VeriSilicon introduces Nano-XR chip subsystem for AI glasses

VeriSilicon introduced a configurable wearable chip subsystem for AI and augmented-reality glasses in a November 18 announcement, combining its new Nano-XR processing designs with memory and data-transfer optimizations.
The subsystem brings together image capture, AI computation, audio, display processing and graphics rendering. Customers can select processing components for different power and performance requirements, with support for RTOS, Linux and Android.
Nano-XR includes designs for photo and video processing, infrared sensing for functions such as eye tracking, low-latency display processing and graphics transformations. The announcement describes components for chip designers, not a finished pair of consumer glasses.
VeriSilicon’s FLEXA interface exchanges data between processing blocks using on-chip memory to reduce external-memory transfers. Its DECNano technology compresses data to reduce memory capacity and bandwidth requirements. Claimed reductions in latency and power are the company’s characterization, not independently tested results in the announcement.
The company says it is collaborating with customers on dedicated smart-glasses chips. The design targets a practical constraint for wearable AI: coordinating sensing, computation and display within a limited power budget.
Sources and citations
Revision history
Published .