This paper provides an initial survey of world models in the context of autonomous driving, highlighting their theoretical underpinnings, practical applications, and ongoing research efforts. World models have emerged as a transformative approach, enabling autonomous systems to synthesize and interpret vast amounts of sensor data, predict future scenarios, and compensate for information gaps. The paper discusses the evolution of world models from control theory to artificial intelligence, emphasizing their role in advancing autonomous driving technologies. Key components of world models, such as perception, memory, control/action, and the world model module, are detailed, along with their architectural foundations and applications in various domains. The paper also explores the challenges and future perspectives in the field, including long-term scalable memory integration, simulation-to-real-world generalization, and ethical and safety considerations. Despite the progress, significant hurdles remain, particularly in long-term memory, simulation-to-real-world adaptation, and ensuring decision-making accountability and privacy. The paper aims to serve as a foundational reference for researchers and practitioners, fostering continued innovation and exploration in the field of autonomous driving.This paper provides an initial survey of world models in the context of autonomous driving, highlighting their theoretical underpinnings, practical applications, and ongoing research efforts. World models have emerged as a transformative approach, enabling autonomous systems to synthesize and interpret vast amounts of sensor data, predict future scenarios, and compensate for information gaps. The paper discusses the evolution of world models from control theory to artificial intelligence, emphasizing their role in advancing autonomous driving technologies. Key components of world models, such as perception, memory, control/action, and the world model module, are detailed, along with their architectural foundations and applications in various domains. The paper also explores the challenges and future perspectives in the field, including long-term scalable memory integration, simulation-to-real-world generalization, and ethical and safety considerations. Despite the progress, significant hurdles remain, particularly in long-term memory, simulation-to-real-world adaptation, and ensuring decision-making accountability and privacy. The paper aims to serve as a foundational reference for researchers and practitioners, fostering continued innovation and exploration in the field of autonomous driving.