Three Technical Paths for AR Optical Waveguides: Which Will Emerge Victorious

In September 2024, Meta unveiled Orion, its AR glasses prototype. The moment Zuckerberg put them on, a full-color image with a 70-degree field of view appeared in the lenses—free of rainbow artifacts or ghosting, with virtual content blending seamlessly into reality like ink in water.

Meta optical scientist Pascual Rivera once described the experience this way: "Wearing glasses with traditional glass waveguides is like being in a disco—your vision is filled with rainbow-like artifacts, making it nearly impossible to see the AR content clearly. But the moment you switch to silicon carbide waveguides, it’s like stepping into a symphony hall—everything goes quiet, and you can finally become fully immersed."

This statement cuts straight to the core pain point of the AR optical waveguide industry: the challenge isn't simply manufacturing them, but manufacturing them well. The AR optical waveguide—a transparent lens less than 1 millimeter thick—represents the "last mile" in determining whether AR glasses can truly make the leap to the consumer market.

The industry has currently converged on three mainstream technical approaches—Surface Relief Grating (SRG) waveguides, geometric array waveguides, and Volume Holographic Grating (VHG) waveguides. Each relies on different physical principles and faces its own unique engineering bottlenecks.

Of these three paths, which one will ultimately prevail?

First, let’s clarify: what problem do optical waveguides actually solve?

Before discussing the three technical approaches, we must first establish the fundamental nature of optical waveguides. The optical system of AR glasses needs to accomplish two things simultaneously: delivering virtual images to your eyes while not obstructing your view of the real world.

This means the optical components must be sufficiently transparent and thin, while also providing a wide field of view (FOV) and a generous eyebox.

Early approaches—such as prisms, freeform optics, and Birdbath designs—all ran up against the same physical trade-off: increasing the FOV required a thicker optical module, whereas making the module thinner meant sacrificing light transmission.

The prism design used in Google Glass offered an FOV of only 15° and less than 50% light transmission. Epson’s freeform solution weighed 550 grams—resembling a helmet—while the Birdbath design used by XREAL offered less than 30% light transmission, making the glasses look like dark sunglasses.

Optical waveguides take a completely different approach: the optical path is folded into a thin sheet of glass and transmitted via total internal reflection, after which specific structures "extract" the light and direct it into the eye. This decouples the optical module's thickness from the FOV, allowing for lenses thinner than 3mm, light transmission exceeding 85%, and a theoretical FOV surpassing 80°.

The challenge lies in how the light is "coupled in" and "coupled out." The distinction between the three different technical routes lies precisely in the physical mechanisms used for this "in-coupling" and "out-coupling."

Surface Relief Grating (SRG) Waveguide: The King of Mass Production, Yet Plagued by Rainbow Artifacts

Principle: Etching "nano-gratings" onto the glass surface. The core of an SRG lies in fabricating periodic, nano-scale relief structures—essentially diffraction gratings—on the surface of a glass substrate.

Light emitted from the micro-display undergoes diffraction upon encountering the input-coupling grating; the diffraction angle satisfies the condition for total internal reflection, thereby "trapping" the light within the waveguide for propagation. Upon reaching the output-coupling grating, the light diffracts again and is "released" to enter the human eye.

Furthermore, by engineering the grating period and duty cycle, "eye-box expansion" can be achieved—enlarging the exit pupil from just a few millimeters to over 10 millimeters—allowing the user to view the image without needing to precisely align their eyes with a specific spot.

Advantage: Mass-production capability far surpasses the other two approaches. The greatest strength of SRG technology is its ability to leverage mature semiconductor manufacturing processes.

Nanoimprint lithography (NIL) enables the high-precision replication of grating master structures onto wafers; a single 12-inch wafer can yield 20 to 25 waveguides.

Goertek Optics has achieved a production capacity of 25,000 units per month (based on 12-inch wafers), resulting in an annual output of approximately 600,000 to 750,000 waveguides. In 2025, global revenue from SRG waveguides reached $245 million, securing a dominant 62% market share.

Chinese companies have driven the per-unit cost below $18 while boosting the yield rate to 86%.

Flagship AR devices—such as Microsoft’s HoloLens, Magic Leap, and products from domestic brands like RayNeo and Meizu—all utilize SRG technology.

A fatal flaw: the physical nature of the "rainbow effect." The most criticized issue with SRGs is the rainbow effect.

This is not a manufacturing defect, but an inherent physical characteristic.

Grating diffraction causes light of different wavelengths to diffract at different angles. When viewing the outside world through an SRG waveguide, the various wavelength components of ambient light are "split" by the grating, creating colored fringes in the field of view. It is akin to the rainbow-like reflection seen on the back of a CD—the same physical mechanism is at work.

Another issue is light leakage.

The out-coupling grating not only directs the image light toward the user's eyes but also scatters some of it externally.

If someone looks at your glasses from the side, they can see the flickering image—a privacy issue that is unacceptable in real-world usage scenarios.

There is also the problem of low optical efficiency. The light utilization rate of SRGs is typically only 1%–5%, meaning that over 95% of the light emitted by the micro-display is wasted; this directly drives up power consumption and increases the brightness requirements for Micro-LEDs.

Technological Breakthroughs in 2025–2026

The industry is not sitting idly by. Etching processes are replacing nanoimprint lithography as the next-generation direction for SRGs.

Nanoimprint lithography utilizes organic photoresists, limiting the refractive index to around 2.0; in contrast, etching processes work directly on inorganic materials, enabling a refractive index exceeding 2.4 and fundamentally enhancing optical performance.

AAC Technologies partnered with Dispelix to launch a single-layer, full-color waveguide—manufactured using semiconductor etching processes—featuring a 25° field of view (FOV), a light efficiency of 1500 nits/lm, a thickness of 0.7 mm, and a weight of just 4 g; it offers over 90% light transmission and suppresses rainbow artifacts to a level that "does not affect the user experience."

Meanwhile, Zhige Technology claims to significantly suppress rainbow artifacts in 95% of usage scenarios and has secured investment from Xiaomi and OPPO.

However, to be frank, rainbow artifacts can only be suppressed, not eliminated—as long as the grating structure exists, dispersion persists. This is an inherent limitation of the SRG (Surface Relief Grating) approach.

Geometric Array Waveguide: The pinnacle of image quality, but a nightmare to mass-produce.

Principle: Using an array of semi-reflective/semi-transmissive mirrors to "relay" light rays.

Geometric waveguides take a completely different approach—relying purely on geometric optics without involving diffraction.

Its structure is simple: an array of tilted semi-reflective mirrors is embedded within a glass substrate. Light enters the waveguide via an input coupling prism and propagates through total internal reflection; each time it encounters a semi-reflective mirror, a portion of the light is reflected out of the waveguide and into the user's eye, while the remaining light continues to the next mirror. It acts like a set of "relay mirrors," "pouring" the light out step by step.

Since they do not rely on diffraction, geometric waveguides are free from rainbow artifacts and dispersion issues, offering the highest image quality among the three approaches.

Advantages: Superior image quality and minimal light leakage.

Geometric waveguides offer far superior optical efficiency compared to SRGs; they eliminate diffraction losses and ensure high optical path efficiency, thereby reducing the power consumption of the optical engine. With a light leakage rate of only 1%–5%—the lowest among the three technologies—they offer excellent privacy, making the display content virtually invisible to observers viewing the glasses from the side. They lead across the board in metrics such as color reproduction, contrast, and resolution; to date, Lumus’s geometric waveguides remain unrivaled in image quality. Projected to hold a global market share of approximately 28% by 2025, they dominate B2B sectors—such as industry and the military—that demand exceptionally high image quality.

Fatal Flaw: The Mass-Production Nightmare of Coating and Bonding

The challenge with geometric waveguides lies in manufacturing.

The core process is precision optical fabrication: cutting glass prisms, grinding, polishing, coating, and gluing or bonding. While each step involves traditional optical techniques, the precision requirements are pushed to the absolute limit.

Coating represents the first major hurdle. Each semi-reflective/semi-transmissive surface requires precise control of the reflection-to-transmission (R/T) ratio. Furthermore, these R/T values must vary across different locations; since light loses energy as it propagates, surfaces further along the path require higher reflectivity to ensure uniform light output. Coating thicknesses are controlled at the nanometer scale, and a single glass substrate may feature dozens of coating layers.

Bonding is the second major hurdle.

Traditional bonding relies on optical adhesives; however, the curing shrinkage of polymeric adhesives introduces stray light and color shifts, and high-refractive-index optical adhesives are scarce. Molecular bonding (direct bonding) offers superior results but demands extremely stringent process conditions—requiring atomically flat surfaces, where even the slightest particle or topographical irregularity leads to bonding failure.

The result: yields remain low, and costs stay high.

In 2025, the mass-production yield for geometric waveguides remains significantly lower than that of SRGs; this is the fundamental reason why their market share is less than half that of SRGs.

Who is staying the course?

Lumus (Israel) is the benchmark for geometric waveguides, leading the field in high-refractive-index glass wafers through its partnership with SCHOTT.

Reports suggest that Meta’s 2025 AR products may utilize an LCoS plus geometric waveguide configuration; if true, this would mark a major breakthrough for geometric waveguides in the consumer market.

Key domestic Chinese players in this space include companies such as Libaiguangjing and Lingxiweiguang.

The position of geometric waveguides is much like that of EUV technology in lithography systems: they offer the highest performance ceiling but also present the highest barriers to mass production.

If manufacturing bottlenecks are overcome, they could represent the ultimate solution for image quality. Until then, however, they remain primarily a high-end option for the enterprise (B2B) market.

Volume Holographic Grating (VHG) Waveguides: A Cost-Killer, but Materials Are the Critical Bottleneck

Principle: "Writing" a refractive-index grating inside the material

The concept behind VHGs differs fundamentally from that of SRGs.

While an SRG is a surface-relief grating formed by physically etching a structure onto the surface, a VHG is created via dual-beam interference exposure. This process induces periodic modulation of the refractive index within a photosensitive material (such as a photopolymer)—effectively "writing" an invisible grating into the material's interior.

Because the grating exists within the bulk of the material (rather than on the surface), it is termed a "volume hologram."

This mechanism offers several inherent advantages:

1. Higher theoretical diffraction efficiency: The Bragg selectivity of volume gratings enables the efficient diffraction of light at specific wavelengths and angles, with a theoretical diffraction efficiency approaching 100%.

2. No surface relief structures: This eliminates dispersion mechanisms that cause rainbow artifacts and avoids issues with surface contamination.

3. Simple manufacturing process: It requires no nanoimprinting or etching—only a laser interference exposure system—resulting in lower theoretical costs and higher production capacity.

4. Minimal light leakage: Nica Optics' N25G product features a light leakage ratio of less than 1:140.

Advantage: Best aligns with the need for cost reduction and mass adoption.

Among the three pathways, VHG offers the greatest potential for low manufacturing costs.

No nanoscale fabrication equipment, precision coating, or molecular bonding is required—only a laser interference exposure system.

In June 2026, Nika Optics launched the world’s first automated production line for volume holographic optical waveguides in Tianjin, capable of producing one million units annually; the combined annual capacity across its Tianjin and Guangzhou facilities reached 1.3 million units.

In 2025, VHG (Volume Holographic Grating) held approximately 10% of the global market share but demonstrated the fastest growth rate.

XINGYIZHINENG’s AR99 glasses feature Nika Optics’ volume holographic waveguides, with the complete unit priced as low as 999 RMB—a price point that sent shockwaves through the industry.

A Fatal Flaw: The "Impossible Triangle" of Materials

VHG's Achilles' heel lies in its materials.

Photopolymer materials must simultaneously meet three criteria: high refractive index modulation (Δn), long-term environmental stability (resistance to aging, temperature, and humidity), and high fabrication precision. These three criteria are inherently contradictory, forming an "impossible triangle":

1. Increasing Δn typically requires the incorporation of high-refractive-index inorganic nanoparticles, which compromises material transparency and stability;

2. Enhancing stability requires cross-linking and curing, which restricts the response of photosensitive groups and reduces Δn;

3. Precisely controlling the refractive index profile demands an extremely narrow exposure bandwidth and strict environmental controls, making it exceptionally difficult to ensure consistency in mass production.

Consequently, VHGs generally offer a field of view (FOV) of only 30°–40°, and their color consistency and color gamut coverage for full-color displays still lag behind those of SRGs and geometric waveguides.

Furthermore, the issue of aging in photosensitive materials has not yet been fundamentally resolved. Will waveguide performance degrade after a few years? There is currently insufficient long-term field data to answer this question.

There is also a constraint at the supply chain level: high-precision laser interference exposure systems lack standardized, mass-produced versions. Consequently, each company must develop its own custom solutions, a process characterized by long development cycles and high investment costs.

Who is placing bets?

Global players include Sony, DigiLens (USA), and Akonia (acquired by Apple in 2018).

Domestic representatives include Nika Optics, Sanji Optoelectronics, Crystal-Optech, and Gudong Technology.

Apple’s acquisition of Akonia itself signals a key assessment: volume holography may well be the ultimate path for consumer-grade AR, though breakthroughs in materials science are a prerequisite.

Silicon Carbide: The Fourth Path? A "Dimensional-Reduction Strike" at the Material Level

After a decade of debate over three competing approaches, Meta dropped a bombshell in 2024 with the Orion: silicon carbide (SiC) optical waveguides.

This is not a fourth technical route; rather, it represents a "dimensional strike" at the material level against the first three routes.

Why silicon carbide is a game-changer

The maximum field of view (FOV) of an optical waveguide is determined by the refractive index of the substrate material: FOVmax∝arcsin⁡(nguide−nair)

The refractive index of ordinary glass is approximately 1.5–1.7, while the high-index glass used for SRGs ranges from 1.9 to 2.0. This directly limits the maximum field of view (FOV) for single-layer waveguides—the 52° FOV of the HoloLens 2 is already approaching the physical limits of glass materials.

Silicon carbide boasts a high refractive index of 2.65–3.2, enabling a 70°–80° field of view (FOV) with just a single-layer waveguide—eliminating the need for multi-layer stacking.

Even more crucially, the dispersion characteristics of silicon carbide allow the grating period to be optimized to 262 nm, physically eliminating rainbow artifacts—not merely suppressing them, but eliminating them entirely.

Then there is heat dissipation: silicon carbide has a thermal conductivity of approximately 490 W/m·K—nearly 500 times that of glass. Heat generated by the optical engine can be conducted away directly through the lens, eliminating the need for dedicated heat-dissipation structures and further reducing weight.

By using silicon carbide waveguides, Meta Orion achieves a 70° field of view (FOV) and a weight of under 100 grams, all while maintaining an appearance similar to ordinary eyewear—feats that glass waveguides cannot match.

However, cost is a major drawback.

The cost of each silicon carbide lens currently runs into the thousands of yuan, far exceeding that of glass waveguides. Meta executives have also acknowledged that a significant portion of Orion's production cost is attributed to the silicon carbide lenses.

Silicon carbide possesses extreme hardness (9.25 on the Mohs scale, second only to diamond), making it exceptionally difficult to process. Furthermore, manufacturing capabilities for large-sized, optical-grade silicon carbide substrates remain limited.

Domestic Chinese companies active in this sector include Sanan Optoelectronics, TankeBlue Semiconductor, Tianke Heda, and Guotan Semiconductor. By 2025, Guotan Semiconductor had successfully developed 12-inch high-purity silicon carbide single crystals and completed small-batch validation with multiple AR clients. Goertek Optics has also launched the F50Se, a full-color optical waveguide display module utilizing silicon carbide etching technology.

The essence of the silicon carbide approach lies in trading higher material costs for a generational leap in performance.

Its transition to the consumer market depends on the pace at which the cost of large-sized silicon carbide substrates can be reduced—a window of opportunity optimistically projected for 2027–2028.

Industry Landscape: Who Is Betting on What

The industry landscape as of mid-2026 has gradually become clear:

Several industry signals worth noting:

1. Manufacturing systems are beginning to converge.

Beyond 2026, industry competition will shift from "who can manufacture the product" to "who can ensure stable delivery." Domestic effective production capacity for SRGs (Surface Relief Gratings) stands at approximately 6–8 million units per year, yet demand exceeds 10 million units annually, resulting in a supply gap of several million units for high-quality waveguides.

2. Collaboration between optical giants and semiconductor equipment giants.

In June 2026, EssilorLuxottica and Applied Materials announced a joint effort to develop next-generation AR optical systems for smart glasses; this partnership—combining a traditional lens manufacturer with a semiconductor equipment provider—signals a shift in AR optics production from "laboratory-scale processes" to "semiconductor-grade manufacturing."

3. Goertek acquires Sunny O-Lye.

In early 2026, Goertek Optics and Sunny Optical completed a share-swap transaction valued at approximately RMB 1.9 billion, integrating full-chain capabilities ranging from design to manufacturing. Industry consolidation is accelerating.

The Final Verdict: Who Will Have the Last Laugh

Our assessment is that while there will be no single winner in the short term, the landscape for the medium term is already taking shape.

Short term (2026–2027): SRG dominates; VHG gains a head start in the mass market.

Leveraging mature nano-imprint and etching processes, an 86% yield rate, and a per-unit cost under $18, SRG will remain the mainstream solution for consumer AR over the next 1–2 years.

Rainbow artifacts are suppressed to "acceptable" levels, and a Field of View (FOV) of 25°–50° covers most consumer use cases.

VHG is rapidly penetrating the mass market (specifically the RMB 999 price tier for AR glasses) and the automotive AR-HUD sector, driven by its cost advantages.

The launch of Nika Optics' million-unit-scale production line serves as a key signal.

However, limitations regarding FOV and full-color display capabilities remain difficult to fundamentally resolve in the short term.

Geometric waveguides will maintain their high-end positioning in the enterprise (B-to-B) market while awaiting validation from consumer-grade products launched by industry leaders like Meta.

Medium term (2028–2030): Silicon carbide reshapes the landscape.

If the cost of large-format silicon carbide (SiC) substrates drops to an acceptable level (projected to take 2–3 years), SiC waveguides will simultaneously resolve the "rainbow effect" associated with Surface Relief Gratings (SRGs) and the Field of View (FOV) limitations of geometric waveguides, emerging as a "game-changing solution" for consumer-grade AR.

At that point, the focus of competition will shift from "grating fabrication processes" to "material platforms"—whoever masters low-cost, large-format, optical-grade silicon carbide will hold the key to the future of AR optics.

Long-term outlook

Ultimately, the competition among AR optical waveguides is not a zero-sum game between three distinct routes, but rather a race between two evolutionary paths:

Path 1: Maximizing manufacturing precision.

Through the continuous iteration of etching processes, high-refractive-index materials, and coating technologies, SRG and geometric waveguides are approaching physical limits within existing material systems.

Path 2: Material revolution.

New materials like silicon carbide are fundamentally changing the game by bypassing existing process bottlenecks and enabling a generational leap in performance.

VHG, however, may take a third path: once photosensitive materials break through the "impossible triangle," the cost advantages of volume holography could make it the ultimate solution for the consumer market. Apple’s acquisition of Akonia represents a strategic bet on this very direction.

The three approaches essentially address the same question from three different dimensions: performance, cost, and feasibility. SRG wins on feasibility, geometric waveguides excel in performance, and VHG offers the greatest cost potential, while silicon carbide attempts to win on all three fronts.

Who will have the last laugh?

It is not about any single specific route, but rather about being the first to simultaneously unlock the "high performance + low cost + mass producibility" triad.

This race has only just begun.

Source:shenduzhuiguang