Snap Inc. has spent years describing its augmented-reality glasses as a long-horizon investment funded out of the cash its advertising business generates, and the patents that issued to the company in the week of April 28 to May 4, 2026 show where that investment lands at the level of an actual product. Across that window the U.S. Patent and Trademark Office granted Snap 13 patents, and the claims fall into two halves of one device: the eyewear hardware on one side and the perception software that runs on it on the other.

The hardware grants are concrete. US12611017B2, "Carry case for rechargeable eyewear devices," claims a case with charging contacts mounted on flexible walls that move relative to the storage chamber, coupling to contact formations on the exterior of the glasses — in some versions provided by the hinge assemblies that join the temples to the frame. It is a granted claim on the off-head charging system that all-day smart glasses depend on, the same role a charging case plays for wireless earbuds. US12613458B2, "LED illuminated projector," reaches the display engine, claiming a projector that builds a full-color frame from an array of color LEDs and collector structures, assembling the frame from subframes formed by spatial movement of the element array — the kind of compact light engine a see-through display needs to put an image in front of the eye.

The perception stack

The larger half of the week's grants is software for understanding the physical world. US12614293B2, "Monocular world meshing," claims generating a depth estimate and a mesh model of an environment from image data and the client device's pose, using a single camera rather than a depth sensor. US12614348B2, "Incremental scanning for custom landmarkers," covers updating a localizer map of a scene by capturing new data points, determining the capture device's poses, and selecting target frames to build an updated map — the persistent, improvable spatial map that lets AR content stay anchored to a place. US12614359B2, "Stationary extended reality device," claims using a machine-learning model to predict tracking information for a real-world object in a fixed field of view and overlaying XR elements based on that tracking.

A carry case for an electronics-enabled eyewear device, such as smart glasses, has charging contacts that are movable relative to a storage chamber in which the eyewear device is receivable.— Carry case for rechargeable eyewear devices, US12611017B2

From the camera to the experience

A further group reaches the imagery itself and the experiences built on it. US12614354B2, "Animatable garment extraction through volumetric reconstruction," claims generating a three-dimensional mesh of a person from a single image, matching its pose, and extracting the garment as an animatable XR item — the technical basis for an AR try-on. US12614260B2 covers correcting the perspective distortion of a close-range selfie by inverting a 3D face model to render an image that appears taken from farther away, and US12614396B2, "AR assisted safe cycling," claims a head-worn device that estimates the distance to a relevant object, computes a braking distance from its GPS speed, and warns the wearer — an application that assumes the glasses are worn in motion outdoors. One grant even reaches Snap's manufacturing toolchain: US12614693B2 covers automatic calibration of a scanning electron microscope on periodic nanostructures, the kind of metrology used in producing the diffractive optics of a waveguide display.

The classification data reflects the split. The week's grants concentrate in G06T 19/006 and G06T 19/20 (augmented-reality rendering and editing), which appear across the world-meshing, garment-extraction, and stationary-XR claims, alongside G06T 17/20 (3D mesh modeling) and G06T 7/70 (pose estimation) on the perception side, and G06F 1/163 (wearable computing) tying the cycling and case claims to a head-worn form factor. The grants consistently describe one object — a pair of camera-equipped glasses — seen from the hardware, the optics, the perception code, and the manufacturing line.

The count is worth situating. Thirteen grants in a single week is a modest figure next to the high-volume assignees in the broader wearable-and-headset slice of the same window, where the week's leaders were Qualcomm at 31 grants and Apple at 29. Snap's significance in the record is not its volume but the span of what those thirteen grants touch: a company far smaller than those two issued patents in one week that reach from a charging case and a light engine through SLAM-style world meshing and persistent scene maps to selfie correction and a manufacturing-metrology method. That a single week's output for one company covers the case, the display, the perception stack, and the production line is the part of the record that rewards reading — it describes a vertically scoped device effort rather than a single subsystem. Quantifying the breakdown — two hardware grants, four perception grants, two imagery grants, a wearable-application grant, and a metrology grant — is more informative than the headline total, because it shows the coverage is distributed across the whole device rather than clustered in one corner of it.

It is worth being exact about what each grant does and does not establish. The carry-case claim in US12611017B2 is a mechanical and electrical-contact claim on a charging case, not a claim on the glasses' internals; the monocular-meshing grant US12614293B2 covers a specific single-camera pipeline, not depth estimation as a field. Each is a granted right on a particular implementation, and a single week of issuances says nothing about how Snap's portfolio in this area compares to others building AR eyewear. What the week's record shows is the shape of the coverage rather than its scale: the case that charges the glasses, the projector that lights the display, the code that maps the room and tracks objects, the imagery features layered on top, and the metrology behind the optics — the same device, captured from end to end in one week of grants. For a reader tracking how Snap's reality-hardware spend is converting into issued patents, that breadth is the signal.