AE-NJ60BT: NVIDIA Jetson Orin NX Industrial Edge AI Computer (157 TOPS)
Fanless industrial Edge AI platform with 4× GbE, 4× GMSL2 camera ingress, 2× CAN FD and DC 9–36V wide input — engineered for autonomous vehicles, AMR/AGV, smart factory and energy AI workloads.
AE-NJ60BT — fanless extruded aluminum unibody with quad GbE and rich industrial I/O.
Product Overview
The AE-NJ60BT is a high-performance industrial Edge AI computer built around the NVIDIA Jetson Orin NX 8GB / 16GB module, delivering scalable AI compute from 70 TOPS up to 157 TOPS (Super mode). Combined with 1024 CUDA cores, 32 Tensor cores, 2× NVDLA v2 and an NVMe storage path, it provides a turn-key platform for production-grade vision, robotics and generative-AI workloads at the edge.
Engineered with a cable-less, jumper-less unibody extruded-aluminum chassis, 4× Gigabit Ethernet(1× SOM + 3× Intel i210-AT), 4-lane MIPI CSI / 4× GMSL2 camera ingress, 2× CAN FD, opto-isolated 8 DI / 8 DO and a DC 9–36V wide-range input, the AE-NJ60BT is JetPack 5.1.2 / 6.0 / 6.2 compatible and rated for -40°C to +70°C deployment.
AE-NJ60BT Workstation-Class Embedded Architecture for Autonomous Infrastructure & Vision AI
157 TOPS NVIDIA Ampere Pipeline with 1024 CUDA Cores
Configurable across Jetson Orin NX 8GB (70/117 TOPS Super) and Orin NX 16GB (100/157 TOPS Super) — 1024 CUDA cores, 32 Tensor cores and 2× NVDLA v2 sustain transformer vision and on-device LLM workloads at the edge.
Multi-Channel Vision Ingress: 4-Lane MIPI CSI & 4× GMSL2 Topography
Adapter board exposes 4-lane MIPI CSI or 4× GMSL2 for stable long-distance HD camera transmission — feeding all four streams directly into the Jetson ISP and 1× 8K30 / 4× 4K30 hardware decoder pipeline.
Independent Quad Gigabit Ethernet Driven by Intel i210-AT MACs
Four independent GbE ports — one from the Jetson SOM and three driven by industrial-grade Intel i210-AT controllers — enable multi-subnet isolation for cameras, PLCs and cloud uplink, with hardware-level network redundancy.
Fail-Safe Power Topography (DC 9-36V Vehicle-Grade Input)
Direct compatibility with 12V / 24V vehicle, AGV and AMR battery systems — no external regulator required. 100W minimum supply with auto power-on and button startup options.
Cableless Unibody Chassis with Dynamic Thermal Management
All internal power and signal connections are board-routed — no internal harnesses to vibrate loose. Unibody extruded-aluminum chassis doubles as the primary heatsink for fanless 100W operation.
4000V Opto-Isolated Industrial I/O & Dual CAN FD Interfaces
8× DI and 8× DO with opto-isolation (selectable 3.3V / 5V / 12V), 2× CAN FD, RS232/RS485, SPI, I²C and 4× GPIO — protect the mainboard from EMI while interfacing directly with sensors, actuators and field bus devices.
AE-NJ60BT Targeted Field Deployment Scenarios for Jetson Architecture
Multi-Camera Autonomous Vision AI & High-Speed Inline AOI Inspection
Aggregate 4× GMSL2 long-distance camera streams into TensorRT-optimized detection / segmentation pipelines for inline AOI, surface inspection and high-speed quality grading.
Autonomous Mobile Robots (AMR / AGV) Navigation & Real-Time SLAM
DC 9–36V wide input and cable-less design tolerate vehicle bus voltage swings and vibration; 157 TOPS + 4× GbE drive SLAM, sensor fusion and Nav2 stacks with sub-100ms perception loops.
Industry 4.0 Smart Factory Edge Gateways & Legacy PLC Bridging
Quad GbE (1× SOM + 3× Intel i210-AT) isolates camera, PLC and cloud subnets; 2× CAN FD and opto-isolated DI/DO bridge field devices to MES/SCADA without congestion.
Energy Infrastructure Monitoring, Substation AI Safety & ESS Site Supervision
Wide -40°C to +70°C envelope, fanless unibody thermals and TPM/TCM secure boot suit unattended substations, PV / wind farms and ESS sites where firmware integrity and 24/7 uptime are required.
Smart City IoT Aggregation, Multi-Protocol Wireless Routing & Edge Sensing
Triple wireless lanes (Wi-Fi 6 + 4G/5G + LoRa) with Nano SIM + eSIM let one unit serve as cellular uplink, LoRa concentrator and local AP — running AI preprocessing before forwarding events upstream.
Core Compute, Memory Subsystem & JetPack Compliance
Configure with Jetson Orin NX 8GB (70/117 TOPS Super) or Orin NX 16GB (100/157 TOPS Super) on the same carrier board — prototype and migrate to production without redesigning integration. The flagship 16GB Super configuration delivers 157 TOPS via 1024 CUDA cores, 32 Tensor cores and 2× NVDLA v2 accelerators.
| AI Module (8GB) | NVIDIA Jetson Orin NX 8GB — 6-core A78 @ 2.0GHz, 70 TOPS / 117 TOPS (Super) |
| AI Module (16GB) | NVIDIA Jetson Orin NX 16GB — 8-core A78 @ 2.2GHz, 100 TOPS / 157 TOPS (Super) |
| GPU | NVIDIA Ampere — 1024 CUDA cores + 32 Tensor cores @ 1173MHz |
| NPU / DLA | 2× NVDLA v2 |
| JetPack | Compatible with JetPack 5.1.2 / 6.0 / 6.2 |
| Operating System | Ubuntu 20.04 (default) |
High-Bandwidth Networking & Wireless Expansion Slots
Four independent Gigabit Ethernet ports — 1× from the SOM and 3× driven by industrial-grade Intel® i210-AT controllers — enable multi-subnet isolation across cameras, PLCs and cloud uplink. Triple wireless lanes (Wi-Fi 6, 4G/5G, LoRa) with Nano SIM + eSIM cover any field deployment.
| Ethernet (SOM) | 1× GbE LAN from SOM (10/100/1000 Mbps) |
| Ethernet (Industrial) | 3× Intel® i210-AT GbE (10/100/1000 Mbps) |
| Wi-Fi / Bluetooth | 1× M.2 KEY-E (PCIe ×1 + USB 2.0, 2230) for Wi-Fi/BT |
| Cellular | 1× M.2 KEY-B (USB 3.0) for 4G/5G modules |
| LoRa | 1× Mini PCIe slot for LoRa modules |
| SIM | 1× Nano SIM + 1× eSIM |
Industrial Fieldbus, Serial Ports & Opto-Isolated DIO
| USB | 4× USB 3.0 Type-A, 4× USB 2.0 Type-A |
| Type-C | 1× Type-C OTG_USB, 1× Type-C Debug |
| Serial Ports | 2× RS232 or 2× RS485, 2× RS485, 1× UART (3.3V level) |
| CAN Bus | 2× CAN FD |
| GPIO | 4× IO, 3.3V level (input/output) |
| Digital I/O | 8× DI + 8× DO with opto-isolation (3.3V / 5V / 12V selectable) |
| SPI / I²C | 1× SPI, 1× I²C (3.3V level) |
| Audio | 1× Dual-channel speaker out (18W 4Ω), 1× headphone (L/R), 2× microphone in |
Display Output, GMSL2 Camera Capture & Video Codecs
| HDMI Output | 1× HDMI Type-A (max 8K @ 30Hz) |
| Camera Input | 4-lane MIPI CSI or 4× GMSL2 (with adapter board) |
| Video Decoding | 1× 8K30, 2× 4K60, 4× 4K30, 9× 1080P, 18× 1080P30 (H.265) |
| Video Encoding | 1× 4K60, 3× 4K30, 6× 1080P30, 18× 1080P30 (H.265) |
System Storage, RTC Configuration & Software Environment
| Storage Slot | 1× M.2 Key-M (PCIe ×4, NVMe SSD, 2242/2260/2280) |
| Storage Module | 1× M.2 2280 NVMe SSD |
| RTC | External RTC backup battery, supports timed power on/off |
| Operating System | Ubuntu 20.04 (BSP + kernel sources provided) |
AE-NJ60BT vs AE-3588BT choosing the Right Edge AI Power
NVIDIA Jetson Orin (157 TOPS) vs Rockchip RK3588J (scalable to 58 TOPS) — pick the right Edge AI class.
Looking for an alternative to Rockchip? AE-NJ60BT offers 2.7x higher AI performance via the NVIDIA Ampere architecture.
| Engineering Factor | AE-NJ60BT | AE-3588BT |
|---|---|---|
| SOM / Processor | NVIDIA Jetson Orin Nano / NX / NX Super | Rockchip RK3588J (4× A76 + 4× A55) |
| AI Compute | Up to 157 TOPS | 6 TOPS NPU, scalable to 58 TOPS via 2× M.2 |
| Camera Ingress | 4-lane MIPI CSI or 4× GMSL2 | MIPI CSI + USB cameras |
| Ethernet | 4× GbE (1× SOM + 3× Intel i210-AT) | 3× GbE |
| CAN / DIO | 2× CAN FD, 8 DI / 8 DO (isolated) | 2× CAN, 4 DI / 4 DO (isolated) |
| Software Stack | NVIDIA JetPack, CUDA, TensorRT | Linux / Android, RKNN, ROS2 |
| Operating Temp | -40°C to +70°C, fanless | -40°C to +80°C, fanless |
| Best Fit | Multi-camera AOI, AMR perception, autonomous driving | Edge gateway, control node, light vision |
Choose this if you need: GMSL2 for long-distance video, 100+ TOPS for transformer models, or native JetPack support.
Benchmarking Matric: Bitech ARM Jetson AE-NJ60BT vs x86 Workstation AE-760EBT
ARM-based dedicated AI inference (157 TOPS) vs x86 workstation-class edge compute (RTX MXM). See which architecture fits your workload.
AE-NJ60BT — ARM AI Inference Node
- NVIDIA Jetson Orin NX up to 157 TOPS
- 4× GMSL2 / MIPI CSI camera ingress
- DC 9–36V, -40°C to +70°C,fanless chassis (optional dynamic internal smart fan for absolute extremes)
- JetPack / CUDA / TensorRT ecosystem
AE-760EBT — x86 Edge Workstation
- Intel i9-13900 + RTX 4060–5090 MXM GPU
- 4× 2.5GbE GigE Vision + USB 3.2
- DC 9–36V, -10°C to +55°C, dual 80 mm fans
- Windows 11 Pro / Ubuntu + CUDA / OpenVINO
Hardened Manufacturing Quality & Supply Chain Compliance
Strict 100% 18-Hour Power-On Full-Load Burn-In Testing
Every unit undergoes an 18-hour full-load power-on burn-in with AQL inspection and per-unit serial number traceability.
Cryptographic Security Layer: Hardware TPM 2.0 / TCM Ready
Optional TPM/TCM encryption module supports secure boot, signed OTA updates and encrypted model storage for fielded fleets.
10 Year Product Lifecycle Commitment & Revision Control
Long-term availability for mainboard, modules, shells and ICs — eliminating EOL supply risk on multi-year deployments.
Lifetime OS Customization & Embedded BSP Engineering Support
OS support, BIOS/firmware updates and CAN/RS485/PLC debugging assistance for the entire product lifetime.
Alternate Industrial Computing Platforms for Deployment Comparison
Compare adjacent models to validate your selection.
AE-3588BT
Best for: Scalable ARM Edge AI with 4000V isolation
RK3588J + dual NPU expansion up to 58 TOPS
AE-3588LBT
Best for: Cabinet-level AI with low-latency I/O
Non-isolated I/O for controlled environments
AE-NJ60BT Technical FAQ for Hardware Integration Evaluation
How does the 157 TOPS "Super Mode" achieve peak execution latency compared to standard NVIDIA Orin NX 16GB profiles?
What is the exact hardware serialization latency across the 4× GMSL2 camera ingress pipeline?
How do we perform physical layer impedance matching and diagnostics for the dual CAN FD interfaces?
The AE-NJ60BT implements dual native CAN FD networks conforming strictly to the ISO 11898-2 standard, supporting data bit rates up to 5Mbps. To verify the network integrity and eliminate transmission reflections across your vehicle or AGV bus:
- Completely isolate the system from primary power to avoid transceiver leakage current from skewing measurements.
- Apply a digital multimeter across the CAN_H and CAN_L terminals on the Phoenix plug.
Is the thermal management purely passive, or does the system utilize an active cooling fan?
How does the independent network topography of the quad Gigabit Ethernet ports eliminate cross-talk congestion?
What are the electrical protection boundaries and sink/source rules for the 16-channel isolated DIO?
How does the system prevent NVMe data corruption during sudden vehicle or factory power drops?
Can Wi-Fi 6, 4G/5G cellular, and LoRa modules be deployed simultaneously without RF cross-talk or trace interference?
How are secure boot, signed Over-The-Air (OTA) updates, and core AI model encryption managed?
What is the exact Board Support Package (BSP) baseline version, and are upstream Linux kernel sources provided?
Evaluate the AE-NJ60BT
Send us your workload and camera topology — an engineer scopes the rest.
Jetson Orin NX 8GB or 16GB Super? GMSL2 or MIPI CSI? Vehicle power, sealed enclosure, wide-temp? Tell us the deployment and we return a configuration, thermal/power budget, and sample terms — not a sales brochure.
- Orin NX module + carrier configuration matched to your AI model
- Camera ingress plan: 4× GMSL2 long-distance or 4-lane MIPI CSI
- Sample units and unlocked BSP for field evaluation
Talk to engineering
Configuration review, customization, and sample requests for the AE-NJ60BT.
Discuss your configuration Download datasheetMOQ & volume pricing on request
Sample units available for evaluation