Radxa announced the 'Radxa C200 Orin Developer Kit'. Combining a Jetson Orin NX 8GB module with a dedicated IO board, it's a dev kit strongly focused on robotics and industrial edge AI. It highlights 'Max 117 TOPS' via Super mode (MAXN_SUPER), though official docs note performance depends on procedures.
Radxa announced the Radxa C200 Orin Developer Kit. Product info is available on the official page and docs. Combining a Jetson Orin NX 8GB module and a custom IO board, it’s a dev kit strongly focused on robotics/industrial edge AI. “Max 117 TOPS” via Super mode (MAXN_SUPER) is a headline feature, but docs explicitly state performance depends on the setup procedure.
| ■ Radxa C200 Orin Developer Kit | |
|---|---|
| CPU | 6-core Arm Cortex-A78AE Max 2.0GHz |
| GPU | Ampere (1024 CUDA / 32 Tensor). Max 1173MHz in Super mode |
| NPU | Normal 70 TOPS, Super 117 TOPS (INT8 Sparse). DLA + PVA |
| Memory | 8GB LPDDR5 6400MT/s (Peak 102GB/s) |
| Storage | M.2 M Key NVMe x2 (PCIe Gen4 x2/x4) + M.2 E Key for Wi-Fi/BT |
| Wireless | Via M.2 E Key (Module required separately) |
| Wired LAN | RJ45 1GbE |
| Video | DisplayPort x1 (Official page says DP 1.4, PDF says DP 1.2 MST) |
| USB | USB 3.2 Type-A x4, Type-C x1 (Product Brief says Gen2, Type-C for Recovery) |
| Expansion | 40-pin GPIO, MIPI CSI 22-pin FPC x2, CAN, PoE header, Fan, RTC |
| Size | 100x79mm |
| Power | DC 5.5x2.5mm, 9-20V |
| OS | Ubuntu 20.04 / 22.04 LTS. JetPack support |
117 TOPS-class Jetson Orin NX 8GB on a 100x79mm board with 2x NVMe. Key checks: Discrepancy in DP specs and installation constraints for MAXN_SUPER.
Jetson Orin NX 8GB (6-core Cortex-A78AE + Ampere GPU). AI performance listed as Normal 70 TOPS / Super 117 TOPS (INT8 Sparse). Verify if your target power mode can handle your workload throughput. Super Mode (MAXN_SUPER) is uncapped, so thermal throttling is a real risk. Radxa C200 assumes a fan (PWM header included). Case design must ensure airflow.
Docs state MAXN_SUPER requires command-line installation (not Balena Etcher/SDK Manager). If performance is your goal, be ready for CLI setup.
8GB LPDDR5 (102GB/s). Prioritizes GPU/DLA bandwidth. 8GB is tight for large LLMs; requires optimization. Storage: 2x NVMe (Gen4 x2/x4). Physically separating OS and Data (or Log/Dataset) is easy. Wireless: Via M.2 E Key. You choose the module to match field RF requirements.
100x79mm board. 4x USB-A, RJ45, DP, Power on one side simplifies desk wiring. M.2 (NVMe x2, Wireless x1) and MIPI CSI take up board surface space. Case design needs to account for SSD heat and FPC routing.
Network: 1GbE + PoE header. PoE usage details (HAT/Backpower) not fully detailed on the product page; check docs. Power: DC 9-20V. Fits 12V/19V field power. USB-C is for data/recovery, not primary power. MicroSD: No slot on C200 module for production. NVIDIA SD images require kernel patching (EDK2). Don’t assume “MicroSD boot like Jetson Nano DevKit”. Supply: Guaranteed until at least Nov 2030 (PDF). Good for industrial planning.
Radxa C200 Orin Developer Kit puts Jetson Orin NX 8GB on a small board with dual NVMe and field-ready I/O (GPIO/CSI/CAN/PoE). Suited for: Fastest prototyping of Robotics/Surveillance/Industrial inference with JetPack. Not suited for: Casual “SD card boot” trialing (MicroSD support is tricky).
Verify the DP spec discrepancy and MAXN_SUPER setup requirements. It’s for those who can handle the “Procedure + Firmware + Thermal” trio.
| Retailer | Price (Ref) |
|---|---|
| Arace Tech | US$499 (Arace Tech) |
Previous Article
GMKtec EVO-T1: Empowering the AI Era with Core Ultra 9 285H
Next Article
83 TOPS "Pocket-Class" Edge AI Dev Kit! Checking UP Xtreme ARL AI Dev Kit Specs
Radxa has published the product page and official documentation for the "Radxa Cubie A7Z." This ultra-compact SBC packs an Allwinner A733 (Cortex-A76×2 + Cortex-A55×6) and a 3 TOPS-class NPU into a 65×30mm form factor, complete with Wi-Fi 6 and Micro HDMI. However, wired LAN and M.2 require expansion, and PCIe is limited to Gen3 x1 via FPC, so checking whether it fits your use case beforehand is recommended.
Radxa has published the "Radxa Dragon Q6A" on their official page and Product Brief (Rev 1.4, updated October 24, 2025). This industrial SBC packs Qualcomm QCS6490, Wi-Fi 6, Bluetooth 5.4, 3 MIPI CSI ports, and M.2 2230 NVMe into an 85×56mm form factor. However, AI performance and cooling requirements depend on OS/SDK and implementation conditions, so checking the official Docs before deployment is recommended.
Banana Pi has released product information for their AI-focused core module "Banana Pi BPI-SM9". Equipped with SOPHGO BM1688 and rated at up to 16 TOPS (INT8), the official documentation emphasizes video analytics applications such as decoding and analyzing multiple HD video streams. However, there are inconsistencies in official sources regarding power (5V/12V), dimensions, and GbE port count, so verify whether you are purchasing the module alone, IO board, or kit before ordering.
As of December 29, 2025, we could not find a dedicated product page or official documentation for the "Radxa NIO 5A" from Radxa official primary sources. This article organizes what we could verify about the NIO series positioning from primary sources, how to find purchase options through official distributor listings, and the checklist of unverified items from an SBC perspective.