AR optical combiners (waveguides): getting the image to the eye

last updated Mon Jun 15 2026 00:00:00 GMT+0000 (Coordinated Universal Time)
Laser-Based Display Light EnginesMetasurfaces & Flat OpticsAR glasses: the component stack and competing technologies (primer)DisplaysAR optical …

The combiner is component 2 of the AR glasses: the component stack and competing technologies (primer): it takes the image from the light engine and overlays it on the real world. This is where most of the cost, the “eye glow,” the flat-vs-curved problem, and the photolithography bottleneck live. It is also the layer the Ar Display Optics thesis bets against in its standalone form (we back fusion instead). Note: this is the free-space near-eye combiner, a different thing from the on-chip PIC waveguides covered by Photonics Material Class War, despite the shared word.

The competing approaches

The two structural problems

  1. Photolithography. Diffractive waveguides are etched in silicon/SiC foundries. Jason Hartlove (Meta) put a number on the wall: ~6 TSMCs of photolith capacity to make the 100M diffractive pairs the market wants. The combiner is the photolith-bound part of the stack, so any photolith-free path (fusion, certain metasurface/replication routes) is a manufacturing-cost wedge.
  2. Flat vs curved. Almost every waveguide approach is flat (foundry-made), but wrap-around curved lenses are what early adopters (cyclists, military, fashion) actually want, and what prescription grinding needs. Flatness is a form-factor dead end for the consumer market.

Value capture (why the standalone combiner is a hard bet)

See also

Ar Display Optics (the thesis) · AR glasses: the component stack and competing technologies (primer) (the full device) · Laser-Based Display Light Engines (the engine that feeds the combiner) · Metalenses (the metasurface platform) · Photonics Material Class War (the other kind of waveguide: on-chip PIC, not AR)

Related concepts