Summary

SRG just completed its 56th 5G benchmark study. For this endeavor, we used a pair of Meta Ray-Ban Display glasses and two different Samsung Galaxy smartphones (S25 Ultra and S23) to evaluate how various AR and AI applications impact cellular data AND signaling traffic. We also looked at how various AI applications that run natively on a smartphone impact battery life / current drain. Highlights of the Report include the following: Our Thanks. We did this study with the support of Accuver (XCAL5, XCAL-Mobile and XCAP). We used XCAL5 to collect chipset data and we used XCAP to post process and analyze those logs. We used XCAL-Mobile and other Android applications to measure and record battery current and CPU usage. The Scope. We tested all applications supported by the Meta Ray-Ban Display glasses with the Galaxy S23 and S25 Ultra smartphones. We also included several other AI / AR applications, including Lens, CoPilot, Amazon Shopping, Doppl, Maps Navigation, Sora, Nano Banana, and Ikea. We looked at downlink/uplink bandwidth when performing various tasks and RRC signaling / connection time when the phone/glasses were supposedly not being used. For native AI applications, we used Meta AI translate, Galaxy AI, and two different LLMs from Google AI Edge Gallery. For these tests, we focused on CPU usage and current battery drain. The Results. Although AI and AR increase uplink data traffic, the impact is inconsequential with meaningful improvements to the user experience possible without further increasing data traffic. The same cannot be said for the impact of Meta AI and the Meta glasses on signaling and RRC connection time, even when the phone and glasses are not being actively used. AI applications that run natively on the smartphone can dramatically increase battery current, but likely not an overall killer of battery life unless these applications are frequently used throughout the day.