Show summary Hide summary
Qualcomm unveiled a new AI audio chip at its Snapdragon Summit in Maui that moves significant intelligence into headphones and other wearables. The platform is designed to run on-device apps, capture environmental audio and video, and let virtual assistants act without a constant phone connection — a shift that promises convenience but raises privacy and battery questions.
What the chip does and how it works
Called Snapdragon Sound Elite Gen 2, the platform embeds an on-device AI accelerator — an eNPU — that can handle speech, transcription, translation and other assistant tasks locally. Manufacturers can store apps in the chip’s memory so the device can operate independently and communicate directly with servers using standard https connections, since each unit can have its own IP address.
Ultra-thin women smoking cigarettes are making a comeback as a fashion trend
Windows Memory Diagnostic explained: what the test does and how to read results
The chip supports both Bluetooth and a low-power Wi-Fi mode Qualcomm describes as micropower Wi‑Fi. That radio consumes far less energy than standard Wi‑Fi but sacrifices peak speed and increases latency. For audio, though, Wi‑Fi can deliver higher-fidelity streaming than Bluetooth because it avoids the same level of realtime compression.
Features users are likely to see
Qualcomm says devices powered by the chip will enable new, always-on experiences. Built-in apps and the eNPU allow two-way conversational assistants that can perform actions — such as creating calendar events or making purchases — without routing every request through a paired smartphone.
On the audio side, the platform can run adaptive and selective active noise cancellation so users hear specific sounds on demand: think amplified voices in a conversation or emergency sirens while filtering other noise. The system relies on conventional machine-learning algorithms rather than large language models for these real‑time audio tasks.
Qualcomm also highlighted support for lossless audio streaming and the ability to interface with sensors and tiny cameras embedded in “hearables.” Company representatives said those cameras would be low resolution and that on-device AI would mask or obfuscate people captured in footage to reduce identification risks.
Device types, battery concerns and partnerships
The company expects the chipset to appear in a wide range of form factors: over-ear headphones, open-ear audio devices, audio glasses, camera-equipped earbuds, hearing aids and other wearable clips or pendants. A short list of target product types includes:
- Over-ear and in-ear headphones
- Open-ear and audio glasses
- Camera buds and hearing devices
- Future personal AI wearables such as pendants or clips
Qualcomm says the design balances audio quality, AI features and power consumption. Still, company staff acknowledged that continuous, always-on use will change how battery life feels to users. Andrew Laister, a Qualcomm staff product manager, told reporters the trade-offs are real and likened the moment to an app-driven turning point: “It’s like the first smartphone moment with the Play Store,” he said, adding that device makers will choose which apps ship on their products.
No specific timelines were given for consumer products, but Qualcomm said it is working with partners including HP and Bose to explore headphones and speakers using the chipset.
Where this fits in the broader landscape
AI-enabled audio is not entirely new. Some makers already sell headphones that record, transcribe and summarize conversations, and big tech vendors offer assistant access via earbuds that still rely on a tethered phone. Qualcomm positions its new chip as a generational advance over the earlier Snapdragon S7 Gen 1, doubling on‑device AI throughput to about 128 billion operations per second from 64 billion, while claiming roughly 40% lower power use and a 30% smaller footprint.
Research groups and other companies have also explored similar concepts. University of Washington teams have embedded computing in consumer earbuds to isolate speakers in noisy settings and to let wearers control selective noise cancellation. Separately, xMEMS unveiled its Cypress audio chip, which uses piezoelectric methods to produce ultrasonic sound; that platform reached mass‑production readiness in September 2025, though consumer devices have yet to be announced. Reports have also suggested Apple may be experimenting with earbuds that include cameras, a topic Bloomberg editor Mark Gurman has raised.
Live Science was invited to attend Snapdragon Summit 2026 courtesy of Qualcomm; the company covered travel and accommodation.
