ADVANTECH AIR-030 — NVIDIA AI PC
Compact AI inference system based on NVIDIA Jetson AGX Orin 32 GB (200 TOPS) or 64 GB (275 TOPS) — 3 × 2.5GbE (PoE option on 2), HDMI 2.0 4K, 4 × USB 3.2 A + USB-C, 4 × RS-232/422/485, 2 × CAN, 16-bit DIO; M.2 E key, M.2 B key storage, SD, optional PCIe x16 slot, 2 × MIPI-CSI, TPM 2.0; 9 – 36 VDC; −10 to 60 °C; IEC 61000-6-4; Ubuntu 22.04 with JetPack 6.2, Edge AI SDK, AWS Greengrass. Listed as “New” by Advantech.
Status: In production
Product: Compact AI inference system — Jetson AGX Orin 32 / 64 GB, 3 × 2.5GbE, 4 × COM, PCIe x16 option
AI module: Jetson AGX Orin 32 GB (200 TOPS) · 64 GB (275 TOPS)
Interfaces: 3 × 2.5GbE (PoE option) · HDMI 2.0 · 4 × USB 3.2 A + USB-C · 4 × COM · 2 × CAN · 16-bit DIO
Expansion: M.2 E key · M.2 B key storage · SD 3.0 · PCIe x16 (optional) · 2 × MIPI-CSI · TPM 2.0
Power / size: 9 – 36 VDC · 200 × 220 × 74 mm · 3.63 kg · wall
AIR-030 is the Jetson AGX Orin model of the AIR line: the 200 × 220 × 74 mm, 3.63 kg chassis holds an AGX Orin 32 GB (8-core, 1792 CUDA + 56 Tensor Cores, 200 TOPS) or 64 GB (12-core, 2048 CUDA + 64 Tensor Cores, 275 TOPS) module and 64 GB eMMC. The 9 – 36 VDC wide input, the −10 to 60 °C (55 °C on 64 GB) operating range and the IEC 61000-6-4 heavy-industry EMC certificate are for factory-floor and on-machine use. Per the datasheet it ships with the Ubuntu 22.04 + JetPack 6.2 image (the web table carries the older Ubuntu 20.04 / JetPack 5 information); it is AWS Greengrass certified and supports Advantech Edge AI SDK and SUSI.
Two of the three Intel I225 2.5GbE ports power cameras with optional PoE; the third shares its PCIe lane with the M.2 2230 E key (one or the other is used). There are HDMI 2.0 4K60, four USB 3.2 Type A and one USB-C, two internal USB 2.0 for a security dongle, four RS-232/422/485, two CAN, 16-bit digital I/O and line-out; two internal 2-lane MIPI-CSI connectors serve camera modules. Storage is on the M.2 2280/3052 B key (PCIe x2 + USB 3.0) and the SD 3.0 slot; the optional dual-layer chassis adds a PCIe x16 slot (x8 signal) and raises the height to 123 mm. TPM 2.0 is built in; the adapter (150 W 19 V or 230 W 24 V) is selected separately. Certifications CE, FCC Class B, CCC, BSMI, UL, CB; there is no RED certificate.
Technical specifications
| AIR-030 | |
|---|---|
| System (Jetson module) | |
| Jetson module | NVIDIA Jetson AGX Orin 32 GB (-B90A1) or 64 GB (-S30A1) |
| CPU | 32 GB: 8-core; 64 GB: 12-core Arm Cortex-A78AE v8.2 |
| GPU | 32 GB: 1792 CUDA + 56 Tensor Cores (939 MHz); 64 GB: 2048 CUDA + 64 Tensor Cores (1.3 GHz) |
| AI performance | 32 GB: up to 200 sparse INT8 TOPS; 64 GB: up to 275 TOPS; TDP 15 – 30 / 15 – 50 W |
| Memory | 32 or 64 GB 256-bit LPDDR5 |
| Flash | 64 GB eMMC 5.1 |
| I/O | |
| Ethernet | 3 × 2.5GbE RJ45 (Intel I225); optional PoE on LAN1 and LAN2; LAN3 shares PCIe with the M.2 E key (one or the other) |
| Display | 1 × HDMI 2.0 (4K @ 60 Hz) |
| USB | 4 × USB 3.2 Type A; 1 × USB 3.2 Type C; 2 × internal USB 2.0 (dongle only); 1 × Micro USB OTG (recovery) |
| CAN | 2 × CAN |
| Digital I/O | 16-bit |
| Serial port | 4 × RS-232/422/485 |
| Audio / camera | 1 × line-out; 2 × internal 2-lane MIPI-CSI connectors |
| Expansion and storage | |
| M.2 / PCIe | 1 × M.2 2230 E key (shared with LAN3); 1 × PCIe x16 slot (optional, PCIe x8 signal; dual-layer chassis); TPM 2.0 built in |
| Storage | 1 × M.2 2280/3052 B key (PCIe x2, USB 3.0); 1 × SD 3.0 slot; the web table also lists M.2 2280 NVMe |
| Power | |
| Input voltage | 9 – 36 VDC, 4-pin terminal block; AT / ATX (ATX default); optional 150 W 19 V or 230 W 24 V adapter |
| Environment | |
| Operating temperature | 32 GB: −10 to 60 °C; 64 GB: −10 to 55 °C (0.7 m/s airflow, 50 W TDP mode) |
| Humidity | 95 % @ 40 °C (non-condensing) |
| Vibration | 3 Grms, 5 – 500 Hz, random, 1 hr/axis |
| Mechanical and software | |
| Dimensions (W × D × H) | 200 × 220 × 74 mm (single layer); dual-layer (PCIe) 123 mm high |
| Weight | 3.63 kg |
| Mounting | Wall mount (2 brackets included) |
| Operating system | Linux Ubuntu 22.04 LTS with JetPack 6.2 (image 20706UJTNS0023); the web table says Ubuntu 20.04 / JetPack 5; Edge AI SDK, SUSI; AWS Greengrass certified |
| Certification | |
| Certification | EMC: CE, FCC Class B, CCC, BSMI · safety: UL, CB · IEC 61000-6-4 heavy-industry EMC (no RED certificate) |
Key features
- Compact, high-performance AI box with up to 275 TOPS of AI computing
- NVIDIA Jetson AGX Orin 32 / 64 GB built in
- 9 – 36 V wide power and −10 to 60 °C wide temperature support
- Multiple I/O: 3 × LAN, DIO, 4 × COM, 4 × USB 3.2 and USB Type C
- M.2 2280 (NVMe), M.2 2230 and optional PCIe x16 expansion
- IEC 61000-6-4 heavy-industry certificate; AWS Greengrass certified
Ordering information
| Part number | Description |
|---|---|
| AIR-030-B90A1 | AGX Orin 32 GB, 64 GB eMMC, 3 × 2.5GbE, 2 × CAN, 4 × COM, 16-bit DIO, −10 to 60 °C |
| AIR-030-S30A1 | AGX Orin 64 GB, 64 GB eMMC, 3 × 2.5GbE, 2 × CAN, 4 × COM, 16-bit DIO, −10 to 55 °C |
The package includes a Micro USB cable for recovery, 2 wall brackets, the 6-pin CAN terminal plug, the 4-pin DC-in block and the Ubuntu / JetPack 6.2 image; the adapter, SSD and PCIe expansion are optional.
Accessories and optional items
| Part number | Description |
|---|---|
| 96PSA-A150W19P4-4 / 96PSA-A230W24P4-3 | AC/DC adapter 19 V 150 W (0 to 40 °C) / 24 V 230 W (−20 to 60 °C); power cord US 1702002600, EU 1702002605, UK 1702031801, PSE 1700000237 |
| 1700019414 / 1700024248-01 | DB9 debug cable 30 cm / Micro USB reflashing cable 40 cm |
Is this AIR system right for your project?
Share your application, the number and type of cameras (GigE / USB / MIPI), the AI power required, the wireless (4G / 5G / Wi-Fi) and storage needs, the ambient temperature and the supply voltage; let’s decide on the module, the model and the accessory set together. FST configures the AIR system with the camera, wireless module and SSD selection and delivers it as one solution with model deployment and remote-management setup services.
All product specifications are subject to change without notice; data compiled from Advantech’s AIR-030 datasheet dated 17 August 2026 and the advantech.com product page.
Source: Advantech product data