In 2025, Forlinx Embedded continued to strengthen its role in the global embedded ecosystem through long-term silicon partnerships and reliable industrial-grade platforms. As an NXP Gold Partner and official i.MX 95 ecosystem partner, Forlinx works closely with NXP to support secure, safety-ready edge computing. At the same time, deeper cooperation with Rockchip and Allwinner Technology has expanded Forlinx’s portfolio of cost-effective solutions for edge AI and industrial control.

Throughout the year, Forlinx Embedded showcased these platforms at major industry exhibitions—from Nuremberg to Guangzhou—demonstrating how advanced processors can be transformed into stable, deployable edge AI systems.

Below is a recap of Forlinx Embedded’s key products and milestones in 2025.

Flagship Platforms for Industrial Edge & AI

FET-MX9596-C System on Module (NXP i.MX 95)

The FET-MX9596-C represents Forlinx Embedded’s next-generation high-end industrial SoM, designed for demanding edge applications requiring sophisticated processing and hardware security.

• Processor Core: Built on NXP’s i.MX 95, featuring a heterogeneous architecture—Six ARM Cortex-A55 application cores. One Cortex-M7 real-time core. One Cortex-M33 security core.

• AI & Graphics: Integrates a 2 TOPS NPU for AI acceleration and an ARM Mali-G310 GPU for 3D/2D graphics.

• Vision Support: Built-in ISP supporting 4K@30fps video and dual MIPI-CSI cameras.

• Target Use Cases: Well-suited for industrial controllers, robotics, smart energy, and safety-critical edge systems, leveraging NXP EdgeLock security features and rich industrial interfaces.

FET1126B-S/FET1126BJ-S System on Module (Rockchip RV1126B/RV1126BJ)

The FET1126B-S/FET1126BJ-S SoM is a low-power, cost-effective solution optimized for edge-side AI computing.

• Processor Core: Rockchip RV1126B/RV1126BJ (4x [email protected]/1.3GHz)

• Vision & Intelligence: Includes a 3 TOPS NPU with RKNN support, 4K video decoding, and a 12M ISP. These features enable efficient execution of AI models for face detection, safety helmet recognition, fire/smoke alarms, and area intrusion detection.

• Industrial Ruggedness: Supports -40°C~+85°C operation, delivering excellent reliability and environmental adaptability in complex settings. Rigorously tested by Forlinx Embedded Laboratory for industrial-grade stability, it ensures reliable performance with a 10-15 year longevity for consistent long-term supply.

• Target Use Cases: Optimized for power transmission, intelligent transportation, gas station monitoring, smart construction sites, and intelligent warehousing.

FET-MX8MPQ-SMARC System on Module (NXP i.MX 8M Plus)

The FET-MX8MPQ-SMARC is an adaptive SoM optimized for AI-enabled HMI and vision-driven edge applications, utilizing the standardized SMARC 2.1 form factor.

• Processor Core: NXP i.MX 8M Plus Quad-Plus (4× [email protected] + Cortex-M7@800 MHz).

• AI & Vision: Includes a 2.3 TOPS NPU and ISP to deliver powerful machine learning, vision, and advanced multimedia processing.

• Form Factor: Built to the SMARC 2.1 standard (82 × 50 mm), offering plug-and-play compatibility with SMARC carriers.

• Target Use Cases: Provides a balanced platform for smart manufacturing, medical devices, and intelligent terminals requiring high-reliability compute and long-term availability.

FCU3501 Embedded Computer (Rockchip RK3588)

The FCU3501 is a high-performance, rugged, and fanless embedded computer, delivered as a ready-to-deploy solution for intensive edge AI workloads.

• Processor Core: Rockchip RK3588, featuring an octa-core architecture (4 x Cortex-A76 up to 2.4GHz + 4 x Cortex-A55 up to 1.8GHz).

• AI Performance (32 TOPS NPU): Offers native 6 TOPS NPU performance, with optional AI acceleration via a Hailo-8 M.2 module (26 TOPS).

• Multimedia: Supports 8K video encoding/decoding.

• Industrial Design: Fanless, passive cooling design ensures stable, reliable operation, verified through rigorous environmental and EMC testing.

• Target Use Cases: Ideal for demanding edge scenarios such as machine vision, intelligent transportation, and AI gateways—delivering performance without custom carrier development.

Cost-Effective & Scalable Industrial Platforms

Rockchip RK3506 Series SoMs

The RK3506J is a high-performance, triple-core Cortex-A7 application processor known for its excellent power consumption and heat dissipation, specifically designed for intelligent industrial control and IoT.

(1) FET3506J-C System on Module

Designed for board-to-board connectivity, the FET3506J-C offers a robust industrial SoM solution with low power consumption and rich I/O. It fits well into applications that require stable mass production, long lifecycle support, and tight mechanical integration.

• 0.7W ultra-low power, supports fanless design and +85°C industrial temperature.
• Compact 29×40mm size with 2mm connector, ideal for small embedded devices.
• Rich industrial I/O, including 2× CAN-FD, UART, RMII, and display interfaces.
• High-speed DSMC bus, enables fast FPGA-to-ARM communication.
• Flexible FlexBUS interface, supports ADC, DAC, camera, QSPI LCD.
• Linux 6.1 + LVGL 9.2, supports AMP and real-time system development.
• 10~15 years supply, ensuring stability for long-life industrial projects.

(2) FET3506J-S System on Module

Featuring stamp-hole design, the FET3506J-S is optimized for cost efficiency and compact designs, supporting rapid prototyping as well as large-scale deployment. It is particularly suitable for industrial control, gateways, and IoT nodes where space and BOM cost matter.

• Edge connector, all pins are led out
• 22nm advanced process
• Display interfaces: MIPI DSI, RGB
• Rich industrial bus interfaces: RMII, CAN-FD, FLEXBUS, DSMC etc
• DSMC can be used to extend PSRAM, FPGA communication
• RM_IO enables matrix configuration for pin functions

(3) OK3506-S12 Mini SBC

The OK3506-S12 Mini SBC is built for rapid evaluation and cost-sensitive development. Compact and developer-friendly, it enables fast function verification and early-stage testing for RK3506-based applications, making it ideal for proof-of-concept projects and entry-level edge designs.

FET536-C System on Module (Allwinner T536)

The FET536-C is a cost-effective, real-time focused SoM based on the Allwinner T536 industrial-grade processor.

• Processor Core: Quad-core Cortex-A55 @ 1.6 GHz and the XuanTie E907 RISC-V core.
• Real-Time & Security: Supports 2 TOPS NPU, secure boot, full-path ECC, AMP, and Linux-RT for deterministic performance.
• Connectivity: Wide range of interfaces including CAN-FD, Ethernet, ADC, and LocalBus.
• Target Use Cases: An excellent choice for cost reduction in key fields such as concentrators, FTU, DTU, charging piles, transportation, robotics, and industrial control that demand reliable, real-time control over raw compute performance.

FET153-S System on Module (Allwinner T153)

The FET153-S is an industrial SoM targeted for intelligent automation applications, such as programmable logic controllers (PLCs) and human-machine interface (HMI).

• Processor Core: Allwinner T153 (4x [email protected]+RISC-V E907@600MHz).
• Vision: Its integrated image signal processor and display engine enable clear real-time visual feedback for managing intricate manufacturing processes.
• Excellent connectivity: It features three Gigabit Ethernet interfaces, two CAN_FD interfaces, and 16-/32-bit local bus, supporting high-throughput networking for complex data-driven use cases.

Special Platform: Ecosystem & Evaluation Support

NXP FRDM i.MX93

As part of the broader NXP ecosystem, the FRDM-IMX93 board is a low-cost and compact platform designed to show the most commonly used features of the i.MX 93 Applications Processor in a small and low cost package. The FRDM-IMX93 board is an entry-level development board, which helps developers to get familiar with the processor before investing a large amount of resources in more specific designs.

Ecosystem Collaboration, Shared Growth

Collaboration and mutual success remained a core theme for Forlinx Embedded in 2025. Throughout the year, we worked closely with leading silicon vendors including NXP, TI, Rockchip, Allwinner, and Nuvoton, co-hosting ecosystem seminars, joining distributor workshops to share critical technical insights.

These efforts were recognized with honors such as Rockchip Annual Outstanding Partner and Allwinner Certified Ecosystem Partner. In parallel, we partnered with industry players like Phoenix Contact to deliver integrated hardware-software solutions, accelerating intelligent transformation across industrial sectors.

Global Footprint, Strengthening Industry Presence

From Nuremberg, Germany to Guangzhou, China, from early spring to late autumn, Forlinx Embedded participated in seven major industry exhibitions in 2025. By engaging with partners across regions and industries, we shared market insights and innovation trends, delivering more robust, intelligent, and scalable embedded solutions for diverse applications.

Through initiatives such as the Edge AI Technology Forum and Embedded Technology Salons, we used technology as a bridge—bringing customers and partners closer to cutting-edge AI and embedded system innovations.

Deepening AI Deployment at the Edge

In 2025, Forlinx Embedded accelerated the adoption of AI at the edge across real-world applications. From driver fatigue detection and battery life prediction, to on-device deployment of DeepSeek and multimodal AI fusion, we embedded intelligence directly into control platforms.

By focusing on real-time performance, low power consumption, and high reliability, we continue to co-build a smarter future together with partners across the embedded and AI ecosystem.

Looking Ahead

Across all 2025 releases, a consistent philosophy stands out: long product lifecycles, industrial-grade reliability, and scalable hardware platforms that reduce integration risk. Whether customers are building next-generation AI edge systems or long-running industrial controllers, Forlinx Embedded’s 2025 portfolio provides a clear path from evaluation to mass production—without compromising stability, supply continuity, or maintainability.