1X Technologies begins commercial sales of its Neo humanoid robot in California, featuring Nvidia Jetson Thor and 18,000 tactile sensors.
A Humanoid Robot You Can Actually Buy
1X Technologies has moved its Neo humanoid from demonstration reels into a product you can purchase, starting with customers in California. That shift matters more than any single spec. A research prototype only has to work in a controlled lab; a robot for sale has to survive shipping, unpredictable homes and offices, and users who did not build it. Limiting the initial rollout to one region is a sensible way to keep support, servicing, and iteration close while the product finds its footing.
Commercial availability also changes the questions people should ask. Instead of "can it walk," the useful questions become "what can it reliably do without supervision," "how does it fail," and "what happens when it needs a repair." Those are the details a regional launch is designed to surface early, before the machine ships to a wider audience.
Onboard Compute: Nvidia Jetson Thor
Neo runs on Nvidia's Jetson Thor, an edge computing platform meant to handle heavy AI workloads directly on the robot rather than in a remote data center. For a humanoid, keeping perception and control local is a practical requirement, not a luxury. A robot that has to phone home for every decision inherits network latency and connectivity risk into its motion loop, which is exactly where you cannot afford delay.
On-device compute of this class supports the parts of the job that have to run continuously and quickly:
- Interpreting camera and sensor input to build a picture of the surroundings.
- Planning and adjusting movements as the environment changes.
- Running learned models for manipulation and navigation without a round trip to the cloud.
Why 18,000 Tactile Sensors Matter
Vision tells a robot where an object roughly is; touch tells it what is actually happening at the moment of contact. Distributing tactile sensing across the body — 18,000 points, in Neo's case — gives the robot dense feedback about grip, pressure, and slip. That feedback is what separates a hand that crushes a cup from one that holds it, and a body that plows into furniture from one that notices contact and backs off.
Dense touch sensing is especially important for a machine meant to operate around people and everyday objects. Many household and workplace items are fragile, oddly shaped, or deformable, and they cannot be handled well by force alone. Feeling contact lets the robot close the loop in real time: apply just enough pressure, detect when something is slipping, and stop before it does damage.
What to Watch as Neo Reaches Real Homes
The interesting story now is behavior over time, not the launch itself. A humanoid earns trust by doing modest tasks dependably and by failing safely when it hits something it was not built for. Buyers and observers should watch how consistently Neo completes routine work, how it handles the messy edge cases a lab never sees, and how gracefully it recovers when it gets something wrong.
A single-region commercial release is best read as the start of a feedback loop rather than a finished product. Real users generate the situations that drive the next round of improvements, and how quickly those lessons translate into better, safer behavior will say more about Neo than any first-day capability list.