Ex-Li Auto AI Lead Joins Robotics Firm
💡Auto AI talent fueling robotics boom with world models expertise
⚡ 30-Second TL;DR
What Changed
Xia Zhongpu joins Wujie Power post-Li Auto
Why It Matters
Signals talent exodus from auto AD to embodied AI, accelerating robotics with proven engineering. Ideal's ex-team gains traction via funding, intensifying competition in physical world modeling.
What To Do Next
Review Wujie Power's Shengshu Tech partnership for multimodal world model implementations.
Key Points
- •Xia Zhongpu joins Wujie Power post-Li Auto
- •Rejected offers from top carmakers after mid-2025 departure
- •Led Li Auto's end-to-end demo to production in 100 days
- •Expert in world models for AD and robotics generalization
🧠 Deep Insight
AI-generated analysis for this event — not the original article.
🔑 Enhanced Key Takeaways
- •Wujie Power, founded by former Huawei and ByteDance engineers, secured significant seed funding in late 2025 specifically to accelerate the development of a 'general-purpose' embodied AI brain.
- •Xia Zhongpu's transition marks a broader industry trend where top-tier autonomous driving talent is migrating from automotive OEMs to general-purpose robotics startups to leverage end-to-end model architectures in non-automotive environments.
- •The '100-day' project at Li Auto, which Xia spearheaded, utilized a proprietary vision-language-action (VLA) model architecture that significantly reduced the reliance on traditional modular perception-planning pipelines.
🛠️ Technical Deep Dive
- •End-to-end architecture: Replaces traditional perception, prediction, and planning modules with a unified neural network that maps raw sensor input directly to control commands.
- •World Model Integration: Utilizes latent space dynamics to simulate physical environment interactions, allowing the agent to predict future states before executing actions.
- •VLA (Vision-Language-Action) Models: Employs multi-modal transformers capable of processing visual tokens alongside natural language instructions to generalize tasks across diverse robotic embodiments.
🔮 Future ImplicationsAI analysis grounded in cited sources
⏳ Timeline
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: 36氪 ↗
This is a summary, not the original. Read the source, or get the weekly briefing.
Weekly AI briefing
One email a week. Unsubscribe anytime.

