In a keynote address in the SID Display Week Business Conference, Meta VP Jason Hartlove laid out the goal the AR Glasses industry should strive for if the want to see wide consumer adoption. The goal is to achieve a $100 premium to add a display module to an average pair of eyewear (with AI functionality too). If this price point can be reached, a volume of 100 million displays will be needed. He did not specify what displays should be developed – just the price and volume goals. To address this challenge directly, a panel discussion was organized on the trade show floor which brought together a purposefully cross-disciplinary group:

• Gary Jones, Moderator and General Chair of the Display Week 2026 program
• Guillaume Chansin, Associate Director, Counterpoint Research (market data)
• Michiel Callens, Display Architecture for AR, Meta
• Hugo da Silva, CEO, Morphotonics (high-volume waveguide manufacturing)
• Satoshi Shiraga, CEO, Cellid (waveguide fabrication and system reference design)
• Jinxin Fu, Sr. Dir., Applied Materials (semiconductor process-based waveguide and display)

Market Context: Where Are We Now?

Guillaume Chansin of Counterpoint Research opened with a data-grounded reality check. The overall AR/VR display market has been in decline for several years, driven primarily by the VR segment — a consequence of excess inventory and headsets still running on single displays rather than the dual-display configurations now entering the roadmap. Counterpoint expects modest growth in 2026, with a more substantial inflection in 2027 as new dual-display products launch, many of them adopting micro-OLED (OLED-on-silicon) for the VR market.

The AR market is a different story. Shipments doubled in 2025, and that momentum is expected to continue — with two caveats: memory price hikes (DRAM costs have risen 8–10x in some segments) and geopolitical supply chain disruption. Looking at market share in the second half of 2025, birdbath and prism optics still dominate, with RayNeo, X-Real, and Viture leading that category. But waveguide-based glasses are gaining fast: from 19% share in H1 2025 to 38% in H2 2025, led by Rokid, Meta, and Even Realities. Meta’s waveguide glasses were only on the market for a single quarter of the tracking period, which contextualizes their ranking.

In China specifically, Chansin noted, the ecosystem is building strongly around microLED as the light engine of choice — a trend the broader industry is watching closely.

The most striking data point came on the smart glasses (display-less) side. AI smart glasses — defined by Counterpoint as requiring an on-device AI-capable processor, voice interaction, camera-based visual assistance, and wireless connectivity — grew 139% year-on-year in H2-2025. In 2026, the market is expected to exceed 10 million units and approximately double again. The long-term forecast reaches approximately 50 million units by 2030.

The uncomfortable implication for the display industry: glasses with displays currently will only represent about 8% of the AI glasses market in 2030. Chansin was direct about this — it is not good news — but framed it as an opportunity. That ratio is not fixed; it is a function of cost, performance, and form factor, all of which are within the industry’s control.

The 100 for 100 Initiative: Meta’s Industry Call to Action

Michiel Callens of Meta presented the conceptual and commercial logic behind the 100-4-100 initiative — and it is more rigorously grounded in economics than the catchy name suggests.

The prescription eyewear market offers the right reference frame. Somewhere between two and four billion people globally require some form of vision correction; roughly two billion of those purchase a new frame each year and are willing to spend approximately $250 on it. Mapping this population onto a price-demand curve reveals that current AR glasses with displays — typically priced above $500 — sit on a very steep, very unfavorable section of that curve. A reduction in price can meaningfully increase demand.

The $100 display module target is the number that moves the needle. To be correct – this is the add-on price a consumer would need to pay to get a single display in a pair of AI glasses vs. no display. This is in addition to the roughly $190 premium consumers are paying today for the AI functionality, according the Meta survey data. Even with both these premiums, you land at roughly 10% of the two-billion-unit addressable prescription frame market. “Ten percent of two billion is 200 million units per year. Even at half that, the market opportunity is transformational,” noted Callens.

Callens was careful to note that getting to $100 is not an incremental improvement problem — it requires an industry-wide structural change. He drew the analogy to ITRS in semiconductors and the successive glass generation standardization in flat panel displays: both succeeded because the entire supply chain aligned on shared roadmaps, enabling coordinated investment that de-risked individual commitments. That is precisely what Meta is now proposing for AR.

Meta identified three specific bottlenecks it considers hardest to crack:

Waveguides — Currently expensive to produce, poorly standardized across processes and materials, and manufactured at low throughput. Alignment on architecture, materials, and process standards would simultaneously drive cost down and enable volume scale-up.

Display engine throughput — High waveguide production capacity is pointless if display engine manufacturing is the bottleneck. Units-per-hour targets need to be coordinated across the full display module, not optimized component by component.

Metrology and calibration — Tool lead times are long, units per hour (UPH) is typically in single digits, and calibration is a node-by-node problem that slows production significantly. Faster measurement techniques are needed throughout the display system.

To address these challenges, Meta wants to help organize the International Technology Roadmap for Augmented Reality (ITAR-AR). It is seeking collaborators across the full stack: waveguides, display engines, eyepieces, materials, equipment, and metrology. It is explicitly a community effort, and Callens was open about wanting input from the audience on how to structure it.

The Manufacturing Debate: Semiconductor vs. Display Infrastructure

One of the more substantive exchanges of the session concerned the right manufacturing paradigm for waveguides — and it surfaced a genuine strategic disagreement worth paying attention to.

Applied Materials’ Jinxin Fu argued for the semiconductor process route. Using semiconductor-grade process flows, Applied has demonstrated waveguides that can and meaningfully improved display performance. The company has publicly announced foundry partnerships targeting high-volume production, and Fu argued that at sufficient volume, semiconductor infrastructure’s cost-per-unit economics become attractive.

Morphotonics’ Hugo Da Silva pushed back — politely but firmly. His argument: a waveguide is geometrically much larger than a semiconductor device. The entire infrastructure of semiconductor manufacturing was built to produce devices where thousands fit on a single wafer. Applying that paradigm to waveguides, which are closer in scale to display panels than to chips, may be structurally incapable of hitting the required cost targets regardless of volume.

Da Silva’s counterproposal was to draw on display manufacturing heritage — the generational glass size scaling from Gen 3 through Gen 12 that brought LCD costs down by orders of magnitude — and apply those lessons to waveguide production. The first LCD TV was not good by today’s standards, he noted; neither was the first iPhone. Starting from a lower-cost-potential platform and iterating is, in his view, a more viable path than starting from an inherently expensive platform and hoping volume cures it.

This is not a resolved debate, and the panel did not try to resolve it. Both approaches are being actively pursued, and it is plausible that different applications and price tiers will ultimately favor different manufacturing paradigms.

Technical Requirements: Form Factor, Performance, and RX Support

Satoshi Shiraga of Cellid laid out what “sufficient performance” means in practice for the 100-million-unit vision. The display module currently accounts for roughly 50% of total system weight, making it the single largest contributor to the form factor problem. The target total system weight is under 50 grams; 40 grams or better is the aspiration. Achieving this requires aggressive lightweighting of the waveguide specifically.

On optical performance, the current commercial baseline — monochrome green, limited field of view, text-only use cases — is not enough. A full-color system with field of view of 50 degrees or greater, full HD or higher resolution, and reasonable projector power consumption is the target. Content delivery — YouTube, Instagram, interactive applications, gaming — requires full-color display capability, and AI applications alone are not sufficient to drive mass adoption.

Callens added a point none of the other panelists had raised: prescription (RX) support. Of the two billion people who buy prescription frames annually, a significant portion cannot comfortably use glasses without vision correction. Without meaningful RX accommodation, the total addressable market is structurally limited. Callens flagged this as an underappreciated area where moving the needle would meaningfully expand market reach — and Fu echoed the point from personal experience, noting that RX is the first feature he personally requires before adopting AR glasses.

Killer Applications: What Gets People to Wear Them?

The panel’s second discussion thread focused on use cases — what applications actually drive adoption at a $100–$200 price point.

The near-term baseline is already set by Meta’s Ray-Ban glasses: camera, speaker, voice assistant, and increasingly AI-powered visual context. Callens made the point that current hardware may already be sufficient for a much more compelling experience than exists today — the bottleneck is software and AI capability, not display image quality. A small field of view device with good-enough display quality can ride the AI wave without requiring a step-function hardware breakthrough.

Da Silva offered a grounded personal example: he commutes by bicycle in the Netherlands, takes meetings on his Meta Ray-Ban Display glasses, and finds them genuinely useful — but misses basic visual output. Even simple on-screen text — translations, navigation instructions, incoming message previews — would unlock meaningful additional value without requiring a premium optical system. He cited translation as a particularly high-value use case for non-English speakers, including his own parents.

Shiraga pointed toward content consumption as the medium-term expansion: from AI assistance to Instagram, YouTube, and interactive content, and ultimately to gaming, all of which require full-color display capability.

Callens offered perhaps the clearest framing of the application trajectory: AI glasses are the optimal interface for giving AI visual context and for AI to return information to the user. The hardware that exists today can support that interaction. What the market needs now is the software layer and the AI models to make it compelling — not necessarily a display revolution.

The “Good Enough” Threshold and the Ramp-Up Imperative

Da Silva closed with what may be the panel’s most useful strategic observation: the technology trap. In his experience across multiple technology cycles, teams tend to continue improving a technology indefinitely rather than declaring “good enough” and shifting effort to manufacturing scale-up. This creates development loops that delay market entry by years.

His argument to the assembled industry: if the technology has reached the good enough threshold for the first generation of meaningful use cases — and there was general panel agreement that it has, at least for AI assistance and basic visual output — then the priority should shift from performance improvement to production ramp. That means agreeing on common platforms, common interfaces, and common standards, in the same way the LCD industry created the generational glass size framework that allowed an entire ecosystem to invest in sync.

This is, at its core, the same argument Michiel Callens made with the ITAR-AR initiative: the display industry already knows how to do this. It has done it before, with LCD and with flat panel generally. The question is whether it can organize fast enough to capture the AI glasses opportunity before the market solidifies around display-less solutions.

Bottom Line

The market data is clear: AI glasses are growing explosively, and the prescription eyewear addressable market is enormous — but display-equipped glasses currently represent a small fraction of the opportunity, and the cost structure is the primary reason. The $100 display module target is not arbitrary; it is the number that, combined with AI value-add, puts display glasses on the right section of the price-demand curve.

Getting there requires three things working in parallel: waveguide manufacturing cost reduction at scale (through whichever manufacturing paradigm proves viable), full-color optical performance at acceptable weight and field of view, and RX support to access the full addressable population. The industry has the manufacturing heritage — in LCD, in flat panel display, in semiconductor processing — to solve each of these. What it has lacked is a shared roadmap.

The ITAR-AR initiative is an attempt to provide one. Whether it attracts the breadth of participation needed to function as a true industry standard-setter, rather than a Meta-centric exercise, will determine whether the analogy to ITRS holds. The standing-room-only crowd at this panel suggests the appetite is there.