helping robots interpret what needs to be done. Interpreting a task and reliably performing it are different problems, the company acknowledged.

When perception maps directly to action from end to end, rich behaviors can emerge straight from data, it said. However, it can be difficult to isolate why a specific motion failed, fix it without retraining the whole task, or apply that behavior to a new task, said Holiday.