Lightspeeur™ is the first Neural Processor over the world to run speech, Natural Language Processing (NLP), image and video processing with Convolutional Neural Networks (CNN).

With superior Energy Efficiency (TOPS/w) level, we have accelerated AI Edge Computing and Datacenter Machine Learning up to several orders of magnitude faster than the competition. Regardless of whether the implementation is in training or inference mode, our unique, proprietary architecture provides massive high-density processing power. Our design supports true, on-chip parallelism, in situ computing, and eliminates memory bottlenecks.

The Lightspeeur™ Neural Processor supports neural networks like CNN, RNN and LSTM. It also supports standard, open frameworks such as Caffe, TensorFlow, and MXNet. Software Development Kits include turn-key designs, system verification hardware, software and tools.


Mobile Edge Computing

Smart surveillance/
video cam

Smart toys/robotics

Smart home

AR/VR products

Face detection/

Speech/voice recognition

Natural Language Processing

Deep Learning enabled devices

Cloud ML/DL System

AI Data Center servers

ADAS/Autonomous driving