Empower Your Infrastructure for AI Solution Deployment

AI-customized inferencing card with 12 TPUs designed in Saudi Arabia for delivering high performance use for neural network workload at production phase which takes you to the next step from AI-possible to AI-proven.


AI-Accelerator PCIe Card

Less Energy, High Efficiency

Deer ITC card

  • ASIC design for AI inference acceleration.
  • Multiple AI models run in parallel.
  • Support up to 48 streams.
  • Achieve low latency with AI models pipelining.
  • Low power consumption of just 21 Watts.

VoltronX Provides AI Acceleration with

High Computing Capability

High Computing Capability

Lower Power Consumption

Lower Power Consumption



Support Wide Temperature

Support Wide Temperature

Simultaneous ML Inferences with Low Latency

Operate multiple AI models in parallel.

In situations where multiple models need to be run, you can assign each model to a dedicated Edge TPU, enabling them to run in parallel, thus achieving peak performance.

Corporate Images


Support TensorFlow Lite machine learning framework

Easy “Plug & Play” installation

Compliant with any motherboard

Compliant with PCIe Specification 2.0 x 16 expansion slot

Boosting ML Performance through Model-Pipelining Innovation

card Images

In scenarios demanding rapid responses or the execution of large models, our pipelining technology allows you to break down models into multiple smaller segments.
Segment and Execute: Distribute smaller models across various Edge TPUs.
Swift Responsiveness: Enhance throughput in high-speed applications.
Latency Reduction: Minimize overall latency for large models.

Do more with less energy

Designed with energy efficiency in mind, VoltronX Card is equipped with excellent thermal stability to achieve inference acceleration with multiple Edge TPUs.

Simultaneous inferencing

Execute your models concurrently on a single or multiple Edge TPU by co-compiling the models so they share the Edge TPU scratchpad memory.

Deploy your AI Model Easily on VoltronX

card Images

Lite converter : Converts TensorFlow models (with .pb file extension) to TensorFlow Lite models (with .tflite file extension).
Compiler : A command-line tool that compiles a TensorFlow Lite models (with .tflite file extension) into files that can be run on an Edge TPU.
Deploy : To execute AI models via PyCoral API (Python) or Libcoral API (C++).

ML Model Requirements

ML Framework Support

TensorFlow Lite

Model Conversion

Tensorflow parameters are Quantizated ( 8 bit fixed-point numbers, int8 )


TensorFlow model to TensorFlow lite model via TensorFlow converter tool

On-Device Intelligence for a Wide Range of Applications

Image segmentation

Identifies various objects in an image and their location on a pixel-by-pixel basis.

Key-phrase detection

Listens to audio samples and quickly recognizes known words and phrases

Object detection

Draws a square around the location of various recognized objects in an image

Pose estimation

Estimates the poses of people in an image by identifying various body joints

Technical Specifications

card Images


Main Chip Core Google® Coral Edge TPU Processor
Interface Technology PCI Express 2.0 x16
Supported Framework TensorFlow Lite
Software Precision INT8
Performance 48 TOPS
Thermal Solution
Power Power Consumption 24 W
Operating System Windows, Linux, OSX Verified by Google
Operating Temperature 0~55°C
Environment Non-Operating Temperature -40~85°C
Relative Humidity 0%~85%
Width 8.1mm
Dimension Height 126.3mm
Depth 186.3mm
Weight Weight 203 g

Experience AI Cutting-Edge Solutions Now