According to a research report released by Dongwu Securities, the development path of the digital AI base model has gradually become clear, and the value lies in certainty; the future growth of this round of AI will increase even more in physical AI, and intelligent driving is a representative scenario where it first runs through a closed loop. The closed-loop path of physical AI from characterization to deployment is clearly different from digital AI, and will form an independent and unified development path. As the closed loop of world models and reinforcement learning matures and cloud vehicle deployment gradually materializes, physical AI is expected to move from ability demonstration to large-scale implementation, becoming the most flexible direction for AI development at this stage.

According to a research report released by Dongwu Securities, the development path of the digital AI base model has gradually become clear, and the value lies in certainty; the future growth of this round of AI will increase even more in physical AI, and intelligent driving is a representative scenario where it first runs through a closed loop. The closed-loop path of physical AI from characterization to deployment is clearly different from digital AI, and will form an independent and unified development path. As the closed loop of world models and reinforcement learning matures and cloud vehicle deployment gradually materializes, physical AI is expected to move from ability demonstration to large-scale implementation, becoming the most flexible direction for AI development at this stage.