Meet Feather, the startup building the ‘Android of robotics’ for developers
Disrupt 2026: OpenAI, Anthropic, Replit, and more take over 6 industry stages. 25% off tickets now Disrupt ticket savings of up to $200 + 50% off a second ends Sept 25, 11:59 p.m.
PT. REGISTER HERE. If there is one thing robotics founders and investors agree on, it’s that the “ChatGPT moment” where a general-purpose robot can seamlessly adapt to any environment has not arrived. Opinions vary on whether the breakthrough is just around the corner or a decade away. Feather Robotics, a humanoid startup founded last year, is developing a robot for either timeline. Unlike Tesla or Figure, which are tackling the lofty mission of building both the body and a foundational ‘brain’ that isn’t yet widely deployed, Feather provides the hardware and software toolkit for developers to tackle real-world tasks now. “You can’t buy a Tesla robot today and develop on top of it,” Feather co-founder Hoa Mai told TechCrunch. “We realized that this is not how most companies became successful. Hardware players like Nvidia or Apple started with a product that worked and was deployable, and then they added complexity over time.” In 2025, after selling his previous humanoid startup to 1X, Mai (pictured left) teamed up with Parsa Bakhtiari (pictured right), a former Tesla Model 3 engineer who at one point reported directly to Elon Musk, to build what the duo hopes will become the Android of robotics. That vision was attractive to Gradient Ventures, which backed Feather at its inception, leading the startup’s previously announced $7.6 million pre-seed round. Since then, Feather has rolled out a modular robotic system that allows developers to customize the hardware for different use cases, such as adjusting arm lengths. The startup has already begun selling to customers, surpassing $1 million in revenue. Although Feather is not disclosing its customers, it said that its robots are working as cooks in restaurants in Japan and cleaning up science labs. On the software side, the hardware can run models from any leading robotics AI provider, including Generalist, Skild, or Physical Intelligence. “We’ve been selling small quantities of these robots, and now that we resolved like almost all the issues over the last year of field testing, we’re getting ready for a big product launch,” Mai said.
