Daily Technology
·20/01/2026
U.S. robotics firm 1X has announced a significant update to its humanoid robot, NEO, integrating an AI model that enables autonomous learning from video content. This update marks a technical leap in robotic learning, shifting from traditional, human-guided training methods to direct acquisition of skills through real-world video exposure.
Historically, humanoid robots required extensive programming or human demonstration to acquire new tasks. Typically, dedicated operators would collect data, and robots would imitate these actions through repeated guidance. This model is labor-intensive, costly, and can result in slow progress due to the constraints of human labor availability and physical limitations.
In contrast, 1X’s updated NEO utilizes the 1X World Model AI, which learns directly from internet-scale videos. Rather than relying on large datasets created by human trainers, the robot analyzes and extracts knowledge from unstructured video material. This allows for rapid scaling of skills and removes much of the tedious, physically demanding labor from the robot operator’s role.
The updated NEO demonstrates the ability to convert simple text or voice prompts into complex action plans, even when encountering unfamiliar objects or scenarios. For example, NEO performed tasks such as packing a lunchbox, opening sliding doors, operating toilet seats, ironing clothes, and brushing human hair, entirely guided by its interpretation of video demonstrations.
NEO’s performance is further validated by its capacity to plan and sequence tasks based on video-derived strategies, displaying adaptability not commonly seen in traditional robots. The model’s grounding in real-world physics enables applications beyond what is pre-programmed or rigidly defined, placing it ahead in generalization capabilities.
The 1X World Model allows the scaling of robotic intelligence with the number of deployed robots, a shift from dependence on the number of human operators or programmers. According to 1X CEO Bernt Bornich, this means each additional NEO robot contributes to the learning ecosystem, exponentially improving the overall knowledge base.
This advancement has the potential to alleviate physically strenuous and sometimes hazardous operator roles, as noted by industry feedback on similar projects, such as Tesla’s Optimus. Operators previously reported injuries attributed to the repetitive physical nature of robot training assignments, highlighting a key benefit of increased robot autonomy.
NEO is available at a one-time purchase price of $20,000 or a $500 monthly subscription, with the first units expected to ship later this year. Early deployments will still utilize remote human operators for tasks beyond current AI capabilities, but the robot’s ongoing data-gathering is anticipated to refine the model further.
With the ability to autonomously learn complex skills from videos, the new NEO represents a noteworthy progression in robotics and artificial intelligence. Its model has the potential to accelerate skill acquisition, reduce operational costs, and improve workplace safety, supporting broader and more versatile adoption in various industries.









