English
sakura-white-blue-icon
Package photo of the EdgeCortix SAKURA-I Edge AI Co-Processor

EdgeCortix SAKURA-I

Energy-efficient Edge AI Co-Processor

Supported Frameworks & Applications

Deep learning accelerator ASIC for fast, efficient AI inference co-processing in boards and systems

EdgeCortix SAKURA-I is a single-chip 40 TOPS implementation of EdgeCortix Dynamic Neural Accelerator IP. Its edge AI co-processor core is run-time reconfigurable using the EdgeCortix MERA compiler and software framework. This enables multiple deep neural network models running concurrently while maintaining critical performance characteristics.

dna-blue-white-logo-v3
High utilization
+80% for concurrent neural network models
edgecortix-high-speed-low-latency-icon
Low latency
< 4 msec when under demanding workloads
edgecortix-power-efficiency-icon
Excellent efficiency
16x inference/sec/W of NVIDIA Jetson AGX

Ready for high-resolution data streams

SAKURA-I is designed for applications needing fast AI inference on streaming data, such as high-resolution video feeds. It is implemented in TSMC 12nm FinFET, packing excellent AI inference performance into a small footprint. Using this advanced process also ensures long-term ASIC availability for extended life cycles typical in defense, industrial, and other application segments.

EdgeCortix SAKURA is an edge AI co-processor built in TSMC 12nm FinFET, delivering 40 TOPS @ 800 MHz in a TDP of 10W
Lithographic view of the EdgeCortix SAKURA-I Edge AI Co-Processor highlighting the Dynamic Neural Accelerator IP, memory, and PCIe interface

Hardware Architecture Overview

  • Up to 40 TOPS (single chip) and 200 TOPS (multi-chip)
  • Run-time reconfigurable datapath
  • PCIe Device TDP @ 10W-15W
  • Typical ASIC power consumption ~5W
  • On-chip 2x64 LPDDR4x – 16 GB
  • PCIe Gen 3 up to 16 GB/s bandwidth
  • JTAG
  • SPI
  • QSPI
  • Temperature Sensor

Powerful AI inference for industrial segments including:

  • Augmented and Virtual Reality
  • Automotive Sensing
  • Communications
  • Defense
  • Smart Cities
  • Smart Manufacturing
The EdgeCortix SAKURA-I achieves over ten times the energy efficiency of GPU-based AI inference solutions at similar levels of TOPS performance
EdgeCortix SAKURA PCIe Dev Card expands a host system with a powerful, yet efficient edge AI chip delivering 40 TOPS

PCIe x16 Development Card

EdgeCortix SAKURA-I is available on a PCIe x16 development card, ready to drop into a host for software development and AI model evaluation tasks.

Get the details in the SAKURA-I datasheet

trust-capitol-co-ltd-logo

EdgeCortix is in a truly unique market position. Beyond simply taking advantage of the massive need and growth opportunity in leveraging AI across many business key sectors, it’s the business strategy with respect to how they develop their solutions for their go-to-market that will be the great differentiator. In my experience, most technology companies focus very myopically, on delivering great code or perhaps semiconductor design. EdgeCortix’s secret sauce is in how they’ve co-developed their IP, applying equal importance to both the software IP and the chip design, creating a symbiotic software-centric hardware ecosystem, this sets EdgeCortix apart in the marketplace.”

Daniel Fujii
President & CEO of Trust Capital Co., Ltd., member of the Executive Committee of Silicon Valley Japan Platform

Learn more about SAKURA-I now