As the demand for AI applications spanning from cloud to edge increases, there is a growing need for high-speed interface IP, low power consumption, and high-performance SoC designs. Faraday is well-equipped to address these requirements with its advanced process technology, comprehensive IP portfolio, and high-bandwidth memory support.

To date, Faraday has successfully completed numerous AI-related projects, including drone vision, medical image analysis, smart appliances, 3D sensing, face and speech recognition, and health status monitoring. These projects have achieved significant power savings, enhanced performance, and reduced system costs, effectively addressing the specific demands of AI applications.

Advanced process technology for AI/HPC

High bandwidth memory Support

To meet the requirement of AI applications for high bandwidth memory access, Faraday supports the memory chips including HBM2 and GDDR6 integrated into a single chip with the package technology of 2.5D or PIP.

Comprehensive IP portfolio

Faraday supports abundant IP solutions for AI applications, including PCIe Gen 3/4/5, SerDes 12/16/25/28/56G, MIPI D-PHY, subLVDS/LVDS, USB, HDMI TX, HBM2, GDDR6, and DDR5/DDR4/3/LPDDR4/3 at advanced FinFET process nodes, which are particularly suited for applications requiring high bandwidth, low power consumption, and low latency interfaces. In addition, Faraday provides the customization service for shrinking the macro area of embedded memory, and implements the self-repair scheme for better SRAM yield in mass production.