Name: Supermicro ARS-E103-JONX Series
Brand: Supermicro
SKU: ARS-E103-JONX-SERIES
Availability: InStock

ARS-E103-JONX-H — Orin NX 16GB / 10GbE

157 TOPS (sparse) / 100 TOPS (INT8)
1024 CUDA Cores · 32 Tensor Cores
8-core ARM Cortex-A78AE @ 2.0GHz
16GB ECC LPDDR5X
1× 10GbE (Marvell AQC113C)
3× 1GbE (Intel I210) + 1× 1GbE (SoC)
4× PoE (optional via jumper)
Super Mode 40W support

ARS-E103-JONX-H2 — Orin NX 16GB / 10GbE

157 TOPS (sparse) / 100 TOPS (INT8)
1024 CUDA Cores · 32 Tensor Cores
8-core ARM Cortex-A78AE @ 2.0GHz
16GB ECC LPDDR5X
1× 10GbE (Marvell AQC113C)
2× 1GbE with PoE+
Alternative PoE configuration
Super Mode 40W support

ARS-E103-JONX-L — Orin NX 8GB / 1GbE

117 TOPS (sparse) / 70 TOPS (INT8)
1024 CUDA Cores · 32 Tensor Cores
6-core ARM Cortex-A78AE @ 2.0GHz
8GB ECC LPDDR5X
4× 1GbE (Intel I210)
+ 1× 1GbE (SoC)
4× PoE (optional via jumper)
Value-optimized edge AI

ARS-E103-JONX-L2 — Orin NX 8GB / 1GbE

117 TOPS (sparse) / 70 TOPS (INT8)
1024 CUDA Cores · 32 Tensor Cores
6-core ARM Cortex-A78AE @ 2.0GHz
8GB ECC LPDDR5X
2× 1GbE with PoE+
Alternative PoE configuration
Value-optimized edge AI
Lower PoE port count

Specifications

Manufacturer

Supermicro
San Jose, CA, USA
Nasdaq: SMCI

AI Performance

H/H2: 157 TOPS sparse
H/H2: 100 TOPS INT8
L/L2: 117 TOPS sparse
L/L2: 70 TOPS INT8

GPU

1024 NVIDIA CUDA Cores
32 Tensor Cores
NVIDIA Ampere Arch

CPU

H/H2: 8-core A78AE
L/L2: 6-core A78AE
6MB Cache

Memory

H/H2: 16GB ECC LPDDR5X
L/L2: 8GB ECC LPDDR5X
Onboard (soldered)

Storage

1× M.2 M-key 2280
PCIe 4.0 x4 NVMe
Up to 2TB NVMe SSD

Networking

H/H2: 1× 10GbE + 4× 1GbE
L/L2: 5× 1GbE
Intel I210 + Marvell AQC113

USB

1× USB 3.2 Gen2 Type-C
3× USB 3.2 Gen2 Type-A

Video

1× HDMI 2.0b
4K@60Hz

Serial / CAN

2× COM RS232/422/485
1× CAN Bus

DIO

4× Digital Input
4× Digital Output
Terminal block

Expansion

M.2 E-key 2230 (WiFi)
M.2 B-key 3052 (5G/LTE)
Dual nano-SIM

Operating Temp

-25°C to 60°C
(with 0.7 m/s airflow)
Fanless passive cooling

Power Input

9–36VDC
Lockable DC connector
180W PSU included

Enclosure

185 × 80 × 140mm
1.47 kg net
Fanless · DIN-Rail
Chassis: CSE-E102L2

Carrier Board

AOM-JSOR-001
Supermicro-designed
Industrial grade

Software

JetPack SDK
TensorRT · DeepStream
NVIDIA Isaac

AI Frameworks

TensorFlow · PyTorch
DeepStream · Caffe
PaddlePaddle · ONNX

Certifications

RoHS Compliant
Industrial Grade
SuperMicro 3-Year

Applications

Imaging & Computer Vision
Industrial Automation
Healthcare · Digital Signage
Smart Retail / PoS

Overview

Supermicro (Nasdaq: SMCI), one of the world's largest server and edge computing manufacturers, brings its enterprise engineering to edge AI with the ARS-E103-JONX Series. These fanless embedded systems integrate the NVIDIA Jetson Orin NX module into Supermicro's custom AOM-JSOR-001 carrier board and CSE-E102L2 industrial chassis — delivering up to 157 TOPS of AI inference in a compact, DIN-rail mountable enclosure measuring just 185×80×140mm.

The series spans four SKUs that pair two Orin NX configurations (16GB 8-core / 8GB 6-core) with two networking options (10GbE or 1GbE). All models share a rich I/O set: USB 3.2 Gen2 Type-C and Type-A ports, dual RS232/422/485 serial, CAN Bus, 4× DI/4× DO terminal block, HDMI 2.0b, and triple M.2 expansion for NVMe storage, WiFi, and 5G/LTE with dual nano-SIM.

Designed for industrial environments, the ARS-E103-JONX operates fanless from -25°C to 60°C with a wide 9–36VDC power input and lockable DC connector. Target applications include imaging and computer vision, industrial automation, healthcare systems, digital signage and point-of-sale — anywhere demanding edge AI workloads meet industrial reliability requirements. As the first Supermicro Jetson product on QS Compute, the ARS-E103 series brings data center-grade engineering to the edge.

Model Lineup