Feather Raises $7.6M for Open Humanoid Robotics

Feather Robotics has raised $7.6M in Pre-Seed financing to scale a modular robot platform built for companies developing physical-AI applications. Gradient Ventures led the SAFE round, with Builder Capital, Geometry, SEED Innovations, and Virgo VC participating.

The East Palo Alto company is selling a configurable robot body and open software-development kit rather than asking every application developer to build the same hardware stack again. That distinction matters because better AI models do not remove the work of sourcing motors, designing arms, integrating cameras, managing batteries, supporting field failures, and keeping a machine useful through a full shift.

What Feather Raised

Feather announced the round on September 24, 2026. SEED Innovations' disclosure describes the financing as a Simple Agreement for Future Equity, or SAFE, and says SEED's $1M participation will convert at a future qualifying round or liquidity event subject to a $60M valuation cap. That cap applies to SEED's conversion terms; it should not be read as a priced valuation for Feather.

The syndicate backs a company founded in 2025 by Hoa Mai, Feather's CEO, and Parsa Bakhtiari, its CPO. Mai has worked on embodied-AI systems and humanoid robotics, while Bakhtiari worked on Tesla Model 3 production and advanced robotics projects. Their operating thesis is straightforward: physical-AI developers should be able to buy a capable body, configure it, and spend their engineering time on the application that makes the machine valuable.

The Platform Behind the Round

Feather's current system is a 105 kg bimanual mobile robot with two 7-degree-of-freedom arms, a torso that travels 600 mm vertically, and a holonomic base that can move in any direction without turning. The base system has a limited-time starting price of $29,990 and supports configurable Nvidia Jetson compute, stereo cameras, grippers or 6-degree-of-freedom hands, built-in teleoperation, and an open Python and ROS2 SDK.

The specifications reflect an argument about deployment rather than spectacle. Dual hot-swappable batteries support a claimed 10-hour operating window, the 1-meter arms extend the working envelope, and modular compute and end effectors let integrators adapt one platform across different jobs. Feather is trying to make the robot body an input to a business, not the entire business.

That is a useful distinction in physical AI. A developer may have a model that recognizes objects, plans motion, or learns a task, yet still face months of mechanical design, procurement, safety work, integration, and service planning before the first customer can use it. Feather wants to absorb more of that repeated infrastructure work so the developer can focus on the workflow, data, and customer outcome.

Early Commercial Evidence

Feather says it built its first prototype in 2 months, fulfilled its first order in 9 months, and has surpassed $1M in revenue. The company also says its robots have worked for more than a year across manufacturing, food service, and laboratory settings, including for an unnamed manufacturer with $4B in annual revenue.

Those figures are company-reported and the customers remain undisclosed. Still, they put the round in a more useful context than a prototype video alone. TechCrunch reported that Feather's machines are performing tasks such as restaurant cooking and laboratory cleaning, while the company prepares for a larger product launch.

The financing gives Feather more room to expand production, engineering, field support, and go-to-market capacity. The company is hiring across engineering, operations, and commercial roles, although it has not published a detailed allocation for the proceeds.

Why the Developer Handoff Matters

Humanoid robotics is attracting capital across vertically integrated systems, household robots, foundation models, and industrial automation. Many of the best-funded companies want to own the intelligence, body, and application together. Feather is making a different economic bet: a broad developer ecosystem will need accessible hardware even if the winning physical-AI models come from several providers.

That approach resembles other computing markets where standardized hardware and software interfaces expanded the number of teams that could build useful products. Robotics is harder because a platform still has to survive contact, maintenance, customer sites, and support calls. An SDK can open the door, but reliability, replacement parts, service response, and predictable unit economics determine whether anyone keeps walking through it.

Feather's published support model includes U.S.-based engineers, a 1-year warranty, service coverage, fleet agreements, and stocked replacement parts. Those operational details are not as cinematic as a robot demo, but they are closer to the buying questions that physical-AI developers and integrators face when a machine leaves the lab.

What the $7.6M Must Prove

Feather now has to show that one configurable platform can support enough applications without becoming a custom-engineering shop for every deployment. The company must also turn early revenue into repeatable production, prove field reliability across larger fleets, and demonstrate that developers can create durable businesses on top of its hardware.

The opportunity is meaningful precisely because the handoff is unfinished. AI teams can choose among increasingly capable models, while the practical path from model to machine still asks them to rebuild too much of the same physical stack. Feather's $7.6M gives the company a chance to make that handoff real, one deployment, service call, and developer application at a time.