Xpeng bets on "Physical AI"

At the start of 2026, XPeng Motors’ organizational transformation is more intense than in previous years.

On February 3rd, Wall Street Insights learned that XPeng officially merged its Autonomous Driving Center with its Smart Cabin Center into a strategic partnership, creating a new department named — the General Intelligence Center. XPeng told Wall Street Insights: “This adjustment is confirmed.”

The two departments were originally parallel first-level units. After the adjustment, Liu Xianming, the former head of autonomous driving, was appointed as the number one position at the center, reporting directly to Chairman and CEO He Xiaopeng.

From an external perspective, this might seem like a simple resource integration, but industry insiders see it as a sign that XPeng Motors is officially ending the past decade of dual-track R&D for intelligent driving and intelligent cabins, shifting fully to a single-track era centered on physical AI. This is also the current direction many automakers are pursuing.

For a long time, the hardware and software architecture of intelligent vehicles has been divided into two high grounds: intelligent driving and intelligent cabins. In XPeng’s R&D sequence, the Autonomous Driving Center is responsible for “hands, feet, and eyes,” while the Smart Cabin Center handles “mouth and ears.” Although this division ensured business specialization in the early days, with the proliferation of end-to-end large models, the boundaries between the two are becoming blurred.

Under traditional distributed architecture, intelligent driving and intelligent cabins each have their own perception algorithms, computing platforms, and data links. This means that when vehicles need to perform cross-domain functions, data must repeatedly jump between different departments and systems. This organizational wall creates technical silos, increasing R&D costs and limiting system response speed.

With the establishment of the General Intelligence Center, XPeng will reorganize its secondary structure around foundational models (Foundation Models), infrastructure, and platform delivery.

This means that whether it’s the decision-making flow for autonomous driving or voice/visual interaction in the smart cabin, they will share the same AI infrastructure at the core. Industry insiders believe: “Future intelligent agents shouldn’t distinguish between AI for driving and AI for chatting; they should all be super-intelligent entities based on a unified understanding of the physical world.”

Behind this organizational change is a grand strategic blueprint for physical AI. Previously, He Xiaopeng officially announced XPeng’s positioning as “a travel explorer in the physical AI world, a globally oriented embodied intelligence company.”

Currently, XPeng plans to use the same AI technology system to drive different forms of robots, such as cars, humanoid robots, and flying cars, bringing them to global markets. Based on the second-generation VLA large model, XPeng is building a full-stack self-developed physical AI system covering chips, operating systems, and smart hardware to support this strategy.

The synergy effects from the department merger will be directly reflected at the product level. Intelligent experience will shift from function stacking to capability integration.

For example, when the system detects driver fatigue, it will no longer just remind on the cabin screen but also actively reduce speed and adjust the driving route through the autonomous driving system. When encountering road construction, the autonomous system will automatically avoid obstacles, while the cabin may broadcast traffic information simultaneously, even adjusting the air conditioning and music to provide a seamless intelligent experience.

This deep collaboration is also reflected in technology reuse. It is understood that XPeng’s robots and cars share up to 70% of AI software, with many key technologies such as perception systems and domain controllers being interconnected.

Li Hengguang, an analyst at Northeast Securities, believes that XPeng Motors has completed a path reorganization over the past decade from a “new force in smart electric vehicles” to a “globally oriented AI automotive intelligent technology enterprise.” Its core is not simply expanding as a new energy vehicle manufacturer but building an integrated intelligent ecosystem of cars, robots, and flying cars centered around AI. With dual-wheel drive of extended range and globalization, deep technical cooperation with Volkswagen, and significant internal organizational efficiency improvements, XPeng is transitioning from a stage of technological leadership with profit pressure to a profit turnaround driven by mass-market hits and technology output.

Currently, XPeng’s organizational adjustments around AI are not isolated but reflect the broader strategic shift in the entire smart automotive industry.

Li Auto has also made similar adjustments, splitting its autonomous driving team into three new teams focused on foundational models, software core, and hardware core. Tesla announced an investment of $2 billion in Musk’s X AI and plans to halt production of some models, redirecting production lines to manufacture humanoid robots like Optimus.

These moves reflect a common strategic direction among new car-making forces: evolving from mere automakers to embodied intelligence companies.

The reason automakers are rushing into embodied intelligence hardware fields is due to supply chain advantages. Citic Securities analyst Li Jingtao believes that the core supply chain overlap between smart cars and humanoid robots exceeds 60%, providing a natural path for cost reduction.

Frequent organizational adjustments in the automotive industry are reshaping the competitive landscape of smart vehicles.

Currently, industry competition is shifting from a focus on individual technological prowess to the construction of full-stack AI systems. Previously, automakers competed over who had more advanced autonomous driving features and larger cabin screens; now, the focus is on who can truly enable these smart modules to work together, forming a “super brain.”

For automakers that rely on external technology procurement and lack full-stack self-development capabilities, industry reshuffling may be even more intense. The landscape of the automotive industry remains in flux.

Risk Warning and Disclaimer

Market risks exist; investments should be cautious. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article are suitable for their particular circumstances. Invest accordingly at your own risk.