Legacy Product


Tera+ is an GPU (Nvidia TX1/TX2) accelerated network switch for edge computer vision processing.

Distributed Intelligence

By connecting multiple TERA+ together, total computational power per second is increased. By setting up multiple TERA+ at various locations and connecting them together (via physical connections or your intranet), you get your very own Distributed-Intelligence Network System. We have seen some clients connect multiple TERA+ to create a little stack, and some clients creating a distributed-intelligence network system with multiple TERA+.

Multiple Camera Support

TERA+ can support up to eight cameras, where images are captured from each camera and processed in parallel.

Ease of Image Capturing

If you are capturing and storing 24 – 60 frames per second from the camera, real-time image processing will be an problem. SmartCow has developed drivers to bypass this step, allowing you to perform real-time computation with ease

Parallel Processing

The beauty of the GPU is that it is able to perform computational tasks in parallel. We have written the camera drivers as such, that it reads the input from all eight cameras in parallel. This is contrary to the traditional method where the inputs of the camera are read in series, one camera at a time. With our managed switch system, you can assign each camera to perform various tasks and neural network models.