Theses›Consumer devices›AR display optics

AR display optics

The smart-glasses optical stack splits into a light engine (generate RGB), a combiner (get it to the eye), and the prescription lens.

  • Bull
  • Early
  • $22B by 2035

The callThe drivers win: glasses go from under a million units to tens of millions and today's optics throw away 95-99% of the light. Early by a product cycle, and an M&A market.

D1D2R1R2C1foragainst D1D2R1R2C1foragainst
Fig. 1The balance. Drivers and catalysts against restraints, each block sized by its weight; catalysts hang dashed because they have not happened yet, sized by their odds. It tips toward the drivers.
  1. D1Display glasses go from about 950k units a year to 32M by 2030, and the optics are the part of the bill that does not commoditise with the phone chain.
  2. D2Today's waveguides throw away 95-99% of the light, so an all-day prescription lens on a tiny battery needs a different light path.
  3. R1The curved-plastic half of the view is falsified for now: the fused engine is a lab demo, red microLED is stuck below 5 µm pitch, and Meta shipped flat glass.
  4. R2Meta keeps the component field alive on NRE to hold prices down, and acquirer appetite is gone for now with the M&A clock at 5-8 years.
  5. C1Apple's N50 glasses ship; the first generation has no in-lens display, so it grows the prescription-lens layer and makes AR a category every OEM must answer.late 2027
Rests onthe beliefs underneath
Supported by · confidence
What would prove itdated tests, open and resolved
2027202820292030203120322033today1234567891011predictionswatch 2027202820292030203120322033today1234567891011predictionswatch
Fig. 2Each ring is a dated test: dashed while open, struck through when overdue, ticked or crossed once resolved.
  1. 1Within 12 months (by 10 Jul 2027) AR/AI glasses become a recognised VC category — a meaningful rise in dedicated glasses-stack rounds led by tier-one / large multi-stage funds vs the prior 12 months, and the category named in mainstream VC commentary the way silicon is in 2026.Jul 2027 · open
  2. 2By end-2028 at least one shipping consumer smart-glasses product uses a photolith-free, non-diffractive light path, not solely diffractive or bonded-glass waveguides.Dec 2028 · open
  3. 3On-demand / distributed prescription-lens manufacturing is adopted in production by at least one Tier-1/2 smart-glasses OEM by end-2028.Dec 2028 · open
  4. 4A shipping consumer glasses product with a photolith-free light path, and a Tier-1 or 2 OEM taking on-demand Rx lenses.Dec 2028 · open
  5. 5Monolithic RGB microLED at sub-5um pixel pitch ships in a volume consumer smart-glasses product (>100k units/yr) by end-2029 - i.e.Dec 2029 · open
  6. 6The sub-5 µm red microLED wall falls; without it the fused engine stays a demo.Dec 2029 · open
  7. 7The next AR-display-optics exit above $200M is a pre-revenue / pre-scale strategic acquisition by an OEM or display-supply major, not an IPO or revenue-scaled standalone, through 2030.Dec 2030 · open
  8. 8No standalone AR-display-optics component company reaches a >$1B venture-scale outcome WITHOUT acquisition by 2030.Dec 2030 · open
  9. 9By end-2030 the volume-leading consumer smart-glasses display engine is EMISSIVE or additive, NOT a modulated architecture (LCoS/DLP) that burns full illumination power regardless of content.Dec 2030 · open
  10. 10More than $200M of AR-optics M&A; below that the exit market is not liquid and the case fails.Dec 2030 · open
  11. 11By end-2032 at least one consumer smart-glasses product shipping >100k units/yr uses a COMBINER-LESS light path, deleting the separate waveguide layer.Dec 2032 · open
Expressionslisted, private, technologies
Private · 7 private companies, 2 of them in China.

BenarioXR · TriLite Technologies · Optinvent · Jade Bird Display (JBD) · Dispelix · SeeYA Technology · Raxium

40 more private companies tracked, not shown

Technologies
Who has already bet on itprivate rounds, from public sources

$721M raised by 5 companies, $543M of it in the last three years. Wave under way: Large rounds in the last three years; the category is funded and getting crowded.

Backed by a top-tier venture firm: Jade Bird Display (JBD), Raxium. Backed by a strategic investor: SeeYA Technology, Dispelix, TriLite Technologies.

  • $441MSeeYA Technology
    • Founded 2016
     IPO (STAR Market), Mar 2026 $317M (Guotai Junan Securities (underwriter)) · Series B, Dec 2021 $124M (China Internet Investment Fund, Shanghai Pudong Innovation & Technology, Xiaomi Industrial Fund)
  • $140MJade Bird Display (JBD)
    • Founded 2015
     Series B2, Oct 2025 $140M (Chaos Investment Group, Citic Goldstone, Lightspeed China Partners) · Series B1, Aug 2025 (Chaos Investment Group, Shanghai Sci-Tech Innovation Center Capital)
  • $85MRaxium
    • Founded 2017
     Series B, 2021 (Vanedge Capital, Khosla Ventures)
  • $33MDispelix
    • Founded 2015
     Series B, Nov 2021 $33M (Atlantic Bridge, CCB Trust, Flashpoint)
  • $22MTriLite Technologies
    • Founded 2019
     Series A extension, Sep 2024 $13M (Continental (CVC unit), APEX Ventures, B&C Innovation Investments) · Series A, 2021 $9M
What changedthe record, newest first
  • 31 Aug 2026A component supplier is not valuable because it is hard to acquire - that often just means no exit. It is valuable when the buyer cannot set its price. Which needs one of three things: owning the only line that makes the thing, selling into more than one end market, or owning IP everyone has to license. On that test the display engine is the weakest layer in smart glasses, because a near-eye microdisplay has about five possible customers in the world.
  • 9 Aug 2026The display architecture shipping today is probably not the one that lasts. Modulated engines burn full power on a mostly-black screen, and the separate combiner throws away 95-99% of the light - so the always-on endgame is emissive, and eventually has no combiner at all. Backing the current volume winner means backing the thing that gets replaced.
  • 20 Jul 2026A senior ex-Meta optics leader's blunt read: the big platforms would now rather have their supply chain build AR display tech and feed it to them than buy a hardware startup — they got burned before. So the optics-startup exits are real but slower and later (think 5–8 years, not 1–2), and the near-term "quick strategic sale" hope is mostly a mirage.
3 earlier
  • 10 Jul 2026Capital rotates memetically — crypto, then AI, then silicon. Next is AR glasses: the industry already knows, VCs do not care yet. On the record: give it a year.
  • 6 Jul 2026The most valuable unsolved problem in AR glasses may not be brightness or cost but FOCUS: today's waveguide displays lock everything to one focal plane, which is why long sessions tire the eyes, and nobody has cracked continuous focus control. Watch which architecture solves it. Separately, the materials that turn a lens into a display are quietly becoming a coating business for the big optical-glass names (Merck, Hoya, Corning), priced per-area or per-licence because the layer is almost weightless.
  • 15 Jun 2026The defensible AR-optics bet isn't a better laser or a better waveguide, those are components that get bought cheap. It's the architecture that fuses them so there's no separate waveguide at all (the display is generated in the lens). That's the only position hard enough to replicate that an OEM has to pay up for it.