Artificial intelligence is increasingly moving to the edge, demanding embedded platforms that combine real-time compute, flexibility, and hardware efficiency. To match these needs, iWave is proud to introduce the iG-RainboW-G58M, a compact System-on-Module (SoM) powered by the Agilex 5 FPGA, a device engineered with AI capabilities built directly into its fabric.
Why Agilex 5 SoM Is Ideal for AI at the Edge
The Agilex 5 Series integrates an AI tensor block, making it one of the first FPGAs designed specifically for inference and AI workloads. This SoM brings together high-performance, low-latency custom logic, and machine learning acceleration — all in a rugged and flexible package.
Further, it features a heterogeneous processing architecture:
- Dual Arm Cortex-A76 cores for high performance
- Dual Cortex-A55 cores for power-efficient tasks
It also supports a rich set of I/O interfaces: high-speed transceivers (up to ~28 Gb/s), PCIe Gen 4, DisplayPort / HDMI, and more.
Software & AI Framework Support
The Agilex™ 5 System on Module is designed to make AI model development and deployment much easier for both software developers and FPGA engineers. To achieve this, the SoM supports popular, industry-standard AI frameworks such as TensorFlow and PyTorch, allowing data scientists to train and refine neural network models using familiar tools.
For deploying these models on FPGA hardware, the SoM supports OpenVINO™, Intel’s widely used optimization toolkit. OpenVINO converts trained models into highly efficient inference graphs that run seamlessly on FPGA-based accelerators, significantly improving performance and reducing latency.
In addition, iWave provides access to the Intel FPGA AI Suite integrated with Quartus Prime, Intel’s FPGA development environment. This suite contains pre-built libraries, optimized kernels, model compilers, and toolchains that simplify the process of mapping AI workloads onto the FPGA fabric. With this combination, developers can:
- Build and train models using mainstream AI frameworks
- Optimize and convert models for FPGA acceleration using OpenVINO
- Implement, fine-tune, and deploy high-performance inference pipelines on Agilex 5 hardware using the Intel FPGA AI Suite
This end-to-end software ecosystem ensures that teams can move quickly from model design to real-world deployment, without needing deep FPGA expertise.
Evaluation Kit & Developer Support
iWave is shipping early samples of the Agilex 5 SoM along with an evaluation kit. The kit includes full user documentation, drivers, and a board support package (BSP) so that developers can begin evaluating and prototyping with minimal friction.
Use Cases & Applications
This Agilex 5-based SoM is well-suited for:
- Edge AI Devices — enabling AI inference in robotics, smart cameras, and industrial automation
- Medical Imaging — where low-latency, real-time data processing is critical
- Vision Systems — leveraging the FPGA’s custom logic, high-speed I/O, and AI capabilities
Longevity & Services
iWave is committed to long-term support: the Agilex 5 SoM is part of a product longevity program, ensuring availability for 10+ years.
Along with the module, iWave offers:
- ODM/design services (carrier board design)
- Thermal simulation
- Mechanical design support
- Customized software support and integration
Conclusion
With the iG-RainboW-G58M, iWave delivers a cutting-edge FPGA SoM tailored for AI at the edge — combining powerful logic, AI tensor acceleration, and flexible interfaces. Whether you’re building robotics, imaging systems, or next-gen industrial devices, this Agilex 5 platform can help you innovate faster and more efficiently.
Contact mktg@iwave-global.com to learn how the Agilex 5 SoM can boost your next AI-driven project.
