Robotics and Automation

Feather Develops a Modular Robotic Platform for Developers Priced at $30,000

The American startup Feather Robotics is betting on a customizable humanoid robotic platform that allows developers to modify the hardware and run models from different artificial intelligence providers. The company has begun selling limited quantities of its robots and generated more than $1 million in revenue, as it prepares to launch a broader product.

2026-09-24
4 min read
20 views
certi.news Editorial Team
Feather Develops a Modular Robotic Platform for Developers Priced at $30,000

The startup Feather Robotics is building a modular humanoid robotic platform aimed at developers and companies, rather than offering a function-locked robot that attempts to handle every environment and task from the outset. The company sells the robot for $30,000, with the option to customize its components, such as adjusting arm lengths according to the intended use.

Feather was founded last year by Hoa Mai and Parsa Bakhtiari. Mai had previously founded a humanoid robotics startup that was sold to 1X, while Bakhtiari worked as an engineer on the Tesla Model 3 and, at one stage, reported directly to Elon Musk. At its founding, Feather raised a $7.6 million seed round backed by Gradient Ventures.

An Open Platform Instead of a Closed Robot

The company believes that the robotics market has not yet reached the stage at which a single general-purpose robot can adapt smoothly to any environment. Feather therefore focuses on providing hardware and software that developers can build on today, rather than waiting for the development of a robotic body and a general-purpose brain capable of performing every task.

The platform can run models from artificial intelligence providers specializing in robotics, including Nvidia, Skild, and Physical Intelligence. Feather says its robots are already performing tasks such as cooking inside restaurants and cleaning scientific laboratories, but it has not disclosed the names of its customers.

What Changes in Practice?

Feather offers a model closer to a development platform: the hardware can be modified to suit the task, and the software model appropriate for it can then be selected. This approach could lower the barrier to experimentation for companies that do not want to develop a complete robot from scratch, but it does not eliminate the challenges of field operation or the need to adapt the system to each environment.

The company has begun selling robots in small quantities and said it surpassed $1 million in revenue after a year of field testing. It also reported that it had addressed most of the problems that emerged during those tests and is now preparing to launch a major product.

Price and Competition

A Feather robot costs approximately $30,000, or nearly half the price of the Unitree H2 Edu. According to Darian Shirazi, a general partner at Gradient, the company has no direct American competitor in the modular humanoid platform category; some competitors focus on home robots, while companies such as Figure and Tesla are developing integrated general-purpose robots.

Feather acknowledges that it drew inspiration for some of its approaches from Chinese robotics companies such as Unitree. At the same time, the company believes that new restrictions on the entry of foreign-manufactured models into the American market give it an opportunity as a local supplier.

A Bet on an Application Ecosystem

Feather's bet does not depend solely on selling the robot, but on the emergence of an ecosystem of physical artificial intelligence application companies that use the platform. The company estimates that the number of these companies could reach thousands within five years, compared with a small market size today.

However, the source does not provide details about the number of units sold, the robot's performance across different tasks, or the terms of support and maintenance, and Feather has not announced the names of its customers. Therefore, the success of its model remains tied to its ability to turn limited tests into reliable, scalable operations, rather than merely providing customizable hardware.

News source
TechCrunch Robotics
Open original source ↗
c
Author

certi.news Editorial Team

In the same category

You may also like

View all news