AI Single Board Computer (AI SBC)
High-Performance AI SBCs for Next-Generation Edge Intelligence.
Built on TI Jacinto 7 series processors, InHand Networks’ AI SBCs deliver 2 TOPS and 8 TOPS edge AI performance for vision inference, video analytics, and intelligent inspection. A standard SBC design, rich I/O, and industrial-grade reliability help accelerate deployment.
Key Value Propositions
2/8 TOPS scalable AI performance
Mainstream model migration
Integrated algorithm-and-hardware delivery
Industrial-grade reliability
TI Jacinto 7 Platform for Reliable Edge AI Compute
Powered by TI AM62A / AM68A processors, the portfolio supports machine vision, object detection, image recognition, and multi-channel video analytics.
Debian Linux for Open, Flexible Development
Debian 13 support, Python and C/C++ environments, plus preinstalled OpenCV, GStreamer, and NumPy help developers build and deploy faster.
Mainstream AI Frameworks for Easier Deployment
Integrated TI TIDL runtime supports TensorFlow Lite, ONNX Runtime, and PyTorch via ONNX for efficient edge inference.
High-Performance Vision Processing for Vision AI
The 7th-generation ISP and VPAC support image pre-processing, WDR, LDC, and other capabilities for intelligent cameras, AI vision boxes, and industrial inspection.
Standard SBC Form Factor with Rich Interfaces
The 85 x 56 mm SBC form factor offers MIPI CSI, HDMI / mini DP, USB, Gigabit Ethernet, and a 40-pin HAT-compatible connector.
Scalable Options from Entry-Level to High-Performance Edge AI
MO 62A fits entry-level and cost-sensitive edge AI with Wi-Fi 5 + BLE. MO 68A supports higher-performance AI and video workloads with USB 3.0 and PCIe 3.0.
Industrial-Grade Design for Stable Deployment
Hardware watchdog, RTC, Secure Boot, TrustZone, and hardware crypto acceleration provide a reliable foundation for long-term deployments.
On-Device AI for Unified Algorithm Management
With AI Hub, InHand Networks’ algorithms and user-owned models can be centrally managed, deployed, invoked, and updated.
Key Parameters
 |  |
|
| Model | MO 62A | MO 68A |
| CPU | TI AM62A74 4× Cortex-A53 @ 1.4GHz | TI AM68A 2× Cortex-A72 @ 2.0GHz |
| AI Accelerator | 2 TOPS | 8 TOPS |
| RAM | 2GB / 4GB / 8GB | 4GB / 8GB |
| Storage | Micro SD | Micro SD |
| Ethernet | 1× Gigabit Ethernet | 1× Gigabit Ethernet |
| USB | 4× USB 2.0 Type-A | 4× USB 3.0 Type-A |
| PCIe | - | 1× PCIe 3.0 |
| Display | 1× micro HDMI | 1× mini DP + up to 2× 4-lane MIPI DSI |
| Camera | 1× 4-lane MIPI CSI-2 | up to 2× 4-lane MIPI CSI-2 |
| Wireless | Wi-Fi 5 (802.11ac) BLE 4.2 | - |
| 40-pin Connector | GPIO / I²C / SPI / UART / PCM HAT-compatible | GPIO / I²C / I2S / SPI / UART / PCM HAT-compatible |
| OS | Debian 13 Trixie (Linux 6.12) | Debian 13 Trixie (Linux 6.12) |
| AI Runtime | TI TIDL (TFLite / ONNX) | TI TIDL (TFLite / ONNX) |
| Security | Secure Boot / TrustZone / OP-TEE Hardware AES-256 encryption | Secure Boot / TrustZone / OP-TEE Hardware AES-256 encryption |
| Dimensions | 85 × 56mm | 85 × 56mm |
| Weight | 47g | 53g |
| Operating Temperature | 0 ~ 50℃ | 0 ~ 50℃ |
| Power Input | USB Type-C 5V/5A DC | USB Type-C 5V/5A DC |
| Power Consumption | 25W (MAX) | 25W (MAX) |
| Resources → | Resources → | |
| Datasheet → | Datasheet → |