Edge AI AcceleratorFFI8805 Mini
Ultra-Compact SRAM-CIM Edge AI Chip
An ultra-lightweight AI inference accelerator designed for IoT and smart home. M.2 2242 / USB 3.0 plug-and-play, under 1W power.
0 TOPS
AI Performance
<0W
Power
179 KB
Inference Latency
7mm
Operating Temp
Core Technology
Three Core Advantages
FFI8805 Mini is built on SRAM-CIM architecture for ultra-low-power edge AI inference.
SRAM-CIM Computing-in-Memory
Performs MAC operations directly in memory, eliminating data movement bottlenecks with 2 TOPS performance.
2 TOPS
INT8 Performance
512 KB
SRAM Capacity
Ultra-Low Power Design
Advanced power management with dynamic voltage and frequency scaling, standby power as low as 10mW, ideal for battery-powered IoT devices.
<1W
TDP Power
10mW
Standby Power
Ultimate Edge Efficiency
Run YOLOv5s, MobileNet and other mainstream edge AI models at under 1W power.
<1W
TDP Power
30 FPS
YOLOv5s
Detailed Specifications
Members-Only Technical Data
Log in to view full specifications, supported model list, and application scenarios.