Home: Motoring > 2026 Beijing Auto Show Spotlights Physical AI as Automakers Accelerate Embodying Intelligence and Full-Scenario Ecosystems

2026 Beijing Auto Show Spotlights Physical AI as Automakers Accelerate Embodying Intelligence and Full-Scenario Ecosystems

From:Internet Info Agency 2026-04-25 07:41:00

At the 2026 Beijing Auto Show, "Physical AI" emerged as the industry's central keyword. Spanning a total area of 380,000 square meters, the exhibition attracted over 2,000 companies from 21 countries and regions, showcasing 1,451 vehicles—including 181 global debuts and 71 concept cars. Technologies such as Physical AI, embodied intelligence, centralized computing platforms, and multimodal large models permeated the entire event. XPeng Group made its debut under its new name, unveiling its humanoid robot IRON, the modular flying car "Land Aircraft Carrier," and several new vehicles, including China’s first fully in-house developed, pre-installed mass-production Robotaxi prototype—the XPeng GX. The GX is equipped with up to four Turing chips, delivering a total computing power of 3,000 TOPS, and features the second-generation VLA autonomous driving system. XPeng announced it will invest RMB 7 billion in Physical AI-related R&D in 2026 and aims to commercialize its humanoid robots by 2027. The company also reported significant growth in both usage rates and order volumes for its Ultra models’ intelligent driving features, underscoring that advanced autonomous capabilities are becoming a decisive factor in vehicle purchasing decisions. Geely Automobile Group set up an independent tech ecosystem pavilion and unveiled Eva Cab, China’s first native Robotaxi prototype. Designed without a steering wheel or pedals and featuring a face-to-face cabin layout, Eva Cab is built on the EEA 4.0 electronic/electrical architecture and integrates NVIDIA’s SuperChip, Thor U, and Qualcomm Snapdragon 8797 chips, achieving over 3,000 TOPS of total computing power. It is also equipped with a 2,160-line digital LiDAR. At its core lies the WAM (World Action Model), enabling Level 4 autonomous driving. Geely plans to launch commercial operations via its mobility service Caocao Mobility in cities like Hangzhou and Suzhou by 2027. QCraft introduced a Physical AI architecture based on “world models + reinforcement learning” and launched its “Chengfeng MAX” advanced driver-assistance solution, offering over 500 TOPS of computing power and supporting urban NOA (Navigate on Autopilot) functionality. The company also shared updates on its L4 Robotaxi and Robovan developments and restructured its product lineup into three tiers—AIR, PRO, and MAX—to address diverse market demands. QCraft announced a refreshed mission statement: “Physical AI, Better Future,” signaling its strategic shift from an autonomous driving solutions provider to a foundational infrastructure enabler for general-purpose Physical AI. Multiple automakers simultaneously advanced their humanoid robotics strategies. XPeng, Li Auto, Xiaomi, Chery, BYD, GAC, Changan, Geely, and FAW all revealed robotics R&D or mass-production plans. XPeng’s IRON and Honor’s “Lightning” robot became standout exhibits. Li Auto plans to launch a dual-wheeled robot for factory applications within the year; Xiaomi’s CyberOne has already begun on-site training in automotive plants; and GAC established a wholly owned subsidiary, HuiLun Tech, dedicated exclusively to robotics. The supply chain landscape is also undergoing restructuring. Core suppliers—including CATL, WeRide, Horizon Robotics, and iFlytek—were showcased alongside OEM brands in the same halls for the first time, reflecting how Physical AI is elevating suppliers’ influence and voice within the automotive value chain. This year’s auto show marks a paradigm shift in the automotive industry—from vehicles as mere transportation tools toward embodied intelligent agents. Physical AI is now widely regarded as the critical framework integrating perception, decision-making, and execution capabilities, forming the foundational cornerstone for defining next-generation intelligent mobility and robotics technologies.

Editor:NewsAssistant