The company claimed that the new AR2G1 is the world’s first purpose-built platform for head-worn augmented reality (AR) devices. Just to be clear, this is a different solution than the existing Snapdragon XR2 Gen 1 and XR2 Plus Gen 1 which were primarily designed for mixed reality (MR) and virtual reality (VR) headsets. [Image: Qualcomm]The AR2G1 on the other hand is made for AR glasses which explains why it is being designed to fit directly inside the glasses’ arm/temples. Among the main objectives of the platform is to deliver AR glasses with under 1W of power consumption despite occupying less space than the XR2 inside Qualcomm’s Wireless AR Smart Viewer Reference Design. The exact details of the new multi-chip solution are not available at the time this article is published but we were told that the AR2G1 was built using 4nm process node and able to deliver 2.5x more AI performance than the XR2G1 while consuming 50% less power. The AR2G1 consisted of three different chips with one of them being the AR processor. [Image: Qualcomm]According to Qualcomm, this particular chip is able to support up to nine cameras simultaneously and it also has a dedicated hardware acceleration engine for motion tracking as well as localization. It also has a reprojection engine as well as an AI accelerator to help reduce latency for critical input interactions such as hand tracking and 6DoF. Aside from that, the AR2G1 also has an AR co-processor that is designed to churn through all the data from cameras and sensors within the AR device while enabling its eye tracking and iris authentication. Essentially, the AR co-processor’s main job is to optimize workload only at the area that the user is looking which helps tone down the AR device’s power consumption. AR2G1’s third chip comes in the form of FastConnect 7800 which is the same connectivity system inside the equally new Snapdragon 8 Gen 2 SoC. This means future AR2G1-based devices are able to support Wi-Fi 7 and it also helps ensure that the latency between the AR device and host device remains under 2ms. [Image: Qualcomm]With that last sentence, you may now notice that AR2G1-based AR devices can’t work on their own though. This is because data that require more complex processing would be offloaded to host machines that can be but are not limited to Snapdragon-based smartphones and PCs. In terms of availability, there is no straightforward answer for AR2G1 though as Qualcomm is currently working with multiple OEMs to deploy on their devices and the development stage that they are on varies from one partner to another. Nevertheless, there are plenty of them around including Lenovo, LG, Nreal, OPPO, Pico, QQNOQ, Rokid, Sharp, TCL, Tencent, Vuzix, and Xiaomi as well as the folks behind Pokemon Go, Niantic.
