On the cozy home robot, pre-orders have opened. But for now, it works remotely.
Startup 1X Technologies has finally launched applications for early testing of its home robot Neo — a soft humanoid in a beige sweater, poised to become the first mass-market robot housekeeper. The cost: $20,000 to buy or $500 per month to rent, with delivery in 2026.
.jpg)
But there's a catch: the robot can't yet operate independently. Neo was demoed to journalists at the company's Palo Alto headquarters — and it turned out every action was controlled by a human in a VR helmet from the next room.
Loading a couple of plates into the dishwasher took five minutes. It folded a sweater crookedly. It fetched a water bottle from the fridge — but only after an awkward dance with the door.
"It was wild," most attendees summed up. And CEO Bernt Børnich was candid: "If you buy this product, you're agreeing to that social contract. Without your data, we can't make the product better."
.jpg)
$20,000 for a 30-kilogram humanoid with AI is peanuts by industry standards. But 1X's business model is built on data collection right now. Neo is essentially an investment in the future of AI robotics, not a finished product. The company promises that by 2026, the robot will handle most tasks autonomously, trained on millions of hours of teleoperation. Models learn not just from teleoperator recordings but from world models — generative systems that simulate real-world physics.
One Twitter user suggested a lifehack: rent Neo for $500 and make the teleoperator grind eight hours on a factory line — at California's minimum wage of $16.50 an hour, it pays for itself in a month. Pure labor arbitrage.
Replies to the announcement mostly hit three themes: privacy, teleoperation, and ethics.
Spying. "People will really pay $20,000 to have a robot spying on them in their own home?" The company promises safeguards: operators connect only with owner permission, you can set no-go zones, faces in video feeds are blurred. But trust is shaky: the mere idea of another human staring out from the robot is unsettling.
Teleoperation as deception. "Every action was manually controlled, even speech. And it was still slow and clumsy," early testers complain. Haters compare 1X to Theranos, calling teleoperation "fraud." The issue: Neo is marketed as an AI robot, not a remote-controlled toy.
Labor ethics. The meme "AI = An Indian" went viral: supposedly low-paid operators from India or other poor countries are behind Neo. Parallels to colonialism and call-center outsourcing are obvious.
The concerns are valid, but this is a necessary stage. Without real user data, improving the robot is impossible — virtual simulation won't cover every scenario. Teleoperation isn't deception; it's a way to rapidly build a dataset while AI matures. The alternative — releasing a robot that confuses hot and cold or drops fragile items — is a dead end.
Yes, the price is subsidized by data. Yes, privacy is in question. But 1X bet on tactility and design — Neo isn't a Terminator; it's a cozy helper in a sweater.
.jpg)
Also read:
- China Deploys DeepSeek on Unmanned Tanks
- Volinga Unveils Suite V 0.2: Faster Splat Training and Seamless Unreal Engine Integration
- Delphi: Digitizing Minds for Multi-Armed Buddhas in a Time-Starved World
The company is backed by OpenAI and other big players, and the launch party featured Silicon Valley's elite. The team is a dream: VP Operations Vikram Kothari spent 8 years at SpaceX, leading supplies for Dragon and Starship. VP Sales Jorge Milburn had 10 years at Tesla, building Supercharger.
VP Design Per Selvaag worked at BMW and Peugeot. Plus a partnership with NVIDIA GEAR Lab — Neo uses specialized foundation models for robots, NVIDIA GR00T. All this adds weight to 1X and explains the hype.
Competition in robotics is fierce — hundreds of startups vie for the market. 1X is taking a risk, but their "data first, autonomy later" approach could pay off. If, of course, they don't burn out in a privacy scandal first.
