Nio Onvo L80 SUV Launches with Advanced AI Driving

See how Nio is deploying proprietary 5nm AI chips and world models in mass-market vehicles.
30-Second TL;DR
What Changed
Equipped with the 5nm Shenji NX9031 autonomous driving chip.
Why It Matters
The integration of NWM into mass-market EVs signals a shift toward model-based autonomous driving architectures in the competitive Chinese EV sector.
What To Do Next
Analyze the performance of Nio's NWM architecture compared to traditional CNN-based perception stacks in autonomous driving.
Key Points
- •Equipped with the 5nm Shenji NX9031 autonomous driving chip.
- •Integrates the Nio World Model (NWM) for advanced driving assistance.
- •Offers a 615km range with a high-efficiency 925V electric powertrain.
- •Features a 17.2-inch 3K immersive cockpit display.
Deep Insight
Background and context from public sources — not the original article. 21 sources cited.
Enhanced Key Takeaways
- •The Onvo L80 is a five-seat variant of the Onvo L90, sharing Nio's NT 3.0 platform and 900-volt high-voltage architecture.
- •The L80 offers a two-tier ADAS strategy: the entry-level Pro version utilizes Nvidia's Orin X chip for pure-vision assisted driving, while higher trims (Max+ and Ultra+) feature Nio's in-house Shenji NX9031 chip alongside high-precision LiDAR.
- •Nio's proprietary 5nm Shenji NX9031 chip contains over 50 billion transistors and offers computing power comparable to four Nvidia Drive Orin X chips.
- •The Nio World Model (NWM) is a multimodal autoregressive generative model that integrates visual, linguistic, spatial, and temporal data, enabling 'mapless' autonomous navigation and natural language interaction for complex scenarios like parking lots.
- •The Onvo L80 is aggressively priced to undercut the Tesla Model Y in China, starting at 242,800 RMB for outright purchase or 156,800 RMB with the Battery-as-a-Service (BaaS) option.
Competitor Analysis
- Nio Onvo L80 (Pro/Max+/Ultra+)
- 242,800 (full purchase), 156,800 (BaaS)
- Tesla Model Y (China Standard Range RWD/Long Range AWD)
- 263,500 (Standard RWD), 313,500 (Long Range AWD)
- Nio Onvo L80 (Pro/Max+/Ultra+)
- 5.7s (RWD), 4.5s (AWD)
- Tesla Model Y (China Standard Range RWD/Long Range AWD)
- 5.9s (Standard RWD), 4.3s (Long Range AWD)
- Nio Onvo L80 (Pro/Max+/Ultra+)
- 615 km (RWD), 570 km (AWD)
- Tesla Model Y (China Standard Range RWD/Long Range AWD)
- 593 km (Standard RWD), 750 km (Long Range AWD)
- Nio Onvo L80 (Pro/Max+/Ultra+)
- Nvidia Orin X (Pro), Shenji NX9031 with LiDAR (Max+/Ultra+)
- Tesla Model Y (China Standard Range RWD/Long Range AWD)
- Tesla HW4 (AI 4)
- Nio Onvo L80 (Pro/Max+/Ultra+)
- 85 kWh (standard)
- Tesla Model Y (China Standard Range RWD/Long Range AWD)
- 60 kWh LFP (Standard RWD), 78.4 kWh NMC (Long Range AWD)
- Nio Onvo L80 (Pro/Max+/Ultra+)
- 2,840
- Tesla Model Y (China Standard Range RWD/Long Range AWD)
- 2,130 (standard Model Y)
- Nio Onvo L80 (Pro/Max+/Ultra+)
- 900V
- Tesla Model Y (China Standard Range RWD/Long Range AWD)
- Not specified, typically 400V class
- Nio Onvo L80 (Pro/Max+/Ultra+)
- Yes
- Tesla Model Y (China Standard Range RWD/Long Range AWD)
- No
| Feature/Pricing/Benchmarks | Nio Onvo L80 (Pro/Max+/Ultra+) | Tesla Model Y (China Standard Range RWD/Long Range AWD) |
|---|---|---|
| Starting Price (RMB) | 242,800 (full purchase), 156,800 (BaaS) | 263,500 (Standard RWD), 313,500 (Long Range AWD) |
| 0-100 km/h Acceleration | 5.7s (RWD), 4.5s (AWD) | 5.9s (Standard RWD), 4.3s (Long Range AWD) |
| CLTC Range | 615 km (RWD), 570 km (AWD) | 593 km (Standard RWD), 750 km (Long Range AWD) |
| ADAS Chip | Nvidia Orin X (Pro), Shenji NX9031 with LiDAR (Max+/Ultra+) | Tesla HW4 (AI 4) |
| Battery Capacity | 85 kWh (standard) | 60 kWh LFP (Standard RWD), 78.4 kWh NMC (Long Range AWD) |
| Max Cargo Space (L) | 2,840 | 2,130 (standard Model Y) |
| High-Voltage Architecture | 900V | Not specified, typically 400V class |
| Battery Swapping | Yes | No |
Technical Deep Dive
- Shenji NX9031 Chip:
- Manufactured on a 5nm process technology by Samsung.
- Contains over 50 billion transistors.
- Features 32 general-purpose Arm cores (Cortex-A78AE and Cortex-A65AE in a big.LITTLE configuration), a neural processing unit (NPU), and a graphics processor (GPU).
- Utilizes LPDDR5X memory at 8.5 GHz with 204 GB/s bandwidth.
- Conforms to the ISO 26262 ASIL-D safety standard, incorporating Cortex-R52 safety cores.
- Offers computing power comparable to four Nvidia Drive Orin X chips, with some sources indicating over 1,000 TOPS or 615 kDMIPS.
- Designed to process data from LiDAR sensors, distinguishing it from vision-only systems.
- Nio World Model (NWM):
- A multimodal autoregressive generative model for smart driving.
- Integrates visual, linguistic, spatial, and temporal data for real-time reasoning and decision-making.
- Employs a closed-loop reinforcement learning framework for continuous self-learning and adaptation.
- Enables "mapless" autonomous navigation in complex environments, such as underground parking lots, by interpreting signage and natural language commands.
- Can simulate 216 potential scenarios within 100 milliseconds to identify optimal driving decisions.
- Training leverages massive human driving data, moving beyond costly expert data.
- The NWM 2.0 update, rolled out in January 2026, introduced pilot-driving among battery swap stations in urban areas, integrating with urban piloted driving.
Future ImplicationsAI analysis grounded in cited sources
Timeline
- 2021Nio initiated in-house chip development efforts.
- 2023-12-23Nio unveiled its Shenji NX9031 autonomous driving chip at Nio Day 2023.
- 2024-05Nio formally introduced its mass-market Onvo sub-brand.
- 2024-09-19Onvo L60, the brand's first model, went on sale in China.
- 2025-05-29Nio announced the rollout of the first version of its Nio World Model (NWM) assisted driving system starting May 30.
- 2026-01-28Nio rolled out the NWM 2.0 update to over 460,000 vehicles in China.
- 2026-05-15Nio's Onvo L80 SUV officially launched in China.
Sources (21)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
Weekly AI Recap
Read this week's curated digest of top AI events →
AI-curated news aggregator. All content rights belong to original publishers.
Original source: IT之家 ↗
This is a summary, not the original. Read the source, or get the weekly briefing.
The weekly digest
One email a week. Unsubscribe anytime.
