Humanoid home robot tested: 1X Neo can vacuum, but it can't crack a walnut

The 1X Neo humanoid home robot, unveiled in 2024, has recently become available for pre-order in the US. This smart humanoid robot is competing with other AI household helpers, such as Figure 03 and Tesla Optimus. However, doubts remain about the actual capabilities of such a relatively novel product aimed at the mass market. In light of the recent spectacular failure of the first Russian humanoid presentation, a certain scepticism is perhaps warranted.

In the case of 1X Neo, Joanna Stern of the Wall Street Journal tested its capabilities and identified areas where it still needs improvement. The robot's developer, Bernt Børnich, also participated in the test. The new household helper is said to be capable of washing dishes, cleaning the kitchen and doing laundry, among other tasks.

Safety, intelligence and physical abilities

When it comes to home robots, safety is a key consideration alongside general capabilities and intelligence. Unlike its industrial counterparts, the robots that work on assembly lines in large halls, Neo wears a textile covering reminiscent of clothing. This textile covering enhances its aesthetic appeal and provides greater safety.

Inside Neo are several powerful yet lightweight motors driven by belts, loosely inspired by muscles, rather than gears. This makes the robot quieter and results in smoother movements. Furthermore, its weight can be reduced to 30 kg (66 pounds). This is particularly important in case it falls and hits something. Despite its light weight, it can lift up to 68 kilograms (150 pounds).

Is 1X Neo a super robot or a toddler?

While one might quickly assume that this is a particularly powerful robot, that's not the case. The 1X Neo failed to crack a walnut by hand. This is apparently intentional, as the 1X Neo prioritises finesse and sensitivity over brute force. It's also worth noting that this is the first model and that there's plenty of room for improvement.

It took the robot a full minute and seven seconds to retrieve a bottle of water from the refrigerator and hand it to the test subject a few steps away. Opening the refrigerator door proved to be the main challenge. The test also examined many other aspects, such as loading a dishwasher and charging cycles. Neo apparently performs very well with a vacuum cleaner.

Control via teleoperation

It is important to note that the model that was tested was still controlled remotely by an AI operator. It is a predecessor to the model that is scheduled for release in 2026 and is expected to feature greater dexterity, increased autonomy and enhanced safety. Rather than being a true household helper, Neo will initially function more like a child, requiring instruction and being connected to human support via an app. 1X refers to its remote control mode as 'expert mode'. In this mode, a human operator connects to the robot to take over tasks that it cannot (yet) solve autonomously.

To train the AI, visual and acoustic data from the customer's home must be sent to 1X's servers via the robot's cameras and microphones. The company assures customers that faces will be pixelated and that they can always see when an operator is active. Nevertheless, for many, this is the biggest compromise regarding data privacy. The developers' goal is for Neo to learn as much as possible through its use in private homes so that, in about five years, it can become a fully developed assistant.

Neo thus represents the beginning of an era of humanoid household robots. According to the developer, once fully developed, Neo will provide great assistance in a wide variety of household tasks, not only for standard applications, but also for people with disabilities, the elderly and other groups. Pre-orders are currently only available in the USA, requiring a $200 deposit. Those interested can choose between a one-off payment of $20,000 or a monthly subscription of $499. However, it is expected to be available for purchase in Europe by 2027. To gain a more comprehensive understanding, we recommend watching the videos listed below.