The era of Physical AI is upon us, as artificial intelligence, once confined to digital screens, now steps into the real world equipped with sensors and robotics. Physical AI is becoming a key driver in reshaping industries worth $50 trillion globally, extending beyond manufacturing and logistics to energy, mobility, and infrastructure, according to Deloitte. In the next decade, humanoids, autonomous robots, self-driving cars, and drones will become commonplace in urban and industrial settings, enhancing safety and productivity. It is essential to prepare the necessary spatial infrastructure to accommodate these advancements.
South Korea boasts one of the highest robot densities and a strong manufacturing base in the world. However, the advancement of Physical AI must now extend beyond factory walls to integrate with urban logistics, transportation, and spatial structures. Urban planning must evolve from merely coordinating residential and transportation needs to creating an 'intelligent spatial ecosystem' where robots and AI agents can seamlessly perceive and act.
This article proposes strategies to incorporate Physical AI into future urban planning to enhance South Korea's national competitiveness.
First, there is a need to build robot-friendly spatial infrastructure and 'urban digital twins.' Current roads and buildings are designed for human use, which significantly limits the movement of robots and drones. Urban planning should include underground logistics networks, dedicated robot pathways, and drone stations. Additionally, a 'digital twin' at the city level should be established to validate real-time spatial information, robot navigation routes, and energy consumption. For instance, Houston is developing dedicated driving data and 3D mapping infrastructure for autonomous mobility and robots in urban areas. Singapore has implemented a digital twin of the entire city to utilize logistics and robot simulations as public infrastructure.
Second, it is crucial to designate regulatory-free zones and testbeds centered around Physical AI clusters. The growth of Physical AI relies heavily on vast 'real-world data' for learning. Major advanced hubs should be designated as Physical AI-specific testbeds, and regulatory sandboxes should be expanded. An ecosystem where humanoids and delivery robots can freely navigate sidewalks and commercial areas while ensuring pedestrian safety is essential. Cities like Tsukuba and Fukuoka in Japan have been designated as robot special zones and experimental cities, allowing delivery robots and autonomous services to collect data and conduct commercialization tests without regulatory constraints.
Third, spatially distributed placement of data centers and edge computing infrastructure is necessary. Physical AI requires ultra-low latency for data processing, necessitating a reduction in reliance on central cloud systems and the establishment of 'edge computing' and power networks on-site. Urban planning must reliably incorporate large AI data centers and edge server hubs, which consume significant power and generate heat, while securing sites for intelligent power grids linked to renewable energy. Northern Virginia (Data Center Alley) and Dublin, Ireland, are examples of regions integrating large data center clusters and renewable energy-linked power infrastructure into urban planning to address the growing demand for AI computation and power issues.
Fourth, moving away from traditional zoning regulations to create 'mixed-use smart spaces' is essential. The strict separation of residential, commercial, and industrial zones poses a significant barrier in the Physical AI era, where manufacturing, logistics, and services converge. Urban areas should introduce mixed-use zones that combine logistics picking, small-scale eco-friendly manufacturing, and robot services. Flexibility in changing the use of commercial building lower levels or basements into robot stations should be increased. Cities like San Francisco and New York are allowing the flexible conversion of vacant commercial spaces and parking lots into small logistics and urban robot hubs.
Urban planning must now shift from a civil engineering mindset focused on constructing physical buildings to designing cities as 'intelligent organisms' where AI and robots can communicate and collaborate seamlessly. A meticulous strategy is needed to validate through simulations, proactively maintain spatial infrastructure, and operate cities based on data. This will serve as a solid foundation for South Korea to emerge as a global leader in Physical AI.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.