NEO – what 1X just launched

NEO is 1X’s consumer-ready home robot built to take on real chores, hold natural conversations, and learn new skills over time. Pre-orders are open today. Early Access units are scheduled to ship in 2026 in the United States first, with broader markets following in 2027. You can choose from Tan, Gray, or Dark Brown.

What NEO actually does at home
Owners give NEO a list of chores, set a time, and come back to a tidier space. Think folding laundry, organizing shelves, putting things away, and answering the door. If a task is new, you can schedule a 1X Expert who remotely guides the robot so the job gets done while NEO learns the skill. Interaction is voice-first and there is a companion app for scheduling, remote check-ins, and teleoperation.

How you talk to it and how it talks back
NEO listens when addressed, thanks to Audio Intelligence, and uses Visual Intelligence plus Memory to ground conversations in what it sees and what it remembers about you. You can control it by voice, manage tasks from the mobile app, or even pilot it through a phone or a VR headset. Status is visible through Emotive Ear Rings that signal attention and battery, and the built-in three stage speakers double as a mobile home audio system.

---

---

Safety and design for living rooms, not labs
The body is wrapped in a soft 3D lattice that cushions contact and hides pinch points. Actuation is tendon-driven for low inertia movement around people. Head Injury Criterion is listed under 250 and acoustic noise around 22 dB, which is quieter than a modern refrigerator. The knit suit and shoes are machine-washable. Hands are rated IP68 for submersion and the body is IP44 for splash resistance.

Human scale hardware and dexterity
NEO stands about 5 feet 6 inches tall, weighs roughly 30 kilograms, and brings 22 degrees of freedom per hand for human-like grasping. Total degrees of freedom across the body come to 75. The robot can lift about 70 kilograms, carry roughly 25 kilograms, and each arm supports an 18 pound payload for everyday manipulation. Fingers can move at up to 8.0 meters per second for quick, precise actions.

Mobility that covers the whole home
Typical walking is about 1.4 meters per second which is roughly 5.0 kilometers per hour, while the listed top running speed is 6.2 meters per second which is roughly 22.3 kilometers per hour. Beyond raw speed, 1X’s Redwood Mobility controller teaches NEO to sit, kneel, lie down and get back up, and to go up and down stairs using stereo vision, all under a unified controller that supports both AI and teleoperation.

Power, charging, and autonomy
A battery capacity of 842 watt-hours targets around 4 hours of run time in typical use. Quick charging is described as about 6 minutes per hour of runtime. When power runs low, NEO navigates to charge on its own, so you do not have to think about it.

Sensing, compute, and connectivity
Vision comes from dual 8.85 megapixel stereo fisheye cameras at 90 hertz, paired with four beamforming microphones. Compute is centered on the 1X NEO Cortex, based on NVIDIA Jetson Thor, with up to 2070 FP4 tera-operations for onboard AI. Wi-Fi, Bluetooth, and 5G are included, and the speaker system is located in the pelvis and chest for room-filling voice and audio.

The brain: Redwood AI plus a built-in LLM
NEO runs a built-in large language model for conversational control and uses Redwood AI, 1X’s vision-language model for end to end mobile manipulation and whole-body control. Redwood is designed to run fully on NEO’s embedded GPU, trained on real world teleoperation and autonomous episodes, and operates at about 5 hertz with around 160 million parameters. The research team also uses a physics-grounded World Model to predict task success and speed up iteration. These ingredients are what let the robot do chores naturally and improve in homes over time.

Why it matters
Humanoids have been inching from research to real deployment. NEO is the clearest pitch yet for a consumer robot you can live with. It does useful work today, it is designed to be safe to share a kitchen with, and its AI stack is built to learn from daily life, not just lab demos. Pricing and timelines are explicit in the launch materials which makes this feel like a product rather than a prototype.

Check out more of NEO here