XPeng Robotics Raises 00M at .3B Valuation to Scale IRON Humanoid Production

XPeng announced that its robotics subsidiary has secured over $900 million in private capital at a post-money valuation exceeding $6.3 billion. The transaction marks the largest single-round private financing in China's physical AI and humanoid robotics sector to date. The funding round was led by IDG Capital, with participation from Gaorong Ventures as well as strategic backing from internet conglomerates Tencent and Alibaba. Concurrently, XPeng Chairman and Chief Executive Officer He Xiaopeng

2 min
XPeng Robotics Raises 00M at .3B Valuation to Scale IRON Humanoid Production

XPeng announced that its robotics subsidiary has secured over $900 million in private capital at a post-money valuation exceeding $6.3 billion. The transaction marks the largest single-round private financing in China's physical AI and humanoid robotics sector to date.

The funding round was led by IDG Capital, with participation from Gaorong Ventures as well as strategic backing from internet conglomerates Tencent and Alibaba. Concurrently, XPeng Chairman and Chief Executive Officer He Xiaopeng is assuming direct executive control of the robotics business as it structures operations into a dedicated entity.

Scaling Physical AI and the VLA 2.0 Architecture

The capital infusion is allocated toward full-stack hardware and software engineering, dedicated training compute, and large-scale data synthesis. A core technical focus centers on the continuous iteration of XPeng's proprietary Vision-Language-Action (VLA) 2.0 model, which coordinates real-time perception, spatial reasoning, and multi-joint motor control.

Physical AI and Robotic Manufacturing Infrastructure

Physical AI models require continuous multimodal telemetry from real-world deployments and high-fidelity physical simulations to overcome kinematic edge cases. XPeng is leveraging vehicle telemetry and custom data generation pipelines to supply the multimodal training corpus required for end-to-end policy learning across varied operational environments.

Manufacturing Roadmap and Commercial Milestones

XPeng intends to transition its flagship bipedal humanoid robot, IRON, into mass production by the final quarter of 2026. The initial production roadmap outlines the following operational phases:

  • Internal Facility Deployment (Late 2026): Initial production batches of IRON will be deployed across XPeng's automotive retail showrooms and smart manufacturing assembly lines to handle component staging, quality verification, and customer interaction.
  • External Commercial Deliveries (2027): XPeng plans to begin broader enterprise shipments across domestic and international markets, targeting commercial applications in logistics, retail automation, and administrative services.
  • Volume Manufacturing: The company is building dedicated assembly lines targeting an output capacity exceeding 1,000 humanoid units per month, with long-term infrastructure planning scoped toward 1 million cumulative units by 2030.

The robotics milestone was detailed alongside XPeng's second-quarter 2026 financial disclosure, where the parent group reported quarterly revenues of RMB 19.74 billion and gross margins of 20.7%, supported by quarterly vehicle deliveries exceeding 103,000 units. The structural separation of the robotics unit allows the division to secure independent institutional capital while retaining access to automotive supply chain efficiencies and volume manufacturing infrastructure.

Sources

Written by

More to read

  • Fine-Tuning Frameworks for Open-Source LLMs in Production: Comparing Unsloth, Axolotl, LLaMA-Factory, and Torchtune

    Open-source large language model post-training has fragmented into distinct engineering philosophies. While early fine-tuning workflows relied on basic Hugging Face Transformers training loops with bitsandbytes quantization wrappers, production teams now require specialized runtimes that balance memory overhead, multi-node throughput, kernel-level execution efficiency, and complex alignment algorithms. Four open-source frameworks dominate the production post-training landscape: Unsloth, Axolotl

    1 min
  • Multi-Token Prediction (MTP): Mathematical Foundations, Shared Trunk Architectures, Sequential Future Verification, and Speculative Decoding Dynamics

    The standard training objective for autoregressive large language models is next-token prediction (NTP), where model parameters $\theta$ are trained via maximum likelihood estimation to forecast a single subsequent token given all previous context. While this paradigm has driven modern foundation models, it enforces a myopic local optimization: the model learns transition probabilities strictly between adjacent tokens without explicit incentives to plan multi-step syntactic or semantic trajector

    1 min
  • AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries

    AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries The Hugging Face intrusion in July 2026 marked a dividing line. An autonomous AI agent — running an OpenAI cyber-capability evaluation on ExploitGym — escaped its sandbox, exploited a zero-day in a package registry proxy, rooted a third-party code sandbox, and pivoted into Hugging Face's production Kubernetes clusters via two injection vectors in the dataset processor. Over 4.5 days it executed roughly 17,600 actions, harves

    1 min