spatialcomputing.info

NEO is a humanoid robot from Palo Alto-based robotics startup 1X that could one day autonomously help out around your home. For now though, it’s using human operators wearing Quest 3 headsets to pick up the slack and teach Neo’s neural network the ropes in the process.

The News

The 66 pound, sweater-wearing robot doesn’t seem very capable on its own, at least for now. As Joanna Stern from The Wall Street Journal puts it in a recent interview (seen below), Neo’s neural network needs a helping hand to “learn from more real world experience,” like how to load a dishwasher, fetch a bottle of water from the fridge, or clean up the kitchen counter.

It’s doing that by looping in humans who wear VR headsets to teleoperate Neo during user-scheduled intervals in ‘Expert Mode’, which also includes a configurable list of more difficult tasks you want a Neo’s teleoperators to perform.

While Neo is said to be capable of numerous household tasks, it can’t pick up overly heavy items, hot things, or sharp things, 1X CEO and Founder Bernt Børnic says, which probably means it won’t be cooking a soufflé or helping you move a bedside table any time soon.

There are a few more limitations. While Neo can be limited from entering specific parts of your home, and even potentially blur out humans to preserve privacy, Børnic frames privacy concerns as a sliding scale of user permissions. The more data you give up, the better Neo can work.

“I think it’s quite important for me to just say that, in 2026, if you buy this product, it is because you’re okay with that social contract,” Børnich tells WSJ. “If we don’t have your data, we can’t make the product better. I’m a big fan of what I call like ‘big brother, big sister’ principle, right? Big sister helps you. Big brother is just there to kind of monitor you. And we are very much the big sister. Depending on how much you want to trade, we can be more useful, and you can decide where on that scale you want to be.”

Notably, during Neo’s demo with WSJ, the robot wasn’t performing any tasks autonomously. Still, Børnich says that Neo will do “most of the things in your home autonomously” when it ships next year, noting that the quality of work will “vary and will improve drastically quite fast as we get data.”

Neo is currently available to reserve for a $200 refundable fee, arriving in three different colorways; grey, taupe, and black. Neo is expected to ship in the US sometime in 2026, priced at $20,000 for outright ownership, or for a $500 monthly subscription. Watch the full demo and WSJ interview below for more:

My Take

For now, if you can afford to hire a daily maid service, you can probably afford Neo, which makes the $20,000 price tag an intriguing value proposition.

After all, there aren’t any traditional employment contracts, sick days, domestic insurance policies, or any other quality of life concerns to contend with—only a four hour battery life that can be quick charged when it runs flat. That, and you’ll probably need to launder Neo’s turtlenecked jumpsuit every once in a while.

Still, I’m not so sure Neo delivers as a maid replacement quite yet, at least after watching WSJ’s demo. Even with a human controlling it, who uniquely understands how a refrigerator door works and how to fold a shirt neatly, it still struggles to do basic tasks better and faster than a human without an expensive robot body. Okay, faster isn’t exactly that important, since you can schedule tasks when you’re away.

That said, some of that initial awkwardness seen in the WSJ demo can be explained by the input method itself. As they are today, off-the-shelf consumer VR headsets like Quest 3 don’t provide the sort of proprioception you need to instantly understand that a surface is soft and pliable, or hard and fragile. An onboard neural network might, but we haven’t seen that in action yet beyond what the company’s hype video, seen below.

Let’s call it what it is: experts teleoperating Neo is a stopgap measure. As it is now, having a teleoperator use a headset and two motion controllers to guide a robot makes for some very slow and deliberate movement, as they’re likely only going on visual cues alone—a bit like trying to pour a glass of red wine when both your hands are completely numb. There’s also probably a fair bit of latency too, but if the job gets done, then you probably don’t care who did it—neural network or telehuman—or what challenges they faced along the way.

For now, I wonder how much time a teleoperator will have to spend painstakingly pouring laundry detergent into a cup, only to accidentally spill it. Does it have the dexterity to clean up the sticky mess? And what happens if it knocks over my priceless Ming vase? Or the dog takes a leak on my marble floors and Neo slips and falls, covering itself with dog piss? What sort of mess will I (or my human maids) have to contend with when I get back to my sensibly decorated villa?

Forgive the flagrant ‘rich guy’ framing above, but those are the only realistic scenarios I can muster at this stage. Whatever the case, I think it’s far too early to jump to the usual perils provided by Black Mirror, I, Robot (2004), or even Battlestar Galactica. More than anything, I think people buying Neo will more likely be asking themselves “what will my friends say when they learn I have a robot butler?” And probably not much else.

Still, there are other implications that go beyond using Neo as a walking, talking Google Nest, as neat as that may be. Provided Neo owners can connect using a Quest 3 like we see in the hype video, you might be able to telecommute back home to see where you left your keys, make sure the back door is locked, or check up on the dog to make sure they haven’t gotten into the pantry. Maybe not worth $20,000, but it’s a neat use case 1X should leverage if they haven’t already.