R
Roel Huis in ‘t Veld
Director, Kriya Materials · Week of October 19, 2026
1.3 million AR glasses is both a doubling and a rounding error
How a specialty coatings company should think about a market that is growing 111% a year and is still too small to matter.
TrendForce published its AR glasses forecast on 15 October. Shipments are expected to go from roughly 620,000 units in 2025 to 1.31 million in 2026, a rise of 111%, with another near-80% in 2027. Meta is expected to take about 20% of this year’s volume, and four Chinese brands (XREAL, Even Realities, Rokid and RayNeo) a combined 53%. On the display side, OLED-on-silicon and LED-on-silicon microdisplays together account for about 80% of the engines; LEDoS is favoured for lightweight information glasses, OLEDoS for entertainment, and LCoS is growing on the back of Meta’s products. (TrendForce)
Two things are true at once about those numbers. A doubling in a year is the signature of an exponential, and exponentials are what I wrote G-Force Disruption about. And 1.3 million units is nothing. Smartphones ship in the billions.
If every AR glasses frame in the world came with a coating from South Limburg, it would still not justify a dedicated line. Both facts are correct, and the mistake most companies make is to pick one and ignore the other.
The pattern is familiar. Every exponential curve looks like a flat line until it does not. The people who waited for the curve to bend before acting found that the suppliers were already chosen, the specifications already written, the qualifications already closed.
The people who bet the plant on the curve before it bent went bankrupt waiting. The question for a materials SME is therefore not whether to be in AR, but how to be present in it without being dependent on it.
The forecast itself hints at the answer. The expensive, contested part of an AR headset is the display engine. That is a fight between very large companies with semiconductor-scale capital, and it is not a fight a coatings company should join. But look at what sits in front of the engine: a polymer or glass lens, a waveguide, and an assortment of layers that have to be anti-reflective, scratch-resistant, fingerprint-resistant and very light. Those layers are mundane. They are also where the user’s experience is decided, and where a small company with deep process knowledge can earn a qualification today, at today’s volumes, and be in place when the volumes are no longer small. That is why AR/VR is still in our "other category" budget wise, despite that every name that matters is "in there".
What I find useful in TrendForce’s split by display technology is that it tells you the industry has not converged. OLEDoS versus LEDoS versus LCoS is still an open question, and the glasses themselves are splitting into entertainment and information categories with different weight and optical requirements. An unconverged market is uncomfortable for a supplier, because every customer wants something different. It is also the only moment when a small supplier can shape the specification rather than receive it. Once the market converges, the spec is written by whoever won, and the coating is commoditised within a few years.
So here is how I would size the bet, and how I try to size ours. Spend engineering hours, not capital. Qualify on the layers that transfer to other products if AR disappoints: an anti-reflective hard coat on a thin polymer lens is useful in a dozen other markets. Keep the exposure at a level where being wrong costs a year of learning rather than a company. And watch the curve, not the volume: the number that matters this year is not 1.3 million, it is 111%.
That is the G-Force logic in one sentence. Position early, size the exposure to what you can survive being wrong about, and let the exponential do the work. Time will do the rest.
Source: Meta and Chinese Brands Drive Doubling of AR Glasses Shipments, 2026 Forecast Expected to Reach 1.31 Million Units, TrendForce, 15 October 2026