# Feather Robotics sells $30,000 robots and lets developers bring their own AI

> Source: <https://runtimewire.com/article/feather-robotics-developer-robot-platform>
> Published: 2026-09-24 19:41:27+00:00

# Feather Robotics sells $30,000 robots and lets developers bring their own AI

**Co-founder Hoa Duc Mai says the platform gives robotics developers a working machine to build on; TechCrunch reports more than $1 million in revenue.**

        By [RuntimeWire Staff](https://runtimewire.com/author/runtimewire-staff)
        · Published 

Primary source: [TechCrunch](https://techcrunch.com/2026/09/24/meet-feather-the-startup-building-the-android-of-robotics-for-developers/)

## Why it matters

Feather is betting that physical-AI companies will buy a shared robot platform rather than build their own hardware. Its early revenue suggests initial demand, but customer names, deployment counts and independent reliability evidence remain undisclosed.

[Feather Robotics](https://feather.dev/?ref=runtimewire) is selling a roughly $30,000 robot to developers who want to put AI into physical work without first building the machine and software stack themselves. Co-founder and CEO [Hoa Duc Mai](https://feather.dev/mission?ref=runtimewire) says Feather is betting that a ready-to-deploy platform can become the base for a broader market of robotics applications.

Mai arrived at Feather after working on humanoid robotics at Kind Humanoid, which [1X acquired in January 2025](https://www.1x.tech/discover/1x-acquires-kind-humanoid?ref=runtimewire). Feather's account says Mai helped build Kind's system architecture and AI control stack. His co-founder, Chief Product Officer Parsa Bakhtiari, brings a manufacturing background: Feather says he worked at Tesla during the Model 3 production ramp and led robotics projects.

The founders are taking a different route from companies such as Tesla and Figure, which are developing both robot hardware and the intelligence intended to control it. Feather sells the hardware and developer tools, then leaves customers room to build or choose the software. Mai put the distinction plainly to [TechCrunch](https://techcrunch.com/2026/09/24/meet-feather-the-startup-building-the-android-of-robotics-for-developers/?ref=runtimewire): "You can't buy a Tesla robot today and develop on top of it."

### A robot platform, not a finished worker

Feather Core is a wheeled, two-armed robot with a height-adjustable torso, modular end effectors and a holonomic base that moves in any direction. Customers can configure parts of the machine, including arm length, batteries and onboard computing. Feather's [technology page](https://feather.dev/technology?ref=runtimewire) lists built-in teleoperation, inverse kinematics, collision detection and simulation support. Its open SDK supports Python and ROS2, according to Feather.

Those are product specifications and company descriptions, not independent performance tests. Feather lists 600 millimeters of vertical travel, a one-meter reach and up to 10 hours of battery life with dual hot-swappable batteries. Its [homepage](https://feather.dev/?ref=runtimewire) lists a $29,990 promotional price and shows $32,990 as the previous price. Feather labels the $29,990 price a promotion while supplies last, so it is best read as an advertised starting price rather than a universal price for every configuration.

Feather targets physical-AI companies, manufacturers, researchers and integrators. The intended division of labor is simple: Feather supplies the robot body and basic operating tools; customers spend their engineering effort on the model or application. [TechCrunch](https://techcrunch.com/2026/09/24/meet-feather-the-startup-building-the-android-of-robotics-for-developers/?ref=runtimewire) reported that the robot can run models from providers including Generalist, Skild AI and Physical Intelligence.

That arrangement puts Feather's business on the hardware side of a market that may eventually produce many software and application companies. Mai told [TechCrunch](https://techcrunch.com/2026/09/24/meet-feather-the-startup-building-the-android-of-robotics-for-developers/?ref=runtimewire) he expects thousands of physical-AI application companies to exist within five years. That is a forecast, not a measured market. Feather's opportunity depends on enough developers needing a shared, adaptable robot platform - and on the platform working reliably beyond demonstrations and limited deployments.

### Early sales, limited evidence

[TechCrunch](https://techcrunch.com/2026/09/24/meet-feather-the-startup-building-the-android-of-robotics-for-developers/?ref=runtimewire) reported that Feather had surpassed $1 million in revenue and had begun selling robots in small quantities. The article did not specify the period covered by that revenue figure or distinguish recognized revenue from orders. Feather told TechCrunch that its robots were being used for cooking in restaurants in Japan and cleaning in science labs, but did not name the customers or disclose deployment counts.

Mai said Feather had spent the previous year field-testing the systems and was preparing for a larger product launch after resolving most of the issues it encountered. That account points to the practical test behind the platform pitch: customers need machines that can be maintained and adapted in real settings, not just an open SDK and a list of compatible models. Feather has not published customer-level results or independent reliability data.

Feather's financing gives the bet some runway, although the public figures do not add up to a clean, confirmed total. [TechCrunch](https://techcrunch.com/2026/09/24/meet-feather-the-startup-building-the-android-of-robotics-for-developers/?ref=runtimewire) reported a previously announced $7.6 million pre-seed round led by [Gradient Ventures](https://www.gradient.com/?ref=runtimewire). Separately, [SEED Innovations](https://www.seedinnovations.co/?ref=runtimewire) said it [invested $1 million in May 2026](https://www.lse.co.uk/rns/SEED/investment-in-feather-robotics-inc-pbqekykxg9a3uiw.html?ref=runtimewire) through a SAFE with a $60 million valuation cap. The [SEC filing index records an April 13th, 2026 filing date](https://www.sec.gov/Archives/edgar/data/2128909/000212890926000001/0002128909-26-000001-index.htm?ref=runtimewire) for Feather's Form D; the [SEC filing reports an offering amount of $6,015,385](https://www.sec.gov/Archives/edgar/data/2128909/000212890926000001/xslFormDX08/primary_doc.xml?ref=runtimewire). That filing does not establish Feather's total capital raised or clarify how it relates to the reported pre-seed and SAFE.

Gradient general partner Darian Shirazi told [TechCrunch](https://techcrunch.com/2026/09/24/meet-feather-the-startup-building-the-android-of-robotics-for-developers/?ref=runtimewire) that Feather had no direct U.S. competitor in its particular category. That is an investor's assessment. The wider field includes robot makers pursuing household use and companies building integrated robot-and-AI systems, while Feather is trying to sell a more open platform to developers. The comparison depends on what customers value: a complete system from one supplier, or hardware they can customize and pair with outside software.

Mai's thesis comes from a problem he says he encountered as a robotics developer: the available platforms did not meet his needs for unit economics, reliability and support. Feather's [mission statement](https://feather.dev/mission?ref=runtimewire) frames the response as making deployable hardware accessible to smaller teams. Mai is betting that developers can build businesses around a shared robot platform. Feather now has to show that affordability and openness can attract repeat customers without leaving those customers to solve the hardest integration and maintenance problems themselves.
