Meta Orion AR Glasses Go on Sale: The Future You Can Wear

Key takeaways

• Orion's significance is the waveguide display and the compute puck architecture, not the demos.
• Silicon carbide waveguides solve the field-of-view problem that glass could not, at a manufacturing cost that is currently prohibitive.
• Offloading compute to a wireless puck is what makes the glasses wearable — and it is the compromise everyone will copy.
• The neural wristband input method matters more long-term than the optics.

Meta's Orion prototype is not a product, and Meta has been unusually clear about that. It is a demonstration that the hard parts of consumer AR glasses are solvable, shown in hardware rather than argued in a keynote. Judged as a shipping device it is irrelevant; judged as an engineering statement it is the most substantial thing in the category.

The display problem, and why it was so hard

Every previous attempt at AR glasses failed the same test: to overlay convincing graphics on the world, you need a wide field of view, and to be wearable, you need the optics to fit in something resembling a spectacle lens. These requirements have been in direct conflict.

Conventional waveguides — the transparent optical elements that route projected light into your eye — are made of glass. Field of view in a waveguide scales with the refractive index of the substrate, and glass tops out at a value that produces the narrow letterbox view that made earlier AR headsets feel like looking through a mail slot.

Silicon carbide

Orion's waveguides use silicon carbide, which has a substantially higher refractive index than glass. That single material change is what widens the field of view into a range where content feels situated in the room rather than pasted onto a small window.

The catch is cost. Silicon carbide of optical quality is expensive to produce and difficult to machine to the tolerances a waveguide requires. This is the primary reason Orion is a prototype: the bill of materials is reportedly in the thousands of dollars per unit, dominated by the optics.

The compute puck

The second architectural decision is at least as consequential. Orion offloads most computation to a separate wireless puck carried in a pocket.

This is not a stopgap. Thermal dissipation and battery mass are the binding constraints on head-worn devices, and both scale with on-board compute. Every watt processed on your face is a watt of heat next to your temple and a gram of battery on your nose. Moving the workload to a device that can be larger, heavier, and thermally unconstrained is what allows the glasses themselves to approach a wearable weight.

Expect this split architecture to become standard. It has an obvious precedent in how smartwatches lean on phones, and the same logic applies more forcefully here.

Input: the part people underrate

Orion pairs with a wrist-worn band that reads electromyography signals — the electrical activity of muscle activation in the forearm — to detect finger gestures before they produce visible movement.

This solves a problem that hand-tracking cameras never did. Camera-based gesture recognition requires holding your hands in the device's field of view, which is tiring, socially conspicuous, and fails in poor lighting. EMG input works with your hand at your side, requires only micro-movements, and is invisible to people around you.

For a device intended to be worn in public, the social acceptability of the input method may matter more than the display quality. Nobody will wear glasses that require semaphore.

What remains unsolved

• Cost. Reducing the optics bill of materials by the order of magnitude required for a consumer price is a manufacturing problem with no announced solution.
• Battery life. Even with offloaded compute, the display and sensors consume meaningful power in a very small volume.
• Prescription support. Integrating corrective optics with a waveguide is non-trivial, and it gates a large fraction of the addressable market.
• Software. Nobody has yet demonstrated a compelling reason to wear AR glasses all day. The hardware problem is closer to solved than the “what is this for” problem.

Why it still matters

Prototypes like this are how a category moves. Orion establishes that wide-field-of-view AR in a glasses form factor is an engineering and cost problem rather than a physics problem, and it identifies which specific problems the industry now needs to solve.

The realistic timeline for a consumer device on this architecture is several years, and it depends far more on silicon carbide manufacturing economics than on anything Meta's software teams do. That is an unglamorous bottleneck, but it is the real one.