
Forlinx FET3572-C: First Rockchip RK3572 System-on-Module Ships With Linux 6.12
Forlinx launched the first Rockchip RK3572 system-on-module and OK3572-C dev board on June 4, 2026 — an octa-core edge-AI building block with a 4 TOPS NPU and Linux 6.12 BSP.
A Fresh Rockchip SoC Lands on Its First System-on-Module
Forlinx launched the FET3572-C on June 4, 2026, the first system-on-module built around Rockchip's new RK3572 SoC, alongside a matching OK3572-C development board. Best of all for makers and industrial integrators, it ships with a current Linux 6.12 BSP out of the gate — fresh kernel support on day one is exactly what you want from a new single-board computer platform.
The FET3572-C SoM pairs an octa-core CPU with a built-in NPU. The CPU combines two Arm Cortex-A73 cores at 2.2 GHz with six Cortex-A53 cores at 2.1 GHz, while a 4 TOPS INT8 NPU handles on-device AI inference and an Arm Mali-G310V2 MC1 GPU drives graphics. Memory options span 2GB, 4GB, or 8GB of LPDDR5, with up to 64GB of eMMC storage. The SoM also supports Android 16 for teams that prefer that stack.
What the OK3572-C Dev Board Exposes
The OK3572-C carrier board breaks out an impressively complete set of interfaces for edge work. It offers dual Gigabit Ethernet, HDMI 2.1, multiple MIPI CSI-2 camera inputs, USB 3.0, PCIe 2.1, a microSD slot, optional NVMe storage, and optional 5G or 4G LTE connectivity.
Multi-Camera Support Is the Standout Feature
The multiple MIPI CSI-2 camera inputs are the detail that makes this board genuinely useful for edge AI. Combined with the 4 TOPS NPU, the OK3572-C can ingest several camera streams at once and run local vision inference on them — ideal for smart security, industrial inspection, and in-vehicle terminals. Dual Gigabit Ethernet adds networking flexibility for gateways and multi-node setups.
A Well-Supported Building Block for Edge AI
Forlinx positions the FET3572-C for edge AI, industrial control, smart security, and in-vehicle terminals, and the combination of fresh silicon and a Linux 6.12 BSP is what makes it attractive. A new SoC paired with up-to-date mainline-leaning kernel support means fewer driver headaches and a longer useful life for products built on it. For developers who want an affordable, multi-camera edge-AI module they can design into real hardware, the first RK3572 system-on-module is a promising starting point.
Sources: CNX Software (June 4, 2026); Rockchip RK3572 platform documentation (2026).
More Mini Computers Stories

ESP32-P4 Desktop OS: Sylphide OS and BMDS Turn MCUs Into PCs
Two open-source desktop OS projects, Sylphide OS and BMDS, give the $25 Olimex ESP32-P4-PC windows, apps, SSH and even Doom at 1280x720 and 60 FPS.

Tektronix 1 Series Oscilloscope: $450 Scope With AI Control
Tektronix's 1 Series DPO starts at $450 with 12-bit resolution, up to 200 MHz, I2C/SPI/UART decoding and an AI connector for natural-language control.

Eight Used V100 GPUs: How a Lawyer Built a Private AI Server
A lawyer's private AI server runs DeepSeek V4.1 Flash on eight used V100 GPUs at up to 215 tokens per second, keeping case files off the cloud.
