VeriSilicon Unveils Configurable Nano-XR Chip Subsystem for AI Glasses
VeriSilicon's November 18, 2026 launch pairs Nano-XR processors with DECNano compression and FLEXA on-chip buffering to cut memory traffic in AI and AR glasses.
By Nathan Brooks
2 min read

What's News
- VeriSilicon unveiled its Nano-XR wearable chip subsystem for AI/AR glasses on November 18, 2026.
- DECNano compresses data bound for external PSRAM/DDR; FLEXA buffers intermediate data in on-chip SRAM.
- The Nano-XR series has four designs covering imaging, IR sensing, graphics and 3DoF/6DoF display processing.
- CSO Weijin Dai said customers are already collaborating on dedicated smart-glasses chips with Nano-XR IP.
- The announcement does not quantify power, bandwidth or latency savings.
VeriSilicon introduced a configurable, low-power chip subsystem for AI and augmented-reality glasses on November 18, 2026, pairing new Nano-XR processor designs with compression and data-transfer technology aimed at external-memory bottlenecks.
The Shanghai-based semiconductor IP supplier's pitch is straightforward: glasses-class devices can't afford the power budget of discrete memory accesses. So the subsystem attacks two costs at once. DECNano applies what VeriSilicon calls visually lossless compression to shrink the capacity and bandwidth demanded from external PSRAM and DDR memory. FLEXA handles the other side, moving intermediate data between processing blocks at the segment level and buffering it in on-chip SRAM.
Together, the company says, the two techniques reduce external-memory accesses and end-to-end latency. The announcement does not quantify the actual power, bandwidth or latency savings, so buyers will want benchmark data before committing.
What does the Nano-XR lineup actually cover?
Four new processor designs extend VeriSilicon's existing low-power Nano IP portfolio into glasses-specific territory:
- ISPNanoUltra-XR — image signal processing that VeriSilicon says delivers smartphone-level photo and video quality while optimizing power consumption and silicon area.
- ISPPico-XR — built for infrared sensors and always-on functions including eye tracking and gesture recognition.
- GCNanoOptimaV-XR — graphics rendering that improves projection transformation and performance over the previous GCNanoUltraV while cutting memory use.
- DCNanoUltra-XR — display processing combining low-latency 3DoF and 6DoF handling with power-gating strategies and multiple high-speed display interfaces.
The subsystem's processing scope spans 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, tuning how data moves between functions as well as each block's power and performance profile.
How flexible is the configuration?
Beyond Nano-XR, customers can configure graphics, neural-network, video, digital-signal, image-signal and display-processing IP against different power and performance targets. The neural-network processors cover low-power wake-up, image enhancement and general-purpose AI computing. The DSPs handle voice processing, noise reduction, audio codecs and Cat.1/Cat.4-related wearable applications.
That range lets a licensee build anything from basic low-power glasses up to advanced AR devices with richer computing and display capabilities. Software support spans real-time operating systems, Linux and Android, so the same design carries across different environments without a port.
Who's building with it?
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 tie the subsystem launch to real silicon programs rather than a catalog release, giving the more advanced AI and AR applications VeriSilicon targets a concrete development path.
For device makers weighing a glasses platform, the open question is measurable: VeriSilicon hasn't published numbers on the memory, latency or power gains. Expect customer chip announcements to surface those figures first.
Source: unite.ai
More from Nathan Brooks
Show full bio
News editor covering regulation, policy and their impact on entrepreneurship, startups and venture funding at Business Bearings.
2,352 articles