Feather Robotics offers $30,000 modular humanoid platform for developers

Feather Robotics, a year-old humanoid-robot startup, announced that its modular robot platform has already generated more than $1 million in revenue and is gearing up for a larger product launch. The company’s approach, providing both hardware and a software toolkit that developers can adapt to specific tasks, offers a practical alternative to the “ChatGPT moment” many in robotics still consider out of reach.

The company was founded by Hoa Mai, whose previous humanoid venture was acquired by 1X, and former Tesla Model 3 engineer Parsa Bakhtiari, who once reported directly to Elon Musk. Their combined experience shaped Feather’s goal of creating a “Android of robotics.” Gradient Ventures led the startup’s $7.6 million pre-seed round, backing the vision of a deployable robot that developers can build on, rather than waiting for a fully autonomous general-purpose machine.

Feather’s robot is built as a modular system that lets users adjust hardware components such as arm length to suit varied applications. The company says early customers are using the units as cooks in restaurants and as cleaners in scientific labs, though it does not disclose specific client names. On the software side, the platform can run AI models from leading robotics providers, including Nvidia, Skild and Physical Intelligence. Priced at $30,000, the robot costs roughly half of the Unitree H2 Edu model that dominates the Chinese market.

Gradient partner Darian Shirazi argues that Feather occupies a unique niche in the United States. He categorizes the current hardware robotics landscape into three groups: domestic-focused startups like Sunday, heavyweight general-purpose efforts from companies such as Figure and Tesla, and Feather’s modular humanoid platform, which he believes has no direct U.S. competitor. Shirazi points to the robot’s price point as a potential catalyst for adoption, noting that hiring a human laborer can cost $50,000 to $60,000 per year plus training, HR and other overhead. A Feather robot, by contrast, offers a one-time capital expense that could replace many of those recurring costs.

Despite the early traction, Feather has spent only a fraction of its pre-seed capital, according to Shirazi, indicating a lean operating model. Mai emphasizes that the long-term value will come from an ecosystem of applications built around the hardware. “The market for physical AI companies is small today, but we expect thousands of application companies to emerge in the next five years,” he said. Feather hopes to become the foundational platform that powers those future solutions.

Looking ahead, the startup plans to expand its customer base beyond the initial pilot deployments and to support a broader range of AI models as the robotics software landscape evolves. If Feather can sustain its revenue growth while attracting developers to its ecosystem, it could set a new standard for modular, developer-friendly robotics in the United States.