LetinAR将AI眼镜光学模组产能扩大五倍

LetinAR CEO Kim Jae-hyeok. (Photo: THE ELEC)

LetinAR is expanding its optical module production capacity fivefold in preparation for the opening of the artificial intelligence (AI) glass market. The move represents a preemptive investment as co-development projects with global Big Tech companies progress beyond technology validation to the mass-production preparation stage. The company also plans to pursue an initial public offering (IPO) under the special technology listing system next year.

"We plan to expand our optical module production capacity from the current 60,000 units per year to 300,000 units," said Kim Jae-hyeok, CEO of LetinAR. "We are preparing facility and equipment investments to begin supplying at the expanded capacity starting late next year."

LetinAR develops and manufactures optical engines and waveguides that deliver display images directly in front of the user's eyes. Leveraging its proprietary PinMR (Pin Mirror) and PinTILT technologies, the company reduces the thickness and weight of optical modules while delivering full-color images.

Unlike competing technologies that primarily rely on diffraction, LetinAR controls the light path using a reflective method. Because diffraction-based methods are wavelength-dependent, they require separate structures or layers for each color to achieve full-color displays. In contrast, the reflective method has relatively low wavelength dependence, making it advantageous for producing multiple colors within a single optical system.

At CES next year, LetinAR will unveil an optical module for consumer AI glasses with a thickness of 3.6mm to 3.8mm. The thickness of its existing industrial products ranges from 6.7mm to 7mm. The company plans to apply plastic materials to the new product to significantly reduce its weight compared with competing products. As for the light source, it primarily uses OLEDoS (OLED on Silicon), a self-emissive display with strengths in contrast ratio and power consumption. Depending on customer requirements, LEDoS or LCoS can also be applied.

"Industrial customers required a wide screen and high reliability, even if the product was somewhat large," Kim explained. "Recently, however, customers prefer products that are thin and light like regular glasses and provide a full-color screen, even if the features are limited."

LetinAR gained mass-production experience of more than 10,000 units by launching the "Miruza" augmented reality (AR) glasses in 2024 with a subsidiary of Japan’s NTT Docomo. It also supplied optical modules for auxiliary monitor glasses to Japanese PC maker Dynabook.

Recently, the company has been co-developing consumer AI glasses with global Big Tech firms. While current volume per project ranges from 100 to 10,000 units, the final product scale under review by customers ranges from 100,000 to over 1 million units. LetinAR is working to transition beyond simple development and technology validation into mass production.

For volumes exceeding 1 million units, LetinAR is considering sharing production with domestic and overseas manufacturers. Under this approach, the company would handle core processes and an annual volume of 300,000 units internally to manage quality, while utilizing external manufacturing networks for larger volumes.

For facility investments, LetinAR will deploy 27.8 billion won raised in a pre-IPO round. Its cumulative funding raised stands at approximately 67.8 billion won. LetinAR has selected its IPO lead manager and is preparing for a special technology listing next year. Last year’s revenue was approximately 2 billion won, and the company expects a similar level this year. The timing of large-scale mass-production revenue will depend on customers' product launch schedules.

LetinAR projects that AI glasses will grow as smartphone-connected companion devices over the next two to three years before evolving into an independent device platform around 2029.

Q&A

Q: What kind of company is LetinAR?

A: We founded the company with the goal of creating a world where smart glasses become part of daily life. Currently, we develop, manufacture, and sell optical modules, which are key components of AI glasses. Smart glasses and AI glasses have faced commercialization hurdles due to the lack of core component technology and high implementation difficulty. LetinAR is playing a role in solving these problems to accelerate market commercialization.

Q: It has been over 10 years since Google Glass came out. What has changed compared to back then?

A: Technology, markets, and services have all changed. In the early days of Google Glass, executable features were limited, and the product shape was vastly different from regular glasses. Consumers questioned why they needed features in glasses when smartphones were so convenient, and there were few services worth using on smart glasses. Today, as AI services expand, awareness has grown that smart glasses are the most suitable devices to use them. In fact, calling them 'AI glasses' has become more appropriate. Social acceptance of taking and sharing photos has also changed compared to the early days of Google Glass, driven by the spread of social media like Instagram and Facebook.

Q: Do you believe the AR glass product category can actually succeed?

A: I believe the AR glass market will eventually open. That is also why I have been doing business in this field for 10 years. Adding information to a person's field of view offers essential utility. Glasses are human's oldest wearable device, used since the 15th century. Compared to other wearable forms like rings, glasses offer far higher utility. There is no disagreement on the direction itself; it is simply a matter of when the market opens. Accelerating that timeline and building the market is the role of technology companies.

Q: Why was Meta's AI glass a turning point in the market?

A: Because it came closer to a consumer device by paying great attention to design. Meta fit as many features as possible into the Ray-Ban brand design. To prevent the product from becoming too large, heavy, or expensive, they partially limited performance and applied the display to only one eye. Nevertheless, they created a working product with necessary features like subtitles and real-time translation. It is significant in that they built a form that regular consumers can wear.

Q: Aren't current AI glasses heavier than regular glasses?

A: Light regular glasses weigh under 10g, and typical ones weigh 20 to 30g. Meta Ray-Ban display products weigh about 70g, which is quite heavy for glasses. In particular, they can constantly slide down when worn by Asians. However, even people who don't wear prescription glasses wear sunglasses when necessary. I understand that over 99% of people wear sunglasses in situations where eye protection is needed. If we provide users with sufficient value and a clear reason to wear them, wearing glasses itself will not be a major barrier.

Q: Can AI glasses go beyond smartphone companion devices to become independent devices?

A: For the next two to three years, I expect them to remain in a role that complements smartphones. Currently sold products are closer to a different form of wireless earbuds rather than full-fledged smart glasses. They provide basic earphone and microphone functions, with cameras occasionally added. In the long term, the reason AI glasses can become independent devices is that they can understand the user's field of view. Necklace- or wrist-type AI devices can record a user's daily life, but it is difficult for them to know what the user is looking at.

Q: Why is it important to understand what the user is looking at?

A: When wearing glasses, the person you are conversing with or the object of interest is located where you are looking. Gaze data can be the most suitable information for AI to understand the user. Once AI understands the user's interests and context, it can offer necessary suggestions without the user prompting first. This is a unique function that only smart glasses can provide compared to other AI devices.

Q: What components does LetinAR develop for smart glasses?

A: We focus on optical modules for smart glasses. Semiconductors can derive from smartwatch products, and batteries or other components can leverage existing electronic device technologies. Optical modules represent a dedicated technology that has rarely been attempted outside of smart glasses. The core is enabling users to comfortably view a large screen while maintaining a design similar to regular glasses.

Q: What is LetinAR's core optical technology?

A: PinMR and PinTILT. When customers build smart glasses or AI glasses, these technologies support delivering full-color images and a sufficiently large screen while maintaining a design similar to regular glasses. Some smart glasses currently on the market cannot provide color screens. LetinAR provides technology that enables building products that combine design, full-color screens, and price competitiveness.

Q: Are there products commercialized using LetinAR's technology?

A: In 2024, we launched 'Miruza' with a subsidiary of Japan's NTT Docomo. It was not a prototype of 100 or 200 units, but a product that went through mass production of over 10,000 units. We exported and sold it in Japan. Japanese PC brand Dynabook also purchased LetinAR products to create auxiliary monitor glasses called 'dynaEdge.' We are already mass-producing and supplying products in special-purpose and industrial markets.

Q: What are the characteristics of the industrial AR glass market?

A: Industrial customers demanded wide screens and high reliability, even if the product was a bit large. Guiding manufacturing or maintenance workers through procedures or providing safety training are representative uses. While products are already being sold in the industrial market, there was a limitation in that market size did not expand significantly.

Q: How have customer demands changed recently?

A: Demands changed significantly after the release of Meta's AI glasses. In the past, there were many requests to make the screen larger even if the product was somewhat large. Now, customers prioritize attractive design over screen size. The requirement is that it must look like regular glasses, even if features are minimal. Requests also emerge that offering full-color alone can differentiate a product when competitors offer monochrome green screens. In response, we are preparing a product lineup for consumer AI glasses.

Q: How does the product to be unveiled next year differ from existing industrial products?

A: Existing products emphasized screen size, resulting in optical module thickness of 6.7mm to 7mm. New products for AI glasses are under 4mm, at 3.6mm to 3.8mm. This is much thinner compared to competitors' optical system thickness of about 7mm. By using plastic materials, we plan to reduce weight to about one-fourth while achieving equivalent or superior performance.

Q: When will the new product be unveiled?

A: We plan to fully unveil it starting at CES early next year. What we have shown so far is a design mock-up of the product to be unveiled next year. The actual product will be manufactured in a similar form. Even after placing the optical module inside the glasses, we secured space to place the battery and other components. We believe it will appeal to consumers in terms of design and performance compared to existing AI glasses.

Q: Do LetinAR products support full-color displays?

A: LetinAR products support full-color. Products recently released by competitors often adopt monochrome.

Q: What is the difference compared to competing optical technologies?

A: LetinAR possesses an optical system that primarily uses reflection. Competitors' products often use diffraction. The reflective method reflects light from surfaces like mirror coatings. The diffractive method forms nanometer-scale protrusions or patterns, and light changes direction according to wavelength as it passes through that structure.

Q: What difficulties does the diffractive method face in achieving full color?

A: The diffractive method is highly wavelength-dependent. To implement multiple colors, separate structures or layers must be built for each color. This complicates the manufacturing process and can create issues in price and performance. The reflective method has relatively low wavelength dependence, making it advantageous for achieving full color.

Q: What displays are used in the optical engine?

A: Currently, we primarily use OLEDoS. Depending on partner and customer requests, we also use LEDoS or LCoS. Customers that have their own display business or display subsidiaries ask us to apply specific displays. LetinAR flexibly designs optical systems to meet customer requirements.

Q: Isn't OLEDoS more expensive than LCoS?

A: The price difference between OLEDoS and LCoS displays themselves is not very large. LCoS panels are relatively cheap, but to make them into an optical engine, LED light sources and structures must be added to the back. Because cost-increasing factors occur during this process, the price gap narrows at the finished optical module stage.

Q: What are the advantages of OLEDoS?

A: OLEDoS is self-emissive, so it does not require a separate light source. It offers a high contrast ratio and low battery consumption. AR glasses are not devices with screens turned on all day, but are used only when needed. In such operating environments, OLEDoS holds a competitive edge.

Q: What was last year's revenue?

A: Last year's revenue was approximately 2 billion won.

Q: What is this year's revenue target?

A: We expect it to be at a similar level to last year. We consider this a current weakness and risk factor for the company.

Q: What is the outlook for next year's performance?

A: It is difficult to provide specific figures for this year and next year's revenue. Customers we co-develop products with are basically reviewing final volumes of 100,000 to over 1 million units. The question is at what point customers decide on mass production and convert to actual supply contracts. We share customer plans and mass production is within sight, but schedules often shift.

Q: At what stage is the development underway with customers?

A: We are working to transition beyond simple product development or technology validation into the mass-production stage. Revenue is also being generated during the co-development process. However, it is difficult to state specifically at present when large-scale mass-production revenue will occur.

Q: When do you expect proper consumer AI glasses to appear?

A: We expect it before 2030. We are discussing the same direction with major customers. Although the timing for LetinAR to reveal products to the public is next year, we have already signed development contracts and are building products with major Big Tech firms. By 2029, we believe products that overwhelmingly surpass current AI glass performance could emerge.

Q: Can AI glasses become the next device after smartphones?

A: Smartphones have already entered a mature stage. Sales volumes are declining and companies are responding by raising unit prices, but a new device platform is needed. While several candidates are mentioned, we believe no product has as much potential as smart glasses. Recent advances in AI technology have made the possibility of smart glasses becoming the next-generation platform clearer than before.

Q: How do you evaluate the moves of global Big Tech companies?

A: Google unveiled screenless AI glasses and announced it would additionally reveal products with screens. Apple also views AI glasses as an inevitable future. Partner ecosystems are also forming, such as Google-Samsung-Gentle Monster and Meta-EssilorLuxottica. These moves show that the AI glass market could form faster than expected. We expect clear changes to appear in the market by 2029.

Q: How much did you raise in the pre-IPO?

A: We raised 27.8 billion won in the pre-IPO. Cumulative funding raised is approximately 67.8 billion won.

Q: Where will the pre-IPO funds be used?

A: The largest destination is capacity expansion. We plan to invest in facilities and equipment to produce optical engines and optical modules. We are preparing from various angles so that technologies and manufacturing processes developed so far can be expanded into mass-production lines.

Q: What is your current optical module production capacity?

A: We can produce approximately 60,000 units per year. Considering volumes discussed with customers and internal plans, we plan to first expand to 300,000 units per year.

Q: Why are you expanding production capacity fivefold?

A: While current project volumes range from 100 to 10,000 units, the final product scale reviewed by customers exceeds 1 million units. We determined that LetinAR must be able to produce around 300,000 units per year internally to maintain stable quality and product standards.

Q: How do you plan to respond to orders of over 1 million units?

A: We are reviewing whether LetinAR will directly produce the entire volume or share production in collaboration with domestic and overseas manufacturing companies. We are looking at handling core volumes of around 300,000 units internally to manage quality, while utilizing partners for larger-scale volumes beyond that.

Q: When will the 300,000-unit annual production system be established?

A: In terms of revenue and supply timing, we are preparing to respond in line with expanded production capacity starting late next year. Investments in production facilities and equipment, as well as preparations to expand existing technologies and manufacturing lines, are already underway.

Q: Does supplying technology to multiple competing customers create conflicts of interest?

A: It is not that there are no interest issues at all. There have been cases where we received certain fees from customers and granted priority rights for a specific period. Conditions when transitioning to large-scale mass production are a matter of individual contracts and negotiations. In some cases, customers request priority investment rights in contracts signed with LetinAR.

Q: What do priority investment rights entail?

A: If a competitor attempts to invest in LetinAR, it is the right of the existing customer to negotiate investment first. Contracts containing these terms exist, while other customers do not have them. Contract conditions vary by customer.

Q: Which side is more proactive among domestic and overseas customers?

A: Currently, overseas customers are more proactive.

Q: Why are overseas customers more proactive?

A: Global companies feel a strong need to take the lead in the AI glass market or beat competitors with better products even as latecomers. They need to secure a next-generation platform after the maturation of the smartphone market, and they also worry about their services becoming dependent on other platforms. For this reason, they invest in various technologies and expand cooperation by entering development contracts early if they deem a technology core.

Q: Is there a risk of becoming dependent on a specific customer?

A: Currently, the market is not fully formed, but at a stage where we are building the market together. If we conduct business in an all-out competition manner with every customer, risk management factors increase from the customer's perspective, lowering business momentum. In the early stage of the market, contracts that give customers priority for a certain period to focus on product development may be necessary. However, contract periods should be set short to reduce the risk of long-term dependence on one or two companies.

Q: How do you respond during the period when priority is granted to a customer?

A: During the promised period, an approach is needed where we focus fully on product development with that customer to raise perfection. Afterwards, as the market grows and demand increases, both the customer and LetinAR must respond flexibly according to changed market conditions.

Q: Can LetinAR secure supply bargaining power when large-scale demand occurs?

A: LetinAR possesses core technology and has invested in technology for over 10 years. Beyond technology, we possess design and manufacturing capabilities. When demand surges, prices and supply conditions can also change. Based on accumulated technology and production capabilities, we believe we can present sufficient competitiveness to customers.

Q: What hardware challenges must AR glasses overcome?

A: Design, price, and performance must be solved together. Because the three elements are interconnected, it is difficult to create a commercial product by improving only one.

Q: Unlike smartphones, don't AI glasses require diverse designs?

A: Existing IT devices evolved by converging into one or two designs. Glasses are both IT devices and fashion products. Diverse products are bound to emerge depending on consumer tastes and services used. Even if not complete high-mix low-volume production, we must produce more diverse products more flexibly than smartphones.

Q: What capabilities do AI glass component makers need?

A: Glasses are small, so there is insufficient space to fit components. Whenever product design and features change, the structure of optical modules can also change. How flexibly component makers can respond to customers' diverse design, feature, performance, and price requirements is the core competitiveness.

Q: What is an acceptable AI glass price for consumers?

A: For full-scale popularization, early products must be released at the $399 to $499 level. Mass sales are possible only at this price point. Recently, domestic telecommunications companies have also begun selling smart glasses through monthly installment plans. If methods of paying prices over multiple months spread like smartphones, consumer burden can be reduced.

Q: Are current AI glasses also at the $399 to $499 level?

A: Current AI glasses that lack screens and have limited features are also mostly positioned in that price range. Meeting this price range is important to form the early market.

Q: What is the required killer service for AI glasses?

A: It can be an AI assistant that understands the user's daily life and surrounding context based on on-device AI. On the premise of complying with legal regulations, it provides necessary content by understanding the user's conversations, field of view, and location. When a user asks, 'What did we talk about in last week's interview and what were the risk factors?', 'What did we decide in the last meeting?', or 'Where did I go yesterday?', the AI can summarize relevant details.

Q: Do translation and healthcare features also get included?

A: It can expand into various functions supporting users' daily life and work, including translation and healthcare. Ultimately, it will become close to an AI assistant like Jarvis in the movie Iron Man.

Q: When will an AI glass like Jarvis be possible?

A: I believe meaningful services can emerge by 2029. While there will be differences depending on the level of information expression, resolution of the device, and performance of AI models, we expect to provide practical AI assistant services starting in 2029.

Q: What are your initial public offering plans?

A: We are preparing with the goal of listing under the special technology listing system next year. We have also completed the selection of IPO lead managers.

Q: What is the most important task ahead of listing?

A: It is converting development projects currently underway with global customers into mass-production supply and expanding production capacity. Before the AI glass market fully opens, it is important to secure a mass-production foundation and customer base, and demonstrate revenue growth potential.

Q: What is LetinAR's long-term goal?

A: It is supplying core optical technology so that smart glasses can establish themselves as a new device platform following smartphones. Based on original technology and design and manufacturing capabilities, we will respond to diverse customers' design and service requirements and accelerate commercialization of the AI glass market.