1X Technologies opens humanoid factory
- 1X Technologies opened its NEO Factory in Hayward, California, saying it is America’s first vertically integrated humanoid robot plant for home robots. - The new site spans 58,000 square feet, targets 10,000 NEO units a year now, and 100,000-plus annually by the end of 2027. - That matters because humanoid hype is shifting from demos to manufacturing — and consumer delivery is now the real bottleneck.
Humanoid robots are leaving the lab phase and entering the factory phase. That is the real news here. 1X Technologies has opened a new production site in Hayward, California, for its NEO home robot, and the company is pitching it as America’s first vertically integrated high-volume humanoid robot factory. The point is not just that another robot startup rented a building. The point is that 1X is trying to solve the ugly part of robotics — making lots of units, with repeatable quality, fast enough that “consumer humanoid” stops sounding theoretical. (1x.tech) ### What exactly opened? The new plant is a 58,000-square-foot factory in Hayward. 1X says it is now online, with another San Carlos facility expected later in 2026. The Hayward site is meant to handle the full stack of robot production more tightly under one roof — parts flow, assembly, testing, and the kind of iteration loop you need when the product is still improving every month. (1x.tech) ### What does “vertically integrated” mean here? Basically, 1X wants fewer handoffs. Instead of acting like a brand that mostly stitches together supplier parts, it is building more of the robot system itself and controlling more of the manufacturing process. For humanoids, that matters a lot. A robot is not one hard problem — it is motors, batteries, sensors, compute(1x.tech)are updates, all coupled together. When one thing changes, five other things break. Vertical integration makes those feedback loops faster. (1x.tech) ### Why is 10,000 units a big deal? Because most humanoid robot news is still demo news. A polished video is easy compared with building thousands of machines that behave consistently in homes. 1X says the Hayward factory has capacity for 10,000 NEOs per year, and with more automation plus the San Carlos site, it wants to push toward 100,000-plus units annually by the(1x.tech)cs experiment and an actual manufacturing ramp. (1x.tech) ### Did people actually buy these? Not exactly bought in the normal retail sense — but demand looks real. 1X says it booked out its first year of production capacity in five days after launching NEO on October 28, with 10,000 units spoken for. The company’s current store also offers Early Access for $20,000 with priority delivery in 2026, plus a $499-a-month subscript(1x.tech)t once — willingness to pay upfront, and willingness to treat a humanoid like a service. (1x.tech) ### Why does software matter so much now? Because once you have a fleet, every bug becomes a manufacturing problem. A humanoid shipped to a home is not like a dishwasher. It keeps learning, gets over-the-air updates, and has to behave safely in messy real-world environments. 1X has been pretty explicit that simulation is part of the stack — using NVIDIA Isaac tools, J(1x.tech)scenarios before those behaviors hit physical robots. That is how you keep the factory from becoming a bug amplifier. (1x.tech) ### Why is OpenAI part of this story? Mostly because it signals where the bet is. OpenAI led a funding round in 1X in 2024, and the company has been positioning NEO as a home robot that gets better through embodied AI, not just fixed automation. So this factory opening is also a test of a bigger thesis — whether frontier-model money can turn humanoids into a consu(1x.tech)bloomberg.com) ### What is the catch? The catch is that scaling a humanoid is brutally harder than unveiling one. Homes are chaotic. Reliability expectations are high. Safety failures are unforgiving. And 10,000 units is still tiny compared with mainstream consumer electronics. The factory solves one bottleneck, but not the hardest one — proving that NEO can do useful work often enough, safely enough, that people keep it around. (1x.tech) ### Bottom line? 1X did not just announce a robot. It announced a production system. That is a more serious move — and a more dangerous one, because now the company has to prove its software, safety, and operations can scale with the hardware. If that works, humanoids stop being a science-fair category and start looking like a real product business. (1x.tech)
