The year 2026 began with organizational changes at XPeng that were more drastic than in previous years. On February 3rd, it was learned that XPeng has formally and strategically merged its Autonomous Driving Center and its Smart Cockpit Center into a new department named the General Intelligence Center. XPeng confirmed to sources that "this adjustment is accurate." These two departments were previously parallel, first-tier departments. Following the restructuring, Liu Xianming, the former head of the autonomous driving business, has been appointed as the leader of this new center, reporting directly to Chairman and CEO He Xiaopeng. To outsiders, this might appear as a routine consolidation of resources, but industry insiders view it as a clear signal that XPeng is officially ending the decade-long, dual-track R&D paradigm of separate development for intelligent driving and intelligent cockpits, and is fully transitioning to a single-track era driven by Physical AI. This is also the direction in which numerous automakers are currently focusing their efforts. For a long time, the software and hardware architecture of smart cars has been divided into two key domains: intelligent driving and the intelligent cockpit. Within XPeng's R&D structure, the Autonomous Driving Center was responsible for the "hands, feet, and eyes," while the Smart Cockpit Center handled the "mouth and ears." Although this division of labor ensured professional expertise in the early stages, the proliferation of end-to-end large models is increasingly blurring the lines between them. Under the traditional distributed architecture, intelligent driving and intelligent cockpits possessed their own respective perception algorithms, computing platforms, and data pathways. This meant that when a vehicle needed to implement cross-domain functionalities, data had to be repeatedly transferred between different departments and systems. These technological silos, created by organizational barriers, not only increased R&D costs but also constrained system response speeds. With the establishment of the General Intelligence Center, XPeng will reorganize its secondary teams around a layered structure comprising the Foundation Model, the Infra base, and platform-based delivery. This signifies that both the decision-making processes for autonomous driving and the voice/visual interactions for the smart cockpit will, at their foundation, share the same AI infrastructure. An industry expert commented, "The intelligent agent of the future should not distinguish between an AI that drives and an AI that chats; they should all be super-intelligent agents based on a unified understanding of the physical world." Behind XPeng's organizational shake-up lies an ambitious strategic blueprint for Physical AI. Prior to this, He Xiaopeng had officially announced an upgrade to XPeng's positioning, defining the company as "an explorer of mobility in the physical AI world, a global embodied intelligence company." Currently, XPeng plans to utilize a single AI technology system to power different forms of robots—namely, cars, humanoid robots, and flying cars—and bring them to the global market. Based on its second-generation VLA large model, XPeng is building a fully in-house, full-stack Physical AI system covering chips, operating systems, and smart hardware to serve as the technical foundation for this strategy. The synergistic effects brought about by the departmental merger will be directly reflected at the product level. The intelligent experience will shift from a mere stacking of functions to a true fusion of capabilities. For instance, when the system detects driver fatigue, it will no longer be limited to a warning on the cockpit screen; it could also coordinate with the intelligent driving system to proactively reduce speed and adjust the route. When encountering road construction, as the intelligent driving system automatically navigates around it, the cockpit might simultaneously broadcast traffic information and could even adjust the air conditioning and music, providing a seamless and coherent intelligent experience. This deep integration is also evident in technology reuse. It is understood that the reuse rate of AI software between XPeng's robots and its cars is as high as 70%, with many key technologies, such as perception systems and domain controllers, being interchangeable. Li Hengguang, an analyst at Northeast Securities, believes that over the past decade, XPeng has completed a path reconstruction from a "new force in intelligent electric vehicles" to a "global AI-powered automotive intelligent technology company." He states that the core of this shift is not a simple expansion of a new energy vehicle manufacturer, but rather the construction of an embodied intelligence ecosystem centered on AI-defined vehicles, integrating cars, robots, and flying cars into a trinity. Accompanied by the dual drivers of extended-range technology and globalization, deep technical cooperation with Volkswagen, and a significant improvement in internal organizational efficiency, XPeng is transitioning from a phase of technological leadership under profit pressure into a channel leading to an inflection point for profitability, driven by mass-market hit products and technology exports. Currently, XPeng's AI-focused organizational adjustments are not an isolated case but rather a microcosm of a strategic pivot occurring across the entire smart car industry. Li Auto has also undertaken a similar R&D restructuring, splitting its former autonomous driving team and integrating personnel into three newly established teams: Foundation Model, Software Entity, and Hardware Entity. Tesla, meanwhile, has announced a $20 billion investment into Musk's X AI and plans to discontinue some vehicle models, repurposing production lines for the manufacture of its humanoid robot, Optimus. These moves reflect a shared strategic direction among new automakers: evolving from pure-play car manufacturers into embodied intelligence companies. The reason automakers are flocking to hardware intelligence domains like embodied intelligence lies in their supply chain advantages. Li Jingtao, an analyst at CITIC Securities, points out that the core supply chains for smart cars and humanoid robots have an overlap exceeding 60%, providing a natural pathway for cost reduction. The frequent organizational adjustments within the automotive industry are reshaping the competitive landscape of the smart car sector. The nature of competition is now escalating from contests over individual technologies to the construction of full-stack AI systems. Where automakers once competed on who had more intelligent driving features or larger cockpit screens, they now compete on who can make these intelligent modules truly work in synergy, forming a super-brain. For those automakers reliant on purchased technology and lacking full-stack in-house R&D capabilities, industry consolidation may arrive more swiftly and forcefully. The structure of the automotive industry remains in a state of flux.