MAUI, Hawaii — Qualcomm has introduced the Snapdragon Sound Elite Gen 2, a novel AI audio processor designed to transform wearable devices into autonomous intelligent systems. Announced at the company’s annual Snapdragon Summit held here from September 21 to 24, the chip aims to free users from the necessity of maintaining a continuous connection to a smartphone.
According to engineers, the new silicon allows devices such as earbuds, smart glasses, and speakers to host applications directly on their onboard memory. This capability enables virtual assistants to execute tasks—ranging from scheduling calendar entries to processing online purchases—based on voice commands, even when Wi-Fi is unavailable. When connected to the internet, these devices can also sync with cloud-based services like ChatGPT.
Andrew Laister, a staff product manager at Qualcomm, described the development as a pivotal moment similar to the launch of the Play Store for smartphones. “It’s like the first smartphone moment with the Play Store,” Laister told journalists during a media Q&A session. “In this sense, that app store for your earbuds is currently owned by the device maker. They will be the guys that will choose which apps they will pre-bundle, and then we’ll go from there.”
The architecture utilizes a specialized micropower Wi-Fi component alongside Bluetooth to establish connections with computers or smartphones for visual interfaces. Qualcomm representatives noted that while this approach incurs higher latency and lower peak speeds than standard Wi-Fi, it consumes significantly less power. For audio applications, however, Wi-Fi offers superior playback quality because it avoids the data compression inherent in Bluetooth transmissions.
The chip also introduces adaptive and selective active noise cancellation powered by standard machine learning algorithms. This feature allows users to filter audio environments on demand, such as amplifying human voices during conversations or preserving critical ambient sounds like emergency sirens while navigating urban areas. Additionally, a built-in subsystem supports lossless audio streaming, a high-fidelity format typically reserved for premium service subscribers.
A core component of the Elite Gen 2 is its “always-on intelligence” engine, driven by an embedded neural processing unit (eNPU). This accelerator powers on-device AI assistants, real-time transcription, translation, and the construction of “contextual awareness.” By continuously monitoring the user’s environment, the system harvests data to build a personal profile, enhancing the relevance of future responses.
Although primarily audio-focused, the chip is designed to interface with sensors and cameras. This includes integration with “camera buds”—headphones equipped with tiny lenses that passively record footage during daily activities. Addressing privacy concerns, Qualcomm stated that future earbud cameras would operate at low resolution and utilize AI masks to automatically obfuscate the identities of individuals captured in the footage.
Laister acknowledged that while device manufacturers no longer need to compromise between audio quality, AI features, and battery life, there will be an impact on power consumption due to the continuous nature of use. The company is currently evaluating the feasibility of running two applications simultaneously in RAM, such as an AI agent and an audio streaming platform, while storing other data in flash memory.
The Elite Gen 2 serves as a spiritual successor to the Snapdragon S7 Gen 1, which appeared in devices like the Xiaomi Buds 5 Pro series three years ago. The new chip delivers double the AI processing power, handling 128 billion operations per second compared to the previous 64 billion, while being 30% smaller and consuming 40% less power.
Competitive landscape analysis shows that other companies are pursuing similar technologies. Apple’s AirPods currently require an iPhone connection for complex AI tasks, though Bloomberg’s Mark Gurman reported that Apple may be developing AirPods with built-in cameras for data harvesting. Meanwhile, Viaim recently launched AI-enabled headphones focused on conversation recording and translation, and research teams at the University of Washington have previously demonstrated deep learning systems for selective noise cancellation.
Qualcomm is collaborating with partners including HP and Bose to integrate the chipset into future headphones and speakers. No official release date has been announced, and information regarding the specific impact on battery life compared to conventional non-AI devices remains pending.
The app store comparison is spot on. This could really fragment the market if every brand locks you into their ecosystem.
Lossless streaming without Bluetooth compression is huge for音质 enthusiasts. Hope the micro Wi-Fi latency doesn’t ruin the experience.
Camera buds are interesting, but I’m still nervous about those AI masks actually protecting identities. Better hope the obfuscation is robust.
Battery life will be the real test here. Running always-on AI and lossless audio simultaneously sounds like a power hungry nightmare.
Finally, earbuds that don’t need my phone for everything? The standalone AI assistant sounds like a game changer for commuters.