ADVANTECH MIC-742-AT — NVIDIA AI PC
AI inference system based on NVIDIA Jetson Thor (T5000 128 GB / T4000 64 GB) with up to 2070 FP4 TFLOPS — QSFP28 (4 × / 3 × 25GbE), 5GbE, 4 × USB 3.2 Gen 2, 4K HDMI, 4 × CAN FD, I2C; optional 8-ch GMSL2 (mini FAKRA), M.2 E key Wi-Fi, M.2 B key 5G, 2 × SATA, TPM; 1 TB NVMe pre-installed; 19 – 36 VDC; −10 to 60 °C; JetPack 7.0. Advantech catalogue status “Release to Market”.
Status: In production
Product: Jetson Thor AI inference system — T5000 / T4000, QSFP28 25GbE, 8 × GMSL2, 4 × CAN FD
AI module: Jetson T5000 128 GB (2070 TFLOPS FP4) · T4000 64 GB
Interfaces: QSFP28 4 × 25GbE · 5GbE · 4 × USB 3.2 Gen 2 · HDMI · 4 × CAN FD · I2C · 8 × GMSL2 (optional)
Expansion: M.2 E key Wi-Fi · M.2 B key 5G + nano SIM · 2 × SATA · M.2 NVMe (1 TB) · TPM
Power / size: 19 – 36 VDC · 217 × 76.4 × 184.5 mm · wall / DIN rail
MIC-742-AT is Advantech’s AI inference system built around the NVIDIA Jetson Thor module; in Advantech’s launch announcement the MIC-AI series is positioned for agentic and physical AI workloads (large language and vision-language models, robotics, multi-camera perception). Depending on the part number it ships with a Jetson T5000 (14-core Neoverse V3AE, 2560 Blackwell CUDA + 96 fifth-gen Tensor Cores, 128 GB LPDDR5X) or T4000 (12-core, 1536 CUDA + 64 Tensor Cores, 64 GB); the A1 SKUs have JetPack 7.0 and a 1 TB NVMe installed. The 217 × 76.4 × 184.5 mm, 3 kg chassis mounts on a wall or DIN rail, runs from 19 – 36 VDC and, per the datasheet, operates from −10 to 60 °C (the web table says −10 to 35 °C; both values are given).
The network side is designed for Thor’s bandwidth: the QSFP28 cage provides four 25GbE links on T5000 and three on T4000, plus one 5GbE RJ45. There are four USB 3.2 Gen 2, 4K HDMI, four CAN FD on T5000, one I2C, speaker out and mic in, and two Micro USB ports for OTG and console. The optional MIC-FG-8G2C1 frame grabber connects a total of 8 GMSL2 cameras through mini FAKRA connectors from four MIPI 4-lane inputs. M.2 2230 E key (Wi-Fi 6) and M.2 3052 B key (5G + GNSS, nano SIM) slots, two SATA connectors, an on-board TPM and M.2 2280 NVMe (PCIe x4) storage are provided; three smart-fan headers serve cooling. Certifications CE, FCC, CCC, KCC, BSMI, UL; there is no RED certificate. Advantech’s MIC-743-AT on the same platform (desktop box, −10 to 35 °C) is preliminary.
Technical specifications
| MIC-742-AT | |
|---|---|
| System (Jetson Thor module) | |
| Jetson module | NVIDIA Jetson T5000 (-AT7) or Jetson T4000 (-AT6) — by part number |
| CPU | T5000: 14-core Arm Neoverse V3AE 64-bit; T4000: 12-core — 64 + 64 KB L1 per core, 1 MB L2, 16 MB L3 |
| GPU | T5000: 2560-core NVIDIA Blackwell GPU with 96 fifth-gen Tensor Cores; T4000: 1536 cores, 64 Tensor Cores |
| AI performance | Up to 2070 TFLOPS (FP4) — Advantech feature list |
| Memory | T5000: 128 GB LPDDR5X; T4000: 64 GB LPDDR5X |
| Ethernet | |
| RJ45 | 1 × 10/100/1000/2500/5000 Mbps (5GbE) |
| QSFP28 | 1 × QSFP28 — T5000: 4 × 25GbE; T4000: 3 × 25GbE |
| I/O | |
| Display | HDMI (up to 3840 × 2160 @ 60 Hz) |
| USB | 4 × USB 3.2 Gen 2; 1 × Micro USB OTG; 1 × Micro USB console (UART, debug) |
| CAN | 4 × CAN FD (T5000 only); 1 × I2C |
| Audio | 1 × speaker out, on-board mic in |
| GMSL cameras (optional) | 4 × MIPI 4-lane — total 8-ch GMSL2 with mini FAKRA via MIC-FG-8G2C1 |
| Expansion and storage | |
| M.2 / SIM | 1 × M.2 2230 E key (PCIe x1 + USB 2.0, Wi-Fi); 1 × M.2 3052/3042 B key (USB 3.0 + USB 2.0, 4G/5G); 1 × nano SIM (board edge) |
| SATA / TPM | 2 × SATA connector; on-board TPM |
| Storage | 1 × M.2 2280 M key NVMe (PCIe x4) — 1 TB SSD pre-installed on A1 SKUs |
| Power | |
| Input voltage | 19 – 36 VDC, 2-pin terminal block; 3 × smart-fan pin header |
| Environment | |
| Operating temperature | −10 to 60 °C (0.7 m/s airflow) — datasheet; the web table says −10 to 35 °C |
| Humidity | 95 % @ 40 °C (non-condensing) |
| Vibration | 3 Grms, 5 – 500 Hz, random, 1 hr/axis |
| Mechanical and software | |
| Dimensions (W × D × H) | 217 × 76.4 × 184.5 mm |
| Weight | 3 kg |
| Installation | Wall mount / DIN rail (kit 98R2S742000); the web table says “desktop” |
| Operating system | NVIDIA JetPack 7.0 (installed on A1 SKUs) |
| Certification | |
| Certification | CE, FCC, CCC, KCC, BSMI, UL (no RED certification); web table CB/UL/CE/FCC/BSMI/CCC |
Key features
- Embedded NVIDIA Jetson T5000 / T4000 with up to 2070 TFLOPS (FP4)
- Supports 1 × QSFP28 (T5000: 4 × 25GbE / T4000: 3 × 25GbE)
- Supports 1 × 5GbE, 4 × USB 3.2 Gen 2, M.2 A/E key (Wi-Fi), M.2 B key (LTE)
- Supports 4 × CANbus, 1 × I2C
- Total 8-ch GMSL2 with mini FAKRA connectors
Ordering information
| Part number | Description |
|---|---|
| MIC-742-AT7A1 | Jetson T5000, 128 GB, JetPack 7.0 installed, 1 TB NVMe pre-installed |
| MIC-742-AT6A1 | Jetson T4000, 64 GB, JetPack 7.0 installed, 1 TB NVMe pre-installed |
| MIC-742-AT7A2 | Jetson T5000, 128 GB, no OS and no SSD |
| MIC-742-AT6A2 | Jetson T4000, 64 GB, no OS and no SSD |
AT7A1 and AT6A1 are listed in the advantech.com table; the A2 versions come from the datasheet dated 14 August 2026. The package includes a Micro USB cable for recovery, the 20-pin CAN terminal plug, the 2-pin power block, 2 mounting brackets and the GMSL / 5G thermal pads; the adapter is optional.
Accessories and optional items
| Part number | Description |
|---|---|
| MIC-FG-8G2C1 | 8-ch GMSL2 frame grabber with mini FAKRA; cables 1751000758-01 (1 to 4 cameras), 1751000805-01 (per camera), internal power cable 1700026907-01 |
| EWM-W179M201E | Wi-Fi 6 + Bluetooth M.2 A/E key module; 2 × antenna cable 1751000620-01, 2 × antenna 1751000460-01 |
| AIW-356-DQ | 5G Sub-6 + GNSS M.2 3052 B key (by region); 4 × antenna cable 1751000623-01, 4 × antenna 1750009372-01 |
| XMIC-FSP230-AAAN3 | AC/DC adapter 100 – 240 V, 230 W, 24 V, 2-pin terminal block; power cord US 1702002600, EU 1702002605, PSE/BSMI 1700029019-01 |
| 96FD80-P2TB-TS1 | SSSTC M.2 2280 2 TB NVMe Gen4 x4 SSD (−25 to 85 °C) |
| 98R2S742000 | MIC-742 DIN-rail mounting kit |
Is this Jetson Thor system right for your project?
Share your application (robotics / visual inspection / generative AI), the number and type of cameras (GMSL2 / GigE / USB), the network needs (25GbE / 10GbE), the CAN and I/O requirements and the ambient temperature; let’s decide on the module (T5000 / T4000), the model and the accessory set together. FST delivers the Thor system together with the camera / sensor layer, network design, model deployment and commissioning services as one solution.
All product specifications are subject to change without notice; data compiled from Advantech’s MIC-742-AT datasheet dated 14 August 2026 and the advantech.com product page.
Source: Advantech product data