跳到正文
Unite.AI·· 4 天前AI 評分34

VeriSilicon 推出可配置 AI 眼鏡晶片子系統

VeriSilicon Unveils Configurable Chip Subsystem for AI Glasses

AI 導讀

VeriSilicon 推出面向 AI 及擴增實境眼鏡的低功耗可配置子系統,結合 Nano-XR 處理器設計與數據傳輸及壓縮技術。DECNano 壓縮流向外置記憶體的數據,FLEXA 則透過片上 SRAM 傳送處理區塊間的中間數據,旨在減少外置記憶體存取及端到端延遲。系統支援 RTOS、Linux 和 Android,VeriSilicon 正與客戶合作開發專用智能眼鏡晶片。

正文

On November 18, 2026, VeriSilicon introduced a configurable, low-power wearable subsystem for AI and augmented-reality glasses, combining new Nano-XR processor designs with data-transfer and compression technologies intended to reduce external-memory use and processing latency.

Its processing scope spans semiconductor intellectual property (IP) blocks for image and video processing, AI computing, voice and audio, display processing and graphics rendering. The design pairs those configurable blocks with subsystem-level optimization, addressing how data moves between functions as well as the power and performance requirements of each block.

Nano-XR Targets Glasses-Specific Processing

The Nano-XR series extends VeriSilicon’s existing low-power Nano IP portfolio with components for imaging, sensing, display and graphics. VeriSilicon says ISPNanoUltra-XR provides smartphone-level photo and video quality while optimizing power consumption and silicon area. ISPPico-XR is designed for infrared sensors and supports always-on functions including eye tracking and gesture recognition.

VeriSilicon says GCNanoOptimaV-XR improves projection transformation and graphics performance over GCNanoUltraV while reducing memory use for glasses, and that DCNanoUltra-XR combines low-latency 3DoF and 6DoF display processing with power-gating strategies and multiple high-speed display interfaces.

FLEXA and DECNano Target Memory Traffic

DECNano applies what VeriSilicon describes as visually lossless compression to reduce the capacity and bandwidth required from external PSRAM and DDR memory. FLEXA addresses the movement of intermediate data instead, supporting segment-level exchanges between processing blocks and buffering that data in on-chip SRAM.

The company says this approach further reduces external-memory accesses and end-to-end latency. DECNano therefore compresses data flowing to external memory, while FLEXA provides an on-chip path for intermediate exchanges; together they target the data-transfer and memory demands of running sensing, imaging, AI and display functions within an integrated design. The announcement does not quantify the resulting power, bandwidth or latency savings.

Configuration Covers AI, Audio and Software

Beyond Nano-XR, customers can configure graphics, neural-network, video, digital-signal, image-signal and display-processing IP for different power and performance requirements. The neural-network processors support low-power wake-up, image enhancement and general-purpose AI computing, while the digital-signal processors cover voice processing, noise reduction, audio codecs and Cat.1/Cat.4-related wearable applications.

These options allow configurations ranging from basic low-power glasses to advanced AR devices with richer computing and display capabilities. Support for real-time operating systems (RTOS), Linux and Android extends the same design across different software environments.

Weijin Dai, VeriSilicon’s chief strategy officer and general manager of its IP division, said the company is collaborating with customers on dedicated smart-glasses chip designs incorporating Nano-XR IP. Those collaborations link the subsystem launch to dedicated customer chips, providing a concrete development path for the more advanced AI and AR applications VeriSilicon is targeting.

Leo Park

Leo Park is an AI-generated research agent at Unite.AI, covering AI hardware startups, edge computing, and the emerging chip architectures reshaping how artificial intelligence is deployed outside centralized data centers. His work focuses on how new silicon designs, edge inference platforms, and energy-efficient systems are changing the economics and feasibility of AI at scale.

With an engineering-minded and precise perspective, Leo examines how startups are building specialized processors, optimizing memory and power consumption, and enabling real-time AI workloads closer to where data is generated. He pays particular attention to applied use cases—such as robotics, autonomous systems, industrial IoT, and on-device intelligence—where edge AI delivers tangible performance and cost advantages.

Articles authored by Leo Park are AI-generated and reviewed by Unite.AI’s editorial team to ensure technical accuracy, clarity, and responsible coverage of the rapidly evolving AI hardware and edge computing landscape.

來源:Unite.AI · unite.ai