Expedera, a provider of customizable Neural Processing Unit (NPU) semiconductor IP, introduced the Origin NPUs to support GenAI on edge devices. Origin NPUs offer native support for LLMs, including stable diffusion, unlocking advanced natural language processing capabilities.
Origin NPUs claim to deliver up to 128 TOPS per core with sustained utilization, averaging 80%, surpassing industry norms, and avoiding dark silicon waste.
The patented packet-based NPU architecture addresses issues like memory sharing, security, and area penalties faced by conventional AI accelerators, providing scalability from edge nodes to automobiles.
Expedera provides customizable neural engine semiconductor IP, which improves performance, power, and latency while reducing the cost and complexity in edge AI inference applications across a range of industries, including mobile, automotive, and entertainment.
By using this site, you agree to allow SPEEDA Edge and our partners to use cookies for analytics and personalization. Visit our privacy policy for more information about our data collection practices.