Satoshi Shiraga, Co-Founder and CEO of Cellid – Interview Series

Satoshi Shiraga is the Co-Founder and CEO of Cellid, where he leads the company’s vision and strategy in developing cutting-edge technologies. He conducted research in particle physics at CERN, Fermilab in the United States, and INFN in Italy before founding the company. He holds a Master’s degree in Physics from Waseda University Graduate School, where he specialized in particle physics.

Cellid is a Japan-based company developing advanced augmented reality technologies, including ultra-light waveguide display modules and spatial recognition engines for next-generation AR smart glasses. Using proprietary optical design and production methods, Cellid creates components as thin as eyeglass lenses and partners with global manufacturers to accelerate the adoption of AR hardware.

What inspired you to pivot from particle physics research at CERN and Fermilab to founding Cellid in 2016?

After graduating from university, I joined the European Organization for Nuclear Research (CERN), where I conducted particle physics research that required immense computing power. At the same time, I explored cutting-edge developments in computer science. What sparked this interest was the idea of applying emerging computing technologies—such as neural networks and simulation tools—to real-world problems. phenomena.

Founding Cellid in 2016 gave me the opportunity to develop technology, which became the foundation of our work. Recognizing that AR glasses are the optimal device for maximizing the potential of spatial recognition services, we identified them as one of our core business areas. We began researching AR glasses as the optimal hardware for providing spatial recognition services. During our research, we spoke with engineers and discovered that the hardware used for the AR glasses’ display was a bottleneck. By observing the evolution of computer devices, we concluded that wearable AR glasses would become the next generation of smartphones. In this emerging market for new devices, we aim to leverage our state-of-the-art computing technologies to lead innovation.

How did your background working with petabyte-scale parallel computing inform your approach to AR display design and optical simulation at Cellid?

My background in particle physics research at CERN involved working with parallel computing systems to simulate and visualize complex physical phenomena. That experience directly shaped how we approach AR display design and optical simulation at Cellid.

Our waveguides are created by forming a diffractive optical element (DOE) on a substrate using nanoimprint technology. Of the three elements—design, materials, and manufacturing—we are responsible for the design. Leveraging my experience with high-performance computing, we developed a proprietary AI-based simulation engine that automates the design process. It not only simulates the material composition of the DOE and substrate but also predicts and optimizes the precision of film formation and processing. Cellid leads or collaborates with partners in the development of materials, production technologies, and processes. (Most of the manufacturing is based on Cellid’s expertise.)

This simulation-driven workflow enables us to explore massive design spaces quickly and efficiently similar to the parallelized simulations I conducted in particle physics. Once we finalize the design, we collaborate with materials and manufacturing partners to bring it into production.

AI plays a dual role at Cellid——on-device user experience and in AI-driven optical design/manufacturing how do you leverage AI for enhancing user experience and how do you use it in manufacturing and design?

Cellid makes maximum use of AI for both. To enhance user experience, we use AI to enable spatial recognition, real-time information presentation, and intuitive operability.

In design and manufacturing, we apply AI to optical design simulation and process optimization, and to support quality stability at scale—helping us design and mass-product  waveguides and optics that are thin, lightweight, and power-efficient for eyewear-style devices.

These two uses of AI aren’t directly linked, but both are essential to mainstream adoption: AI-driven design makes AR glasses wearable; AI-powered UX makes them useful..

What are some real-world examples of seamless AR use cases made possible by Cellid’s SDK?

We’re focused on tangible, everyday use cases where AR glasses can add real value now:

King Salmon Project – Remote Work Support in Construction & City Administration: In Tokyo’s King Salmon Project, Cellid partnered with Shibuya Ward to pilot AR glasses for remote QA in construction and local government settings. Field workers wore Cellid’s lightweight AR glasses (using the Reference Design). The use of Reference Design improved construction site operations by enabling remote supervision, real-time communication, and accurate material inspections. For supervision, managers and specialists could monitor progress, provide instructions, and consult on construction methods without being on-site. The glasses also displayed order details, allowing direct comparison with delivered items to detect discrepancies quickly. On-site efficiency improved through seamless communication between staff and remote experts, ensuring immediate feedback and guidance.

Convenience Store Pilot – Smart Shopping Experience: In collaboration with world known convenience store in Japan and SMBC Group, Cellid conducted a trial at an actual convenience store to demonstrate the shopping experience using AR glasses. More than 300 participants joined the test, which featured automatic product recognition, personalized recommendations based on items added to the cart (including showing discount coupons and directions to product locations), and payment through the AR glasses.

CREST Medical AR Project – Validating AR Glasses for Medical Use: Using AR glasses in clinical settings, this project conducts demonstration experiments at clinical sites to evaluate display methods that surface vital information to clinicians during procedures and  magnify images of treatment areas. Findings from these trials inform application software and interface methods for practical medical workflows. The work is part of Japan’s multi-year JST CREST program (“Realization of medical AR glasses using metamaterial technology,” FY2024–2029; grant JPMJCR24R1), a consortium with the Institute of Science Tokyo, Cellid, and Mitsui Chemicals spanning materials development, optical design, manufacturing process establishment, application-software implementation, and clinical verification.

Cellid is contributing its optical simulation, waveguide design, and software development expertise to create ultra-lightweight AR glasses suitable for clinical environments. The project includes clinical validation with surgeons to evaluate display methods that support—but do not obstruct—medical procedures. Prototype testing will inform both optical design and interface usability based on real-world surgical feedback.

These examples demonstrate how Cellid’s hardware and SDK deliver pragmatic, day-one value—improving efficiency, decision-making, and user experience in both retail and workforce environments.

How do your proprietary plastic-based waveguides compare to traditional glass alternatives in terms of weight, power, and comfort?

At Cellid, we develop both plastic and glass waveguides to address different AR use cases. Our plastic waveguides are significantly lighter than glass, approximately 5 grams per monocular unit versus around 8.2 grams for glass. This reduced weight contributes to overall wearability and comfort for everyday use.

In pilot evaluations, our plastic waveguides—featuring a 30° field of view—received positive feedback from users. At this FOV level, optical performance between plastic and glass is comparable.

Durability is another advantage of plastic. Cellid’s plastic waveguides have passed both U.S. and Japanese impact tests for eyewear without cracking. By contrast, glass typically requires additional reinforcement to meet similar safety standards. This is evident from the fact that ordinary eyeglass lenses are made of plastic.

While plastic is well suited for many everyday use cases where lightweight, compact AR devices are essential, glass still offers benefits in some scenarios in industrial environments where higher optical clarity is prioritized. By offering both material platforms, Cellid can address to accelerate further drive adoption of AR glasses to meet various usage.

What were the biggest engineering or materials science challenges in integrating these components into a wearable form factor?

The key challenge was miniaturization and fitting them into stylish eyewear form without compromise. We had to rethink how optical elements like waveguides and projectors could be compacted without degrading performance. Materials had to be thin yet durable, optically efficient yet manufacturable at scale. In addition, developing specialized materials for the waveguide and a proprietary manufacturing process could present challenges. Addressing all these factors required close integration between simulation and manufacturing.

Can you walk us through how your simulation platform works and how it reduces iteration time and manufacturing costs?

At Cellid, simulation plays a critical role in accelerating development and reducing the cost of bringing AR hardware to market. We use proprietary tools to evaluate complex optical components—like waveguides, diffractive optical elements, resins, and coatings—before any physical prototypes are made. By simulating how these materials perform under different conditions, we can optimize for clarity, field of view, and efficiency early in the design process. This helps minimize reliance on trial-and-error fabrication, shorten iteration cycles, and lower production costs.

What features does Cellid’s SDK offer developers, and what’s next on the roadmap?

Cellid provides SDKs for importing camera-recognized images, displaying AR images in specified spaces, displaying smartphone app data, and operating hardware implemented in AR glasses, such as microphones and speakers. In addition, the Cellid SDK for the reference design includes an API for generative AI, which enables its use in a wide range of use cases. For example, by asking through the reference design how to use a device in front of you, the device can be automatically recognized, and the usage and operating instructions can be displayed. In addition, by registering content to AR markers, it is possible to recognize the AR markers and display the content in real space, allowing users to instantly check equipment manuals and other hands-free information. Furthermore, additional eye tracking and gesture recognition functions are planned for the future.

These efforts are focused on enabling faster development, broader device compatibility, and tighter alignment between hardware performance and software responsiveness—all critical to delivering next-generation AR wearables. As a developer of AR glasses and spatial recognition software, Cellid is uniquely positioned to provide an environment that includes the SDK for rapid development of applications suitable for AR glasses.

Which industries are you most focused on, and why are they best suited for near-term AR deployment?

The AR glasses market is currently in its early stages, with adoption being driven by practical applications for display information, such as notifications, weather updates, translations and integration with generative AI. In these cases, small, lightweight devices that can provide sufficient user value even with relatively narrow viewing angles are expected to support initial market growth..

Looking ahead, demand is expected to increase for more immersive experiences such as video viewing, 3D content and interaction with real space (spatial computing). Consequently, the importance of wide viewing angles and high-definition display performance will continue to grow.

From this perspective, we identify two near-term segments in which waveguide-based AR glasses offer clear benefits:

1. Ophthalmic / everyday use

Waveguide advances make eyeglass-shaped (“ophthalmic”) AR glasses practical—same general shape as regular eyewear—positioning them as the likely next major device after smartphones. As with smartphones, this form factor supports broad, day-to-day use by individuals and companies across many applications. As waveguides become more compact and cost-effective, they drive everyday consumer use—navigation, subtle prompts, and other on-the-go tasks. In retail, our pilot is an example of our focus

2. Enterprise & industrial operations

Efficient waveguides enable real-time data overlays for technicians and field staff, helping reduce errors and improve productivity in sectors such as manufacturing, logistics, and healthcare. These “information-display” scenarios fit today’s power and field-of-view constraints while keeping devices lightweight.

What does “everyday AR” look like five years from now?

With the advent of multimodal generative AI, AR glasses will experience rapid growth as one of their input/output interfaces.

We envision a future where AR glasses become an intuitive, always-on companion—blending the digital and physical world through seamless interaction and lightweight design.  Driven by the evolving usage of generative AI, AR glasses will also evolve to become a natural interface for accessing real-time insights, automating tasks, and receiving contextual assistance..

At Cellid, we believe achieving everyday usability hinges on three pillars: comfort, intelligence, and power efficiency. That’s why our reference designs prioritize ultra-lightweight optics alongside slim form factors that resemble regular eyewear. This ensures AR glasses can be worn for extended periods without fatigue.

Looking ahead, we’re continuing to co-develop an open ecosystem with global partners to ensure AR glasses not only look and feel like everyday accessories but also deliver lasting value across industries. Whether navigating cities, supporting field work, or enhancing retail experiences, AR will increasingly shift from a novel interface to a practical tool that fits naturally into everyday life. Ultimately, AR will redefine how people engage with both the physical and digital worlds, becoming a powerful companion to human experience.

Thank you for the great interview, readers who wish to learn more should visit Cellid.