In the patent record for the grant week beginning June 2, 2026, Snap Inc. was the single largest source of newly issued consumer-electronics patents, ahead of NVIDIA and Adobe by volume in a sector keyword sweep. Across the week the company was granted roughly a dozen patents, and read together they map a clear area of coverage: the rendering pipeline, display behavior, and physical-fit sensing that sit behind a camera-and-glasses product line. For a business desk, the interest is not any single claim but what the cluster, taken as issued coverage, describes about where Snap has secured a footprint.

A granted patent differs from a published application in one practical respect: it is enforceable. The claims that issued this week are the ones Snap can now assert. The batch is weighted toward augmented reality. US12646268B2 covers real-time replacement of an upper-body garment in video, swapping clothing worn by one person for clothing worn by another while matching pose. US12646266B2 covers overlaying generative visual content onto a real-world object from a user prompt and an image template. Both carry CPC class G06T 19/006 — the classification for augmented- and mixed-reality rendering — and both describe the kind of camera-effect feature that is the core of Snap's consumer surface.

From research into issued coverage

The cluster extends past effects into the apparatus and the rendering economics. US12646135B2, "Cached cloud rendering," describes offloading frame rendering to a cloud server and caching the results to reduce the graphics workload on a mobile device and, per the abstract, extend its battery life — a recurring constraint for head-worn hardware. US12646487B2 covers normalized brightness control across layered visual content on a display, and US12646240B2 covers an avatar dance-animation system driven by a neural network that reads tempo and "dance energy" from a live audio signal. The throughline is a device that renders overlays efficiently, manages its own display, and animates content in real time.

The wearable-hardware coverage is the part most directly tied to a glasses product. US12645280B2, "Detecting wear status of wearable device," describes determining whether a device is being worn by transmitting a radio signal between two communication points on the device and measuring its received signal strength against a threshold. Its claim language is direct:

The system transmits a radio signal from a first communication device of a wearable device to a second communication device of the wearable device and measures a signal strength associated with the radio signal received by the second communication device.— Detecting wear status of wearable device, US12645280B2

Wear-detection is a power-management primitive: a device that knows when it is on a user's face can decide when to wake displays and sensors. Filed alongside it, US12645936B2 covers fixed-point quantization techniques to compress a neural network into a smaller representation — the kind of model-compression work that matters when inference has to run on a constrained wearable rather than in the cloud. The two together describe both sides of the on-device problem: knowing when the hardware is in use, and shrinking the models it has to run.

What the footprint describes

Mapping the week's grants by subject, the coverage concentrates in three adjacent areas — AR rendering (G06T 19/006 and related classes), display and power management, and wearable sensing. That distribution lines up with a smart-glasses and AR-camera product rather than a general software portfolio. The records also show Snap securing coverage across the full stack of such a device: the effects a user sees, the rendering and brightness systems that draw them, the model-compression that lets them run locally, and the radio-based sensing that tells the device it is being worn.

For the competitive picture, the relevant fact is volume and concentration in a single grant week. Snap's dozen-patent batch outpaced the other consumer-electronics assignees issued the same week, and it clustered tightly rather than spreading across unrelated technologies. The other notable AR-and-audio names in the same week's record — Meta Platforms Technologies, issued patents including a speaker force-cancellation design, and Sonos in audio — appeared at lower volume in this sweep. Issued coverage of this kind defines the area in which a company can assert claims and the freedom-to-operate questions rivals building comparable head-worn hardware will encounter. What the June batch shows is Snap turning a body of AR and wearable research into that enforceable footprint, one grant week at a time.