SAN JOSE, CA, May 1, 2026 – Altera announced the release of FPGA AI Suite 26.1.1 to help developers run trained AI models on FPGAs in edge systems, including robotics and autonomous machines. The update adds a spatial compiler that maps AI models onto Agilex silicon and uses streaming dataflow execution for AI inference with low latency and reprogrammable deployment.
The spatial compiler architecture in FPGA AI Suite 26.1.1 maps neural networks onto FPGA hardware instead of relying on sequential processing. The approach delivers ASIC-like performance for enhanced AI inference while preserving FPGA re-programmability for changing workloads. The spatial compiler controls data movement and parallel execution to improve throughput, reduce power use and keep latency predictable.
The software targets edge AI applications such as vision, video analytics, language models and sensor processing in physical AI systems. These systems require low-latency sensor processing, AI inference and responses in changing operating conditions.
Altera supports these applications with AI-enabled Agilex FPGAs and FPGA AI Suite. Agilex FPGAs offer different capacities, peripherals and interfaces for edge AI applications across multiple form factors. The hardware and software toolchain supports FPGA-based AI deployment across the Agilex portfolio, including systems with safety, security and reliability.
“FPGA AI Suite 26.1.1 reflects our broader strategy to make FPGA-based AI implementation more accessible and scalable,” said Venkat Yadavalli, head of Altera’s business management group. “As AI moves closer to the edge, developers need easy-to-deploy solutions that combine performance, efficiency and flexibility. With FPGA AI Suite, customers can scale seamlessly across Altera’s Agilex portfolio and target a broad range of end applications and use cases with a unified development approach.”
About FPGA AI Suite
FPGA AI Suite is a software platform for AI development and deployment on FPGA-based systems. Developers can use FPGA AI Suite with PyTorch, TensorFlow and model optimization capabilities from OpenVINO. The software helps run AI models with low latency and deterministic performance. FPGA AI Suite supports applications from edge systems to data centers.
Availability
FPGA AI Suite 26.1.1 is available for download and supports Quartus Prime Pro Edition 26.1. FPGA AI Suite supports development with license-free operation for up to 100,000 consecutive inferences.
Source: Altera
About Altera
Altera is a semiconductor company founded in 1983 and based in San Jose, CA. The company designs programmable logic devices, including field-programmable gate arrays and complex programmable logic devices, along with related software tools. Engineers use these products to build configurable hardware for communications, computing, industrial, automotive and aerospace systems. Altera serves system designers, equipment manufacturers and technology companies that integrate programmable logic into network infrastructure, data centers and embedded applications. The company became part of Intel’s solutions business and later reestablished operations with majority investment from Silver Lake, while Intel retained a minority stake. Altera serves about 14,000 customers worldwide and employs approx. 3,000 people across North America, Europe and Asia. The company supplies chips and development tools that support flexible hardware design in electronic systems.