Will 2026 be the make-or-break year for AR ? Meta answers with a 4-product lineup

Meta unveils its new smart glasses featuring artificial intelligence, marking a turning point in its technology strategy. Raphael Mermet-Lyaudoz, a Technology & Market Analyst, Display at Yole Group, attended at the announcement and analyzed the impact of this innovation. This launch is part of a series of new strategies implemented by Meta to strengthen its position in the field of augmented reality and AI.

This article is related to the following reports: AR/VR 2025 – From Optics and Displays to Advanced Sensing and Processing – Vuzix Z100 smart glasses and JBD microLED Display – Rayneo X3 Pro AR glasses – MicroLED – Markets, Applications and Competitive Landscape 2025 – MicroLED – Technology, Equipment and Manufacturing 2025…

On September 17th, Meta’s CEO Mark Zuckerberg introduced a brand-new lineup of products at the Meta Connect keynote, showcasing new Meta AI capabilities. With no new VR headsets — but several new glasses — and EssilorLuxottica highlighted early on as a key partner, Meta’s 2025 focus is clearly shifting toward lightweight AR.

The products consist of 3 pairs of glasses embedding an open ear speaker and camera, and a 4^th one with a display in it – the first time for a commercial product by Meta. Each of these devices serves as a gateway to Meta’s expanding suite of AI-powered features.

The lineup includes a new pair of Ray-Bans with updated frames, the Oakley HSTN (introduced a bit earlier this year), and the Oakley Vanguard — a pair of glasses clearly aimed at outdoor and performance use. The Vanguard is positioned as a GoPro replacement, with a central, stabilized imaging camera. That said, it weighs 67g — nearly twice the weight of a regular Oakley outdoor frame without electronics. Will trail runners, cyclists, and other athletes accept the tradeoff ? The next few months will tell.

Meta Ray-Ban glasses are on an exponential sales trajectory — with more than 2 million units sold in 2025 and triple-digit YoY growth. Until now, these were audio-only products, and a version with a display had been long anticipated to bring the category to a new dimension, thanks to a richer app environment. Now, the Meta Ray-Ban Display is here — shipping by September 30th. The display is monocular (right eye), 2D, offers 42 PPD resolution within a 20° field of view, and can deliver up to 5,000 nits to the eye, making it usable even in bright sunlight.

The display technology itself is composed of an engine Zuckerberg said was “custom” in the keynote – most likely LCoS-based (as detail in Yole Group analysis) and the integration at the temples closely resembles Avegant’s architecture (also well detailed coming teardown). The display engine is paired with a geometric waveguide, recognizable by its visible stacked glass slabs used for light extraction. It is even possible to recognize a Lumus-type of geometric waveguide, which is manufactured by Schott via a non-semiconductor process. The lenses are photochromic so they can adapt to dimmed or sunny environments, and up to ±4 unifocal prescriptions can be supported — although details on integration and user flow are still lacking.

It is noteworthy that just a year ago, Meta showcased a prototype based on microLEDs and diffractive waveguides (Orion). Yet, their 2025 commercial product opts for a more mature LCoS and geometric waveguide architecture — a choice that reflects both performance priorities and technological readiness. Indeed, geometric waveguides are currently the most power-efficient option, enabling high brightness without excessive power consumption from the display engine.

Battery life is said to last all day — although the live AI features are limited to about one hour, which aligns with our modeling.  This limitation suggests that activating AI services likely triggers all subsystems: display, camera, audio, sensors, and on-device processing.

The frame is Ray-Ban-branded and echoes their iconic design language. Yet, unlike the audio-only models, no additional frame options are available. It could underline a more complex integration and/or a more modest expectation in terms of sales for these $799 70g glasses.

As for processing, Meta says on their website, a smartphone with the Meta AI app installed will be necessary to access all the features, meaning some of the computing will be handled elsewhere than the glasses themselves – as was the case with the Orion suite’s compute puck.

The glasses are controlled through the EMG-based “Neural band”, which could be a product on its own. Indeed, the handwriting demo during the keynote — where text input was generated via wrist motion alone — was arguably the most impressive moment, alongside the choreographed entrance. For the record, Meta – still Facebook at that time – acquired CTRL-Labs (the company developing this technology) in 2019.

Meta is clearly leveraging its momentum in audio wearables to normalize and prime the market for visual AR. The transition from passive audio engagement to heads-up visual interaction signals an expanding continuum of modality in ambient computing. Other OEMs will follow and kind of rush into the breach opened by Meta’s product and marketing push: Google and Snap are already positioned, Amazon is likely to enter, and even Apple may be forced to react. In China, both Alibaba and Xiaomi have released smart glasses this year — the former with a display, the latter audio-only.

Looking ahead, Yole Group forecast an 18-million-unit market for AR and smart eyewear by 2030. While that figure remains modest compared to smartphone volumes, it marks a steady escape from the “Death Valley” AR was stuck in for over a decade. 2026 could be the inflection point expected in the last edition of the report — when technical maturity, ecosystem support, and consumer readiness finally converge.

Stay tuned!

Raphaël Mermet-Lyaudoz, PhD is Technology & Market Analyst, Display at Yole Group.

He is a member of Display team. Raphaël covers technological evolutions and market trends in the display field to deliver his analysis in Yole Group’s products.

Raphaël obtained an Engineering degree from Institut d’Optique (FR) and a PhD from Université de Lyon (FR) in collaboration with INL-CNRS. During his academic curriculum he particularly studied Optics, Material Appearances, Nanophotonics, Emerging Materials and lasers and published some of his works in peer-reviewed scientific journals.

Source: www.yolegroup.com