> ## Content Index > Fetch the complete content index at: https://nexi.fund/llms.txt > Use this file to discover other available public pages before exploring further. # Feather Robotics Is Selling a Humanoid Worker for $29,990 - URL: https://nexi.fund/feather-robotics-humanoid-2026/ - Published: 2026-10-01T10:30:17.000Z - Updated: 2026-10-01T10:30:17.000Z - Description: Feather Robotics raised $7.6M in pre-seed led by Gradient Ventures and opened orders for a $29,990 wheeled humanoid developer platform — roughly one-third the price of rival research machines. It has booked more than $1M selling to kitchens and labs, and bets the body, not the model, is scarce. - Author: Nexi.fund Labs - Tags: Defence & Robotics, #mode-5, #hook-paradox, #track-E On a Tuesday afternoon in late September, a robot about the height of a person rolls across a restaurant kitchen in East Palo Alto, picks up a pan and sets it back down. It is not a research demo and it is not a render. You can order one from a web page for $29,990. The machine is called Feather. It is a wheeled, two-armed humanoid built by Feather Robotics, a company that opened for business this month with $7.6 million in pre-seed funding led by Gradient Ventures, Google's AI venture arm. Builder Capital, Geometry, SEED Innovations and Virgo VC joined the round. The pitch is narrow and deliberate: a reliable humanoid body that third-party developers can program, sold at a price that makes the hardware a line item rather than a research budget. ``` CHRONICLE: Feather Robotics ───────────────────────────────────────────────────────────── Founding ── Prototype ── First order ── Revenue ── 24 Sept 2026 🔬 🧪 📦 💰 ◉ LAUNCH ex-Google/ 2 months 9 months >$1M $7.6M pre-seed Tesla team to build to ship booked + $29,990 on sale ``` Company timeline, compiled from TechCrunch, Humanoids Daily and SEED Innovations disclosures, September 2026. Feather is not chasing the fully autonomous general-purpose robot that Tesla and Figure market. It is selling the body and the software toolkit, and leaving the applications to everyone else. The founders describe the goal as an Android for robotics: a common platform other companies build on. ## What $29,990 buys The specifications are modest in the places that matter for cost and useful in the places that matter for work. Feather reaches about one meter, travels 600 millimeters vertically to reach a floor or a high shelf, runs on two hot-swappable batteries for a ten-hour shift, and sits on a holonomic wheeled base that moves in any direction without turning around. The reach and the battery are the two numbers Feather leads with, and they are the two that decide whether a robot survives a real workplace. A machine that cannot pick something off the floor, or that has to stop for an hour of charging, is a demo. A machine that works a full shift without a break is a tool. $29,990 Feather unit price #### The list price Orderable online now, described as a discounted launch price. · *feather.dev, 2026* 10 h dual hot-swap battery #### Shift length Two batteries, swapped without stopping the task. · *Humanoids Daily, 2026* On the software side, the robot runs models from outside suppliers — Nvidia, Skild and Physical Intelligence are named — through Feather's own controls and a software development kit. That choice is the whole strategy. Feather is not betting that its motion software will out-perform a lab. It is betting that developers will pay for a body that does not break, and that the best control models will arrive from somewhere else. ## The price is the product Humanoids have spent a decade being impressive and unaffordable. The breakthrough Feather is selling is not autonomy. It is unit economics. A developer who has to spend six figures per unit cannot test twenty of them; at $29,990, that math changes. The comparison the company invites is with the low-cost humanoid makers that set the market's floor. Feather prices itself at roughly one-half to one-third of those systems, and the gap is wide enough to restate the entire deployment question. | System | Price (approx.) | Form | Aimed at | | ------------------ | --------------- | ------------------------ | ---------------------- | | **Feather** | ✔ $29,990 | Wheeled, dual-arm | Third-party developers | | **Unitree H2 Edu** | ◐ \~$68,000 | Bipedal | Research and education | | **Dexmate Vega** | ◐ \~$83,000 | Bipedal, dexterous hands | Research | | **Agibot G2** | ✗ \~$135,000 | Bipedal | Industrial pilots | List prices cited by Feather's launch materials and Humanoids Daily, September 2026\. Bipedal systems carry the balance and actuation cost that Feather's wheeled base avoids. The wheeled base is the cost decision that makes the rest possible. Legs are the single most expensive, least reliable part of a walking humanoid, and they exist to climb stairs and cross rough ground. A factory floor, a commercial kitchen and a lab bench are flat. Drop the legs, and the bill drops with them. #### Why a wheeled base is a feature, not a compromise A bipedal robot spends energy and control complexity staying upright. Every falling risk is a safety case, an insurance question and a maintenance event. A holonomic wheeled base is stable by construction, moves sideways and diagonally, and returns the saved battery capacity to the arms. Feather's designers are explicit that they traded the ability to climb stairs for reliability and price. For the buyers they name — restaurants, labs, light manufacturing — that is a trade worth making. **The bet:** most paid humanoid work in 2026 happens on flat floors, so legs are a cost, not a capability. ## Who is buying, and why now Feather says it has been selling in small numbers for about a year, has passed $1 million in revenue, and has robots working as cooks in restaurants and cleaning in science labs. It declines to name the customers. Both of those details are the story. A humanoid that has booked revenue from commercial kitchens is past the demonstration stage; a company that will not name its buyers is still proving that the demand is repeatable. > "When I looked for a hardware platform to build on top of, I realized that none of them had the unit economics, reliability and support logistics that I needed to succeed as an independent developer."— Hoa Mai, co-founder and chief executive, Feather Robotics The team behind the price is part of the pitch. Hoa Mai is a serial founder whose previous company, Kind Humanoid, was acquired by 1X. His co-founder, Parsa Bakhtiari, ran part of Tesla's Model 3 production ramp in 2018 and reported to Elon Musk during the company's manufacturing crisis. The lesson the two carry is a manufacturing one, not an AI one: cost, yield and field support decide whether a robot ships, and the software can come from the ecosystem. ## The bet, stated plainly Feather is making a specific wager about where value sits in physical AI. If control models become a commodity — many suppliers, all good, all interoperable — then the scarce asset is the reliable, cheap body everyone runs them on, and the company that owns the body owns the platform. If instead the winning control model stays proprietary and tightly bound to its own hardware, Feather becomes a component supplier to someone else's stack. The funding is sized for the first case. A $7.6 million pre-seed does not buy a general-purpose robot. It buys actuator prototyping, a little low-volume production, and a developer ecosystem — enough to find out whether outside developers actually show up. Price is the pitch. That framing is why the number matters more than the specification sheet. Every humanoid maker claims improving autonomy. Feather claims a lower cost of failure, and lets its customers take the autonomy risk on hardware they can afford to break. ## What it changes, and what it does not A $29,990 humanoid does not make robots cheap. It makes the first robot cheap enough to be a business decision at a small company, which is a different threshold. Twenty years of industrial automation ran on the premise that the machine has to last a decade to justify its cost. A fleet you can replace, upgrade and experiment with changes the buyer from a large factory to a startup. As we wrote in September, the humanoid field is still divided over whether to teach robots the work or to copy it, with [Vesoma building its case around learning on the job](https://nexi.fund/vesoma-humanoid-2026/). Feather sidesteps that debate. It does not need to win the argument about how robots learn if it can sell the hardware both sides need. The open questions are the unglamorous ones. Can the company hold the price once the launch discount ends and volumes rise? Can it support a global developer base from a small team? And will the control models that make the arms useful arrive fast enough to justify buying the body first? 📊 **Signals to watch** Whether Feather holds $29,990 after the introductory price, or raises it as production scales. Whether named enterprise customers appear, or the company stays quiet about who is buying. Whether revenue compounds past the reported $1 million as the developer base grows. Whether the next $50,000-class rival matches the price, which would turn a one-company advantage into a market floor. The honest read is that Feather has not solved the humanoid problem. It has changed which part of the problem is worth paying for, and put a number on it that a small developer can say yes to. In a field full of promises about what robots will eventually do, a $29,990 price tag is the most falsifiable claim on the page. [ Meet Feather, the startup building the Android of robotics for developers TechCrunch's launch interview — the modular system, the $1 million revenue figure and the customers Feather will not name. TechCrunch ](https://techcrunch.com/2026/09/24/meet-feather-the-startup-building-the-android-of-robotics-for-developers/?ref=nexi.fund) The primary account of the launch and the company's own framing. [ Feather Robotics raises $7.6M pre-seed to build a humanoid platform The full company announcement as disclosed by investor SEED Innovations — round structure, founders and the rival price comparisons. FT.com / SEED Innovations ](https://markets.ft.com/data/announce/full?dockey=1323-17806511-2M3I94OC4OOELCOR3VD65QHK1B&ref=nexi.fund) The investor-side disclosure, useful for the hard numbers on price and thesis. [ Feather launches a $30,000 wheeled humanoid for developers A trade-press read on the specification, the dual hot-swap battery and the limits of what the company has disclosed. Humanoids Daily ](https://www.humanoidsdaily.com/news/feather-launches-30000-wheeled-humanoid-developers-7-6m-funding?ref=nexi.fund) Specialist coverage that separates the launch claims from verified detail.