Tesla Optimus vs. BYD Xiao Di vs. LG CLOiD: Nvidia enters AI robot segment

LG and Nvidia are significantly expanding their robotics partnership in Seoul. A new data factory at LG's research campus in Yangjae is expected to generate a total of 100,000 hours of training data by the end of 2026. LG will combine real-world data from manufacturing, logistics and robot deployments with synthetic datasets, while Nvidia will provide technologies including Omniverse, Cosmos and the Isaac robotics platform. LG's CLOiD robots will be trained in recreated living spaces and simulated manufacturing and logistics workflows. One key difference from the approaches taken by Tesla and BYD is that CLOiD uses a wheeled AI robot base instead of walking on two "legs."

The data will be incorporated into LG's Robot Foundation Model to improve the capabilities of humanoid robotic systems. CLOiD is already practicing household tasks such as retrieving milk from a refrigerator, operating an oven, and folding and stacking clothes. The robot has two arms with seven degrees of freedom each and hands with five independently powered fingers. Unlike Tesla Optimus or Figure 03, CLOiD moves on a wheeled base rather than two legs. The 1X NEO is also designed for household use, while Agility Digit and Apptronik Apollo are more focused on logistics and industrial tasks.