XPENG’s physical AI unit has secured more than $900 million at a $6.3 billion valuation to expand its IRON humanoid robot platform. The Chinese electric-vehicle manufacturer says the financing is the largest single private round yet recorded in China’s physical AI industry, measured by both the money raised and the resulting valuation. That description is XPENG’s own claim.
IDG Capital led the round, with Gaorong Ventures participating alongside strategic investors Tencent and Alibaba. XPENG will retain a controlling stake after the transaction closes, and the robotics unit will remain consolidated in the group’s financial statements. The structure gives the business external capital without separating it from its parent.
Funding targets the entire robotics stack
XPENG says the money will support long-term development across software, hardware and physical AI models. Planned spending includes model training and iteration, generation of high-quality data, and construction of end-to-end mass-production facilities. Part of the funding is also intended for expansion outside China and incentive arrangements for senior executives and other robotics employees.
The plan reflects XPENG’s emphasis on full-stack development. Chairman and CEO He Xiaopeng says the company has spent 12 years building capabilities spanning a physical-world foundation model, Turing AI chips and AI infrastructure. According to the company, that base now enables it to move toward mass production and commercial deployment of humanoid machines.
The financing arrives against a more difficult picture for the parent company’s listed shares. AI News reported that XPENG’s stock was down 7.22% at the time of publication and had declined 51.24% over 12 months. Its vehicle business faces intensified competition from Chinese manufacturers and Tesla. Investors are therefore assigning a $6.3 billion valuation to the robotics operation while the market value of XPENG’s core business remains under pressure.
IRON combines in-house hardware and local AI
IRON uses a fully enclosed flexible lattice structure designed by XPENG to balance a human-like appearance with safety. The company lists 76 degrees of freedom across the robot’s body and 21 in each hand. Those specifications indicate a design aimed at dexterous movement, although they do not by themselves establish reliability in daily commercial tasks.
XPENG developed the robot’s chips, controllers, motion modules and hand mechanisms internally. It says manufacturing will borrow quality standards and production processes from its electric-vehicle operations, with the goal of reaching automotive-scale output. The company’s experience in supply chains and high-volume production is central to the case made by participating investors, but the target scale has not yet been achieved.
Three in-house Turing AI chips provide up to 2,250 TOPS of effective computing power, according to XPENG. The robot runs the company’s physical AI foundation model on the device, an arrangement intended to lower inference latency and keep data processing local. XPENG argues that a human-like body can also collect behavior data from environments and tools designed for people, creating a cycle in which deployment generates data for further model improvements.
Mass production is promised before broad delivery
XPENG expects IRON to enter mass production by the end of 2026. The initial deployment is planned for the company’s stores and campuses, giving it controlled environments in which to introduce the system. Customer deliveries in China and overseas markets are scheduled to begin during 2027.
Investors framed the timing as a transition from technical demonstrations toward scalable manufacturing and practical value. IDG Capital highlighted the connection between XPENG’s edge processors, foundation models, complete robot systems and its existing vehicle and autonomous-driving work. Gaorong Ventures similarly pointed to automotive-grade safety, quality controls and manufacturing experience.
Those statements explain why the round attracted major financial and strategic backers, but they remain projections about execution. The financing does not prove that IRON can be produced at the planned volume, perform consistently outside XPENG facilities or win customers internationally. The next meaningful evidence will come from the promised production ramp and the limited internal deployments that precede broader sales.
