AEC-6100-32
The AEC-6100-32 is a fanless embedded edge AI computer built on the NVIDIA® Jetson AGX Orin™ system-on-module, combining an 8-core Arm® Cortex-A78AE CPU with a 1,792-core NVIDIA Ampere GPU, 56 Tensor Cores and dual NVDLA deep-learning accelerator engines. With 32 GB of LPDDR5 unified memory and a fanless thermal design rated from −25 °C to +70 °C, the AEC-6100-32 delivers datacenter-class AI inferencing in a compact, vibration-resistant chassis built for continuous deployment in transport, automotive and industrial environments.
Engineered for applications where camera density, ruggedised connectivity and vehicle integration are non-negotiable, the AEC-6100-32 supports eight simultaneous GMSL automotive cameras via Fakra-Z connectors alongside isolated CAN FD, RS-232/422/485 serial ports and 8 DI / 8 DO GPIO. A wide-range 9–48 V DC input with ignition control enables seamless integration into 12 V and 24 V vehicle and rolling-stock electrical systems, whilst a PCIe Gen4 x8 expansion slot and M.2 sockets for 5G and Wi-Fi modules allow the platform to grow with application requirements.
Key Features
- NVIDIA Jetson AGX Orin Compute Platform: 8-core Arm Cortex-A78AE v8.2 CPU, 1,792-core NVIDIA Ampere GPU with 56 Tensor Cores and 2× NVDLA engines; 32 GB LPDDR5 unified memory supports large inference models and high-throughput vision pipelines without offloading to the cloud
- 8× GMSL Automotive Camera Inputs: Eight Fakra-Z connectors for GMSL cameras deliver synchronised multi-camera perception over shielded coaxial links — no frame grabber required; HDMI 2.1 output supports display resolutions up to 8K@60Hz for operator consoles and HMI integration
- Rich Industrial I/O on a Single Platform: 2× RS-232/422/485 serial (DB9), 2× isolated CAN FD, 8 DI / 8 DO isolated GPIO, 5× USB 3.2 and 2× Micro-USB cover PLC integration, automotive sensor buses, fieldbus communication and USB3 Vision camera arrays in a single unit
- Fanless Design for Harsh Environments: Passive cooling eliminates moving parts and ingress points; operating range −25 °C to +70 °C (TDP 30 W) and −25 °C to +55 °C (TDP 40/50 W); shock and vibration qualified to MIL-STD-810G; humidity 5–95 % (non-condensing)
- Wide-Voltage Power with Vehicle Ignition Support: DC 9–48 V input via terminal block covers 12 V, 24 V and 48 V vehicle and industrial supply architectures; ignition (IGN) control enables clean power sequencing for in-vehicle and rolling-stock deployments
- Flexible Mounting and Connectivity Expansion: Wall, DIN rail and VESA mounting options; 1× PCIe Gen4 x8 expansion slot; M.2 Key-B for 4G/5G LTE; M.2 Key-E for Wi-Fi; 2× SIM sockets; 6× antenna openings for wireless modules — adaptable to fixed and mobile infrastructure
- Industrial Certifications: CE, FCC, EN 50155, EN 50121-3-2 — qualifies the AEC-6100-32 for railway rolling-stock, trackside and road-transport EMC compliance programmes
Application Use Cases
Autonomous Vehicles and Advanced Driver Assistance Systems
The AEC-6100-32 is purpose-built for ADAS and autonomous vehicle perception stacks. Eight GMSL camera inputs provide synchronised surround-view coverage, dual isolated CAN FD handles vehicle network communication at up to 5 Mbit/s, and ACC/IGN control ensures reliable power sequencing from engine start to shutdown. The 9–48 V input range accommodates both 12 V passenger vehicles and 24 V commercial trucks without additional power conversion hardware.
Intelligent Traffic Systems and Roadside AI
For roadside and overhead gantry installations, the eight GMSL connectors feed real-time video streams from traffic cameras directly to the Jetson AGX Orin inference engine, enabling licence plate recognition, vehicle classification and incident detection without a separate host PC. The HDMI 2.1 output drives operator displays or digital signage at up to 8K resolution, and the wide operating temperature range covers outdoor deployments across all European climate zones.
Railway and Public Transport
EN 50155 and EN 50121-3-2 certification positions the AEC-6100-32 for on-board rolling-stock and trackside applications. DIN rail mounting fits standard railway cabinet formats, and the MIL-STD-810G shock and vibration qualification satisfies the mechanical stress requirements of mainline and metro rolling stock. The fanless enclosure eliminates filter maintenance cycles and meets cleanliness requirements for enclosed passenger compartment deployments.
Industrial Machine Vision and Automation
In factory environments, the AEC-6100-32 connects up to five USB 3.2 cameras for quality inspection, combined with the PCIe Gen4 x8 slot for frame grabber cards or additional GPU accelerators. Isolated RS-422/485 and GPIO interfaces integrate directly with PLCs and motion controllers over standard fieldbus wiring, and the DIN rail mounting option fits directly into control cabinet infrastructure without custom brackets.
Edge AI Inference for Industrial IoT and Predictive Analytics
The dual NVDLA deep-learning accelerators run inference workloads in parallel with the Ampere GPU, allowing the AEC-6100-32 to execute multiple models simultaneously — anomaly detection, optical character recognition and thermal imaging classification — on the 32 GB unified memory pool. A 5G cellular module in the M.2 Key-B slot delivers low-latency telemetry to cloud dashboards without exposing raw sensor data over public networks.
Why Choose the AEC-6100-32
Jetson AGX Orin AI Performance Without Active Cooling
Combining a fanless thermal design with the NVIDIA Jetson AGX Orin's 1,792-core Ampere GPU and dual NVDLA engines, the AEC-6100-32 is one of the few platforms in its class that sustains high-throughput AI inference across the full −25 °C to +70 °C operating range without relying on fan-based cooling. This makes it directly deployable in environments where dust ingress, vibration or maintenance access constraints rule out fan-cooled alternatives — from roadside enclosures to rolling-stock computer cabinets.
Native GMSL Camera Ecosystem Integration
Unlike embedded AI platforms that require third-party frame grabbers or camera serialiser boards, the AEC-6100-32 integrates eight GMSL Fakra-Z connectors directly on the chassis. This allows systems integrators to connect automotive-grade GMSL cameras over standard coaxial cabling at distances up to 15 metres per channel, reducing system BOM cost, connector count and potential failure points in multi-camera perception arrays for ADAS, traffic and surveillance programmes.
Cross-Market Deployment from a Single Qualified Platform
Wall mount, DIN rail and VESA mounting options, a 9–48 V DC wide-range input, IGN control and industrial certifications covering CE, FCC, EN 50155 and MIL-STD-810G allow a single AEC-6100-32 qualification to span automotive, railway, smart city and factory automation projects. Engineers reduce re-qualification effort and procurement teams consolidate to a single SKU across programmes — a material operational advantage for organisations deploying across multiple market verticals simultaneously.
Frequently Asked Questions
What processor does the AEC-6100-32 use?
The AEC-6100-32 is powered by the NVIDIA Jetson AGX Orin system-on-module, which integrates an 8-core Arm Cortex-A78AE v8.2 64-bit CPU, a 1,792-core NVIDIA Ampere GPU with 56 Tensor Cores, dual NVDLA deep-learning accelerator engines, and 32 GB of LPDDR5 unified memory.
How many cameras does the AEC-6100-32 support?
The AEC-6100-32 supports up to eight GMSL automotive cameras simultaneously via Fakra-Z coaxial connectors. This makes it suitable for surround-view ADAS perception, multi-lane traffic monitoring and wide-area surveillance applications without additional frame grabber hardware.
What is the operating temperature range of the AEC-6100-32?
The AEC-6100-32 operates from −25 °C to +70 °C in fanless 30 W TDP mode, and from −25 °C to +55 °C at 40/50 W TDP. Storage temperature range is −40 °C to +85 °C. The unit meets MIL-STD-810G shock and vibration requirements.
Is the AEC-6100-32 suitable for vehicle and railway deployment?
Yes. The unit accepts DC 9–48 V input with IGN ignition control, covering 12 V and 24 V vehicle architectures. It carries CE, FCC, EN 50155 and EN 50121-3-2 certifications and MIL-STD-810G qualification, meeting the electrical, EMC and mechanical requirements for road-vehicle and railway rolling-stock installations.
What mounting and expansion options are available?
The AEC-6100-32 supports wall mount, DIN rail and VESA mounting (VESA optional). Expansion is provided by a PCIe Gen4 x8 slot, an M.2 Key-B socket (4G/5G LTE), an M.2 Key-E socket (Wi-Fi), two M.2 Key-M slots for NVMe SSD, two SIM card sockets and six antenna openings for wireless modules.
The AEC-6100-32 is a fanless embedded edge AI computer built on the NVIDIA® Jetson AGX Orin™ system-on-module, combining an 8-core Arm® Cortex-A78AE CPU with a 1,792-core NVIDIA Ampere GPU, 56 Tensor Cores and dual NVDLA deep-learning accelerator engines. With 32 GB of LPDDR5 unified memory and a fanless thermal design rated from −25 °C to +70 °C, the AEC-6100-32 delivers datacenter-class AI inferencing in a compact, vibration-resistant chassis built for continuous deployment in transport, automotive and industrial environments.
Engineered for applications where camera density, ruggedised connectivity and vehicle integration are non-negotiable, the AEC-6100-32 supports eight simultaneous GMSL automotive cameras via Fakra-Z connectors alongside isolated CAN FD, RS-232/422/485 serial ports and 8 DI / 8 DO GPIO. A wide-range 9–48 V DC input with ignition control enables seamless integration into 12 V and 24 V vehicle and rolling-stock electrical systems, whilst a PCIe Gen4 x8 expansion slot and M.2 sockets for 5G and Wi-Fi modules allow the platform to grow with application requirements.
Key Features
- NVIDIA Jetson AGX Orin Compute Platform: 8-core Arm Cortex-A78AE v8.2 CPU, 1,792-core NVIDIA Ampere GPU with 56 Tensor Cores and 2× NVDLA engines; 32 GB LPDDR5 unified memory supports large inference models and high-throughput vision pipelines without offloading to the cloud
- 8× GMSL Automotive Camera Inputs: Eight Fakra-Z connectors for GMSL cameras deliver synchronised multi-camera perception over shielded coaxial links — no frame grabber required; HDMI 2.1 output supports display resolutions up to 8K@60Hz for operator consoles and HMI integration
- Rich Industrial I/O on a Single Platform: 2× RS-232/422/485 serial (DB9), 2× isolated CAN FD, 8 DI / 8 DO isolated GPIO, 5× USB 3.2 and 2× Micro-USB cover PLC integration, automotive sensor buses, fieldbus communication and USB3 Vision camera arrays in a single unit
- Fanless Design for Harsh Environments: Passive cooling eliminates moving parts and ingress points; operating range −25 °C to +70 °C (TDP 30 W) and −25 °C to +55 °C (TDP 40/50 W); shock and vibration qualified to MIL-STD-810G; humidity 5–95 % (non-condensing)
- Wide-Voltage Power with Vehicle Ignition Support: DC 9–48 V input via terminal block covers 12 V, 24 V and 48 V vehicle and industrial supply architectures; ignition (IGN) control enables clean power sequencing for in-vehicle and rolling-stock deployments
- Flexible Mounting and Connectivity Expansion: Wall, DIN rail and VESA mounting options; 1× PCIe Gen4 x8 expansion slot; M.2 Key-B for 4G/5G LTE; M.2 Key-E for Wi-Fi; 2× SIM sockets; 6× antenna openings for wireless modules — adaptable to fixed and mobile infrastructure
- Industrial Certifications: CE, FCC, EN 50155, EN 50121-3-2 — qualifies the AEC-6100-32 for railway rolling-stock, trackside and road-transport EMC compliance programmes
Application Use Cases
Autonomous Vehicles and Advanced Driver Assistance Systems
The AEC-6100-32 is purpose-built for ADAS and autonomous vehicle perception stacks. Eight GMSL camera inputs provide synchronised surround-view coverage, dual isolated CAN FD handles vehicle network communication at up to 5 Mbit/s, and ACC/IGN control ensures reliable power sequencing from engine start to shutdown. The 9–48 V input range accommodates both 12 V passenger vehicles and 24 V commercial trucks without additional power conversion hardware.
Intelligent Traffic Systems and Roadside AI
For roadside and overhead gantry installations, the eight GMSL connectors feed real-time video streams from traffic cameras directly to the Jetson AGX Orin inference engine, enabling licence plate recognition, vehicle classification and incident detection without a separate host PC. The HDMI 2.1 output drives operator displays or digital signage at up to 8K resolution, and the wide operating temperature range covers outdoor deployments across all European climate zones.
Railway and Public Transport
EN 50155 and EN 50121-3-2 certification positions the AEC-6100-32 for on-board rolling-stock and trackside applications. DIN rail mounting fits standard railway cabinet formats, and the MIL-STD-810G shock and vibration qualification satisfies the mechanical stress requirements of mainline and metro rolling stock. The fanless enclosure eliminates filter maintenance cycles and meets cleanliness requirements for enclosed passenger compartment deployments.
Industrial Machine Vision and Automation
In factory environments, the AEC-6100-32 connects up to five USB 3.2 cameras for quality inspection, combined with the PCIe Gen4 x8 slot for frame grabber cards or additional GPU accelerators. Isolated RS-422/485 and GPIO interfaces integrate directly with PLCs and motion controllers over standard fieldbus wiring, and the DIN rail mounting option fits directly into control cabinet infrastructure without custom brackets.
Edge AI Inference for Industrial IoT and Predictive Analytics
The dual NVDLA deep-learning accelerators run inference workloads in parallel with the Ampere GPU, allowing the AEC-6100-32 to execute multiple models simultaneously — anomaly detection, optical character recognition and thermal imaging classification — on the 32 GB unified memory pool. A 5G cellular module in the M.2 Key-B slot delivers low-latency telemetry to cloud dashboards without exposing raw sensor data over public networks.
Why Choose the AEC-6100-32
Jetson AGX Orin AI Performance Without Active Cooling
Combining a fanless thermal design with the NVIDIA Jetson AGX Orin's 1,792-core Ampere GPU and dual NVDLA engines, the AEC-6100-32 is one of the few platforms in its class that sustains high-throughput AI inference across the full −25 °C to +70 °C operating range without relying on fan-based cooling. This makes it directly deployable in environments where dust ingress, vibration or maintenance access constraints rule out fan-cooled alternatives — from roadside enclosures to rolling-stock computer cabinets.
Native GMSL Camera Ecosystem Integration
Unlike embedded AI platforms that require third-party frame grabbers or camera serialiser boards, the AEC-6100-32 integrates eight GMSL Fakra-Z connectors directly on the chassis. This allows systems integrators to connect automotive-grade GMSL cameras over standard coaxial cabling at distances up to 15 metres per channel, reducing system BOM cost, connector count and potential failure points in multi-camera perception arrays for ADAS, traffic and surveillance programmes.
Cross-Market Deployment from a Single Qualified Platform
Wall mount, DIN rail and VESA mounting options, a 9–48 V DC wide-range input, IGN control and industrial certifications covering CE, FCC, EN 50155 and MIL-STD-810G allow a single AEC-6100-32 qualification to span automotive, railway, smart city and factory automation projects. Engineers reduce re-qualification effort and procurement teams consolidate to a single SKU across programmes — a material operational advantage for organisations deploying across multiple market verticals simultaneously.
Frequently Asked Questions
What processor does the AEC-6100-32 use?
The AEC-6100-32 is powered by the NVIDIA Jetson AGX Orin system-on-module, which integrates an 8-core Arm Cortex-A78AE v8.2 64-bit CPU, a 1,792-core NVIDIA Ampere GPU with 56 Tensor Cores, dual NVDLA deep-learning accelerator engines, and 32 GB of LPDDR5 unified memory.
How many cameras does the AEC-6100-32 support?
The AEC-6100-32 supports up to eight GMSL automotive cameras simultaneously via Fakra-Z coaxial connectors. This makes it suitable for surround-view ADAS perception, multi-lane traffic monitoring and wide-area surveillance applications without additional frame grabber hardware.
What is the operating temperature range of the AEC-6100-32?
The AEC-6100-32 operates from −25 °C to +70 °C in fanless 30 W TDP mode, and from −25 °C to +55 °C at 40/50 W TDP. Storage temperature range is −40 °C to +85 °C. The unit meets MIL-STD-810G shock and vibration requirements.
Is the AEC-6100-32 suitable for vehicle and railway deployment?
Yes. The unit accepts DC 9–48 V input with IGN ignition control, covering 12 V and 24 V vehicle architectures. It carries CE, FCC, EN 50155 and EN 50121-3-2 certifications and MIL-STD-810G qualification, meeting the electrical, EMC and mechanical requirements for road-vehicle and railway rolling-stock installations.
What mounting and expansion options are available?
The AEC-6100-32 supports wall mount, DIN rail and VESA mounting (VESA optional). Expansion is provided by a PCIe Gen4 x8 slot, an M.2 Key-B socket (4G/5G LTE), an M.2 Key-E socket (Wi-Fi), two M.2 Key-M slots for NVMe SSD, two SIM card sockets and six antenna openings for wireless modules.
Get a Quote
