Altera Releases New AI Software to Accelerate Edge AI Deployment
2026-05-02 17:50
Favorite

en.Wedoany.com Reported - Altera recently announced the release of FPGA AI Suite version 26.1.1, a major update to its AI software platform. The suite is designed to simplify the process of deploying trained AI models onto FPGAs, specifically tailored for edge AI applications, providing support for physical AI systems such as robotics and real-time autonomous machines. The new version introduces a spatial mapping-based compiler technology that delivers ASIC-like performance for optimized AI inference, while maintaining rapid time-to-market, reprogrammability, and the deterministic low-latency execution inherent to FPGAs.

FPGA AI Suite 26.1.1 Spatial Compiler Architecture Diagram

The spatial compiler architecture in FPGA AI Suite 26.1.1 can map neural networks directly onto FPGA hardware, employing a streaming dataflow approach rather than traditional sequential processing. By optimizing data movement and parallel execution, this architecture enhances efficiency, achieving higher throughput, lower power consumption, and deterministic low latency, making it particularly suitable for real-time edge applications such as vision, video analytics, language models, and sensor processing in physical AI systems.

Edge AI is transforming how physical AI systems, such as robots and autonomous machines, interact with the real world, enabling them to perceive, think, and act in real-time within dynamic environments. These systems require high-performance, low-latency sensor processing, efficient AI inference, and deterministic responses to ensure safe and intelligent operation.

Altera supports these applications through its combination of AI-enabled Agilex FPGAs and the FPGA AI Suite, delivering deterministic low-latency performance and the flexibility to adapt to evolving AI workloads. By combining adaptable hardware with optimized AI development tools, developers can build and deploy scalable, real-time edge AI solutions. Agilex FPGAs offer a variety of capacities, peripherals, and interfaces to meet the needs of different edge AI form factors, and provide built-in security, safety, and reliability for next-generation physical AI systems.

FPGA AI Suite 26.1.1 is now available for download and supports Quartus Prime Pro Edition 26.1. This version continues to support early development, allowing up to 100,000 consecutive inferences to run for free, lowering the barrier to adoption.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com