DeepRoute.ai (hereinafter referred to as DeepRoute.AI) attended the 2025 Seoul Mobility Show that opened on the 3rd and introduced its solution to the domestic market for the first time.

DeepRoot.AI introduced its advanced Visual Language Action (VLA) model and announced that more than five vehicles equipped with its VLA model will be launched in the Chinese market within the year.

The VLA model supports both pure camera solutions and LiDAR camera fusion configurations, and is compatible with various chip platforms to meet the needs of various automakers, it emphasized. This advanced system is capable of long-term contextual understanding and analysis of driving scenarios up to tens of seconds, and provides insight into the surrounding environment as well as predicted and planned actions by explaining the decision-making process step by step.

The SUV mass-production model in collaboration with Jangseong Motor was launched on the consumer market in August of last year, and DeepRoot.AI is said to be eyeing Korea as a key market, paying attention to the recent developments in the electric vehicle and smart driving fields in Korea. This is also the reason why we decided to attend this mobility show and introduce the VLA model to the Korean market.

At the exhibition that day, DeepRoot.AI demonstrated solutions such as △Level 2 smart driving system △Seamless parking to parking △Automatic emergency braking.

DeepRoot.AI is a Chinese startup founded by Dr. Ju Guang in 2019, headquartered in Shenzhen, and operates in various regions around the world. It has achieved a cumulative investment of $500 million to date and is dedicated to building general-purpose artificial intelligence for the physical world.

Through the realization of mass-produced smart driving vehicles and large-scale operation of robotaxis, DeepRoot.AI is accumulating high-quality de-identified data from the physical world, and announced that it has signed mass production cooperation agreements with major automobile manufacturers such as Great Wall Motors and Smart.