As cities increasingly pursue smart development and carbon neutrality, autonomous driving technology has emerged as a critical force reshaping urban mobility and spatial planning. By optimizing driving behaviors and enhancing energy efficiency, autonomous vehicles (AVs) are projected to reduce transport-related emissions by up to 34%. Yet beyond emissions reduction, AVs are driving a broader transformation—from car-centric planning paradigms to integrated systems that emphasize coordination among people, vehicles, and the urban environment. This paper conducts a systematic examination of autonomous driving technology’s developmental trajectory and its spatial consequences, with particular emphasis on urban street environments. Tracing the development of AVs from early conceptual prototypes in the 20th century to recent breakthroughs in artificial intelligence, deep learning, and vehicle-to-everything (V2X) communication, it identifies four key phases: emergence (pre-1970s), initiation (1980s–1990s), expansion (2000s), and breakthrough (post- 2010s). Each phase corresponds with evolving urban design theories, shifting the role of streets from utilitarian transport corridors to multifunctional public spaces that support diverse social and mobility functions.The global landscape of AV integration into urban planning is illustrated through a series of diverse case studies. San Francisco’s Smart City initiative reallocated road space from cars to pedestrians and cyclists while introducing autonomous ride-hailing services. Although these interventions were designed to promote more human-centered mobility, empirical evidence indicates that the deployment of AV-based ride-hailing may have inadvertently contributed to increased congestion due to a rise in vehicle miles travelled. In contrast, Toyota’s Woven City in Shizuoka represents a purpose-built prototype that establishes a differentiated street hierarchy comprising AVdominant corridors, pedestrian promenades, and ecological linear parks. This modular and scalable framework exemplifies how AVs can be seamlessly integrated with recreational and environmental priorities. Meanwhile, Toronto’s “City of Tomorrow,” proposed by Sidewalk Labs, envisioned a highly connected smart district driven by AVs and data-centric infrastructure. Despite its ambition, the project was ultimately cancelled due to growing public concerns over data privacy, opaque governance processes, and potential monopolistic control. Finally, Chongqing’s AI-integrated transit system in China offers a more pragmatic approach, featuring real-time traffic signal coordination, vehicle-to-infrastructure (V2I) technologies, and autonomous minibuses. While demonstrating the operational feasibility of AVs in mixed-use urban settings, it also exposes current limitations in deploying AVs within complex and dynamic open-road environments. These cases illustrate the diverse impacts and trade-offs of AV-led planning. While some strategies enable ecological and pedestrian-oriented design, others raise new challenges around congestion, social acceptance, anddigital governance.This paper explores how autonomous driving technology is transforming street space design and urban planning. 1) AVs can reclaim road space by reducing lanes and eliminating parking, reshaping street functions and elements. 2) They prompt a shift in right-of-way priorities, requiring integration of the “complete streets” concept with emerging technologies, as cities move toward more pedestrian- and cyclist-friendly environments. 3) Street governance must evolve to include dynamic lane use, flexible loading zones, and smart infrastructure. Further, the integration of end-to-end autonomous driving systems and large language models introduces new challenges and opportunities for planning. Physical street elements need to be translated into semantic data via multimodal perception and structured representation, requiring planners to engage in the full semantic modeling process. Human behavioral responses to complex environments should also be captured through semantic networks. At the infrastructure level, proactive development of digital twin systems and integration of V2X technologies are essential for supporting multimodal traffic coordination within the urban fabric.