Qualcomm CEO Cristiano Amon disclosed that his company is now working on more than 40 different AI wearable designs — from camera-equipped earbuds to jewelry, pins and watches — and positioned Qualcomm as the default silicon layer beneath whatever device category eventually displaces the smartphone.
To support that ambition, Qualcomm announced two products: Snapdragon Reality Elite, a chip platform for mixed-reality glasses, and START (Scalable Turnkey AI-Ready Toolkit), a hardware-plus-software bundle plus white-label program designed to compress the path from concept to shipping product for hardware manufacturers.
The two announcements
Snapdragon Reality Elite delivers up to 60% GPU performance gains, up to 30% CPU gains, and up to 160% NPU gains over Qualcomm’s previous XR platform. The company says it can run a 3-billion-parameter language model at 45 tokens per second and supports 4.4K per-eye resolution at 90 fps — a marginal step up from the XR2+ Gen 2’s 4.3K. XREAL’s Project Aura and an upcoming device from Play for Dream are among the first products built on the platform.
START packages an AR chip, software stack, companion apps and three reference designs (an audio-plus-camera setup similar to Meta’s Ray-Ban smart glasses, a monocular display, and a binocular display). Eyewear manufacturers Inspecs and O’Neill are named as initial white-label partners.
Why glasses, why now
Amon reportedly said that smart-glasses shipments are currently in the tens of millions per year and reportedly predicted shipments could reach hundreds of millions within a couple of years and rival smartphones in scale. For context, smartphone shipments in 2025 totaled approximately 1.26 billion units, roughly 3% above the prior year.
Amon’s framing is that AI agents — not apps — will become the center of digital life, and that whichever devices act as always-on sensory endpoints for those agents will accrue the strategic value. According to Amon, the key principle is that these are wearable devices that stay with users constantly and can observe their surroundings.
The data layer underneath the hardware story
The more revealing portion of Amon’s interview concerns why non-traditional hardware entrants are pushing into consumer devices. Amon said the motivation is data: wearables will generate vastly larger signal volumes than the corpora used to train current frontier models, and AI companies need ownership of those endpoints to train future models and build bespoke user experiences.
That reframes the product category. Mixed-reality glasses and camera-equipped pins are not merely a new screen — they are a new sensor network pointed continuously at users and their physical environments. The economic logic that makes them attractive to OpenAI, Meta and Samsung is the same logic that makes them strategically uncomfortable: whoever controls the endpoint controls the training data pipeline.
Qualcomm’s structural position
Qualcomm’s strategy is to avoid picking a winner. By offering START as a turnkey program with reference designs and white-label partners, the company lowers the capital and engineering barrier for new entrants — which means more shots on goal for the post-smartphone category, and Qualcomm silicon inside the majority of them regardless of which form factor wins.
It is the same playbook that made Qualcomm indispensable to the Android ecosystem: rather than compete with Apple’s vertical integration, sell the foundational layer to everyone trying to. Amon said that Qualcomm’s entire roadmap is being upgraded because current devices aren’t prepared for an agent-centric computing model.
The structural bet is straightforward. If the next platform shift produces dozens of viable form factors instead of one dominant device, the silicon supplier sitting beneath all of them captures more value than any single hardware brand competing on top.
