Back to Advantech

Advantech MIC-742-AT AI Inference Systems

Advantech MIC-742-AT AI Inference Systems are powered by NVIDIA® Jetson Thor™ T5000 and T4000 modules. The systems deliver up to 2070 TFLOPS FP4 AI computing performance for demanding edge workloads that require high-performance processing, advanced vision connectivity, and industrial deployment support.

The Advantech MIC-742-AT family supplies Jetson T5000 and Jetson T4000 processor options with Arm® Neoverse™ V3AE CPU architecture, NVIDIA Blackwell GPU architecture, fifth-generation Tensor Cores, LPDDR5X memory, and 1TB pre-installed NVMe storage on A1 SKUs. The MIC-742-AT AI Inference Systems also include QSFP28 Ethernet, 5GbE, USB 3.2 Gen 2, CAN FD, I2C, GMSL2.0, M.2 expansion, SATA, TPM, and support for 19V to 36V power input.

These systems are ideal for industrial edge AI deployments that need high compute density, multiple camera connections, fast networking, and flexible expansion. The devices support NVIDIA JetPack™ 7.0, wall-mount or DIN-rail installation, and -10°C to +60°C operation with 0.7m/s airflow.

Features

  • Edge AI inference system powered by NVIDIA Jetson Thor T5000 and Jetson T4000 modules
  • AI computing performance of up to 2070TFLOPS FP4
  • NVIDIA Blackwell architecture GPU with fifth-generation Tensor Cores
  • QSFP28 networking support
    • 4x 25GbE with Jetson T5000
    • 3x 25GbE with Jetson T4000
  • Rich I/O support with 5GbE, USB 3.2 Gen 2, CAN FD, I2C, HDMI, audio, and microUSB debug connectivity
  • Total 8-channel GMSL2.0 support with mini FAKRA connectors
  • Expansion support through M.2 E key, M.2 B key, M.2 M key, SATA, and Nano SIM interfaces
  • Onboard TPM support for security-enabled edge systems
  • Wall-mount and DIN-rail installation support

Applications

  • AI vision systems requiring GMSL2.0 camera connectivity
  • Industrial AI deployments requiring high-speed networking and rich I/O
  • Edge AI inference systems
  • Embedded AI systems requiring NVIDIA Jetson Thor T5000 or T4000 performance

Specifications

  • Processor platform options
    • MIC-742-AT7A1 and MIC-742-AT7A2 with NVIDIA Jetson T5000
    • MIC-742-AT6A1 and MIC-742-AT6A2 with NVIDIA Jetson T4000
  • GPU and memory
    • Jetson T5000 2560-core NVIDIA Blackwell GPU with 96 fifth-generation Tensor Cores
    • Jetson T4000 1536-core NVIDIA Blackwell GPU with 64 fifth-generation Tensor Cores
    • 128GB LPDDR5X memory for Jetson T5000 models
    • 64GB LPDDR5X memory for Jetson T4000 models
  • Connectivity and I/O
    • One RJ45 Ethernet port supporting 10Mbps, 100Mbps, 1000Mbps, 2500Mbps, and 5000Mbps
    • One QSFP28 interface with 4x25GbE for T5000 or 3x25GbE for T4000
    • HDMI output with maximum resolution of 3840 x 2160 at 60Hz
    • Four USB 3.2 Gen 2 ports
    • One microUSB OTG port and one microUSB console port for UART-to-USB debug
    • One onboard audio speaker output and microphone input
    • Four CAN FD interfaces on T5000 models
    • Four MIPI 4-lane interfaces with support for 8-channel GMSL2.0 via mini FAKRA connectors
    • One I2C interface
    • One Nano SIM slot on the side of the board
  • Processor system
    • Jetson T5000 14-core Arm Neoverse V3AE 64-bit SMP CPU architecture
    • Jetson T4000 12-core Arm Neoverse V3AE 64-bit SMP CPU architecture
    • 64KB + 64KB L1 cache per core, 1MB L2, and 16MB L3 cache
  • Expansion and storage
    • One M.2 2230 E-key interface with PCIe x1 and USB 2.0 signals
    • One M.2 3052/3042 B-key interface with USB 3.0 and USB 2.0 signals
    • Two SATA connectors
    • One M.2 2280 M-key NVMe PCIe x4 storage interface
    • 1TB pre-installed SSD on A1 SKUs
  • Power and environment
    • Power input range of 19V to 36V
    • 2-pin terminal block power input
    • Three SMART FAN internal pin headers
    • Operational temperature range of -10°C to +60°C with 0.7m/s airflow
    • Operating humidity of 95% at 40°C, non-condensing
    • Vibration rating of 3Grms from 5Hz to 500Hz, random, 1hr/axis
  • Mechanical and certifications
    • Dimensions of 217mm x 76.4mm x 184.5mm
    • Weight of 3kg
    • CE, FCC, CCC, KCC, BSMI, and UL certifications with no RED certification

Kit Contents

  • MIC-742-AT AI inference developer kit
  • 1700023619-01 micro USB cable for system recovery
  • 1652008236-01 terminal plug 20P for CANbus
  • 1652002202 plug-in block 2P for power
  • 1960116156N001 mounting bracket for MIC-742, quantity of 2
  • 1939000410 screws for mounting bracket, quantity of 4
  • 1990019482N010 thermal pad for GMSL, quantity of 2
  • 1990040296N010 thermal pad for 5G module
  • 1930005673-11 screws for 5G/Wi-Fi module, quantity of 2

Dimensions

Mechanical Drawing - Advantech MIC-742-AT AI Inference Systems
Advantech MIC-742-AT AI Inference Systems
Published: 2026-07-13 | Updated: 2026-07-16