ByteDance is exploring autonomous driving technologies as part of its broader artificial intelligence research, according to a report by Chinese technology media outlet 36Kr, although the company said it has no plans to enter the smart driving business.
Citing multiple industry sources, the report said the initiative is being led by the world model team within ByteDance’s Seed AI research division. The team is working on technologies related to environmental perception, prediction and planning, capabilities that are also central to autonomous driving systems.
Autonomous Driving Research Underway
According to 36Kr, ByteDance is studying applications including autonomous logistics, with the project operating under the automotive business unit of its cloud computing platform, Volcengine.
The report said the company has held discussions with established autonomous driving teams and has begun preparations for the project. It also claimed ByteDance has approached engineers and researchers with expertise in advanced driver assistance and autonomous vehicle technologies.
Responding to the report, ByteDance said its frontier AI research includes physical AI but denied plans to commercialize a smart driving business.
“Our frontier large-model exploration areas include physical AI, where there are many early-stage research efforts and explorations, but we have no plan to build a smart driving business.”
36Kr said the company’s response suggests it currently views the effort as research rather than the development of a commercial autonomous driving platform.
World Models Become Strategic Focus
Seed, ByteDance’s large-model research organization established in 2023, is responsible for foundational AI research within the company.
The report noted that Zhou Chang, who joined ByteDance in 2024, oversees several AI research areas including multimodal large models, world models, visual generation and embodied intelligence.
ByteDance has already expanded into the automotive software sector through Volcengine and its Doubao large language model, supplying cloud services and intelligent cockpit technologies to vehicle manufacturers.
In June, automotive technology company Aiva announced a partnership with Volcengine to integrate the Doubao large model into its intelligent cockpit platform. However, Aiva is expected to source its advanced driver assistance system from autonomous driving startup DeepRoute rather than ByteDance.
Industry attention has increasingly shifted toward world models as a foundation for future autonomous driving systems, with developers seeking AI models capable of understanding and predicting complex real-world environments.
According to the report, ByteDance could eventually use data generated through autonomous driving research to improve its world models and support future embodied AI applications, including robotics.
The report also suggested that ByteDance’s financial resources, AI expertise and computing infrastructure could position the company as a new competitor if it chooses to expand further into autonomous driving technologies.