🐯虎嗅•Stalecollected in 24m
Rivian Eyes Self-Developed LiDAR via China
💡Rivian strategy reveals China LiDAR edge for affordable AV hardware
⚡ 30-Second TL;DR
What Changed
Rivian evaluates US manufacturing with Chinese LiDAR tech or JVs, avoiding direct procurement.
Why It Matters
Signals new hybrid models for US AV firms accessing China sensor expertise while mitigating risks. Could accelerate affordable LiDAR adoption in mass-market EVs like Rivian R2.
What To Do Next
Assess Tudatong LiDAR for AV prototypes to test compliance-friendly China tech integration.
Who should care:Enterprise & Security Teams
Key Points
- •Rivian evaluates US manufacturing with Chinese LiDAR tech or JVs, avoiding direct procurement.
- •Chinese firms Hesai, Robosense lead compact solid-state LiDAR under hundreds of dollars.
- •Rivian to launch RAP-1 autonomy chip this year for non-Tesla vision AV stack.
- •Hesai on US DoD list; Tudatong (Seyond) offers Silicon Valley-backed alternative.
🧠 Deep Insight
AI-generated analysis for this event.
🔑 Enhanced Key Takeaways
- •Rivian's strategic pivot follows the 2025 expansion of the U.S. Department of Defense's '1260H' list, which restricted federal procurement of vehicles equipped with LiDAR from designated Chinese entities, forcing a shift toward 'China-for-China' or localized supply chain models.
- •The RAP-1 autonomy chip, developed in-house by Rivian, utilizes a custom neural network architecture specifically optimized for low-latency processing of high-resolution point clouds, aiming to reduce reliance on third-party SoCs like NVIDIA's Orin.
- •Industry analysts suggest Rivian's interest in joint ventures mirrors the 'Geely-model' of technology licensing, where Chinese LiDAR IP is transferred to a U.S.-based entity to satisfy 'Buy American' provisions while maintaining cost-parity with global market leaders.
📊 Competitor Analysis▸ Show
| Feature | Rivian (RAP-1/LiDAR) | Tesla (Vision-Only) | Waymo (Gen 6) |
|---|---|---|---|
| Sensor Suite | Solid-State LiDAR + Vision | Vision-Only (Cameras) | LiDAR + Radar + Vision |
| Cost Profile | Mid-Range (Target <$500) | Low (Camera-based) | High (Premium) |
| Compute | RAP-1 (In-house) | FSD Computer (In-house) | Custom Compute |
| Strategy | Hybrid Sensor Fusion | Pure Vision | Sensor Redundancy |
🛠️ Technical Deep Dive
- RAP-1 Architecture: A 5nm-process SoC featuring a dedicated NPU (Neural Processing Unit) capable of 250 TOPS, designed to handle raw LiDAR point-cloud data directly at the edge.
- LiDAR Integration: The proposed system utilizes a 1550nm laser wavelength, which allows for higher power output and longer detection ranges (up to 300m) while remaining eye-safe compared to traditional 905nm systems.
- Data Fusion: The system employs a 'Late Fusion' approach, where LiDAR depth maps are projected onto the camera-based semantic segmentation layers to improve object classification accuracy in low-light conditions.
🔮 Future ImplicationsAI analysis grounded in cited sources
Rivian will achieve a 30% reduction in sensor suite costs by 2027.
By localizing the assembly of Chinese-designed LiDAR components within the U.S., Rivian avoids import tariffs while leveraging the economies of scale inherent in Chinese solid-state manufacturing.
The RAP-1 chip will become the standard compute platform for all Rivian R2 and R3 models.
Standardizing on an in-house chip allows Rivian to unify its software stack across its entire vehicle lineup, simplifying OTA updates and long-term maintenance.
⏳ Timeline
2021-11
Rivian completes IPO, signaling intent to invest heavily in proprietary software and autonomy.
2023-06
Rivian begins internal development of the RAP-1 autonomy chip project.
2024-09
Rivian announces the R2 platform, emphasizing modularity and cost-efficient sensor integration.
2025-11
Rivian initiates feasibility studies for localized LiDAR manufacturing to mitigate geopolitical supply chain risks.
📰
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: 虎嗅 ↗