Flex Logix Announces InferX™ High Performance IP for DSP and AI Inference

Integrate Leadership AI in your SoC

Very high performance batch=1 Vision AI.

Low power and low cost due to InferX’s efficient architecture and software.

High accuracy using INT8, INT16 & BF16 as needed.

Reconfigurable architecture efficiently runs every new YOLO model & even Transformers. New operators are implemented accurately & efficiently.

Run multiple models.

Easy to integrate into your SoC and Software Stack.

Flex Logix has a proven track record in dozens of chips.

The first generation of InferX is proven in silicon and production qualified.