From Eyewear to Silicon: RISC-V for Low-Power AI in Next-Generation Smart Glasses

on YouTube ↗30m

" summary

Luxottica builds a RISC‑V based SoC using the PULP ecosystem plus custom IP to hit sub‑35 mW system budgets for 150 mAh glasses running 16 hours. ISA extensions deliver 10x DSP speedup versus baseline RISC‑V, multicore parallelism adds another 10x over single core, the Smart Filtering Unit runs ANC below 5 mW, six‑mic beamforming shows ~10x latency and ~2x power versus multicore, and a reconfigurable integer conv engine gives ~5x end to end energy on object detection.

AI-written summary · about

" chapters (13)

1. 0:00Introduction to smart glasses
2. 1:15Smart glasses vision and applications
3. 5:26Power constraints in smart glasses
4. 7:56Approach to silicon design
5. 10:05RISC-V as the key pillar
6. 12:09Extending RISC-V ISA for efficiency
7. 14:14Parallelization with multi-core RISC-V
8. 15:39Specialized computing for AI
9. 17:03Smart filtering unit for audio
10. 21:03AI inference acceleration
11. 23:55Synergy between components
12. 25:25Conclusion and future outlook
13. 26:36Q&A session

" related talks

• 13mKeynote: Taking RISC-V Intelligence EverywhereCharlie Su · RISC-V Summit North America 2023
• 19mWhy RISC-V?Keith Graham · RISC-V 101 2023
• 22mKeynote: Unlocking Innovation with RISC-V and QualcommZiad Asghar · RISC-V Summit North America 2023
• 10mRISC-V Revolutionizing in AIVince Wu · RISC-V at embedded world 2024
• 40mAdvancing AI Application Development on RISC-V PlatformsYuning Liang · Open Source Summit Europe 2025
• 13mBoosting RISC V ISA With Open Source Peripherals: An SoC For Low Power SensorsElkim Roa · 7th RISC-V Workshop 2017
• 14mKeynote: RISC-V Opportunities at the Edge of AIMakeljana Shkurti, Ed Doran · RISC-V Summit North America 2025
• 01mThe impact of smart wearablesChris Bergey · CES 2026