The consumer electronics market, with smartphones at the forefront, is witnessing a significant resurgence. As domestic optical products strive towards high-end development, the entire industrial chain is gradually coming into place. Chinese manufacturers are racing to lay out their plans for VR and AR optical lenses and accessories, while the localization rate in sectors such as automotive products and machine vision is steadily increasing.

Upgrades in smart phone capabilities, particularly in photography and biometric functions, have escalated the technical demands for optical lenses. Innovations in zoom, high resolution, optical stabilization, and overall reliability have become imperative. As a result, periscope cameras have emerged as a prominent trend in development. The competition among optical lens manufacturers in the mobile and security industries has intensified, while the increase in automotive optical products has led to a decrease in average pricing. Companies are under pressure to maintain competitiveness across various parameters, including research and development, production processes, scale, key clients, and margins. Consequently, some firms are extending their business activities toward automotive optics and VR/AR technologies, seeking out a secondary growth curve.

To meet the burgeoning requirements of advanced driver assistance systems (ADAS), car manufacturers are continually incorporating a multitude of cameras for forward, surround, rear, and interior vision. The growth of domestically produced new energy vehicles has also played a significant role in reducing the costs associated with automotive heads-up displays (HUDs), boosting production rates, and enabling the mass production of domestic AR-HUD products that now fully cover mainstream technology pathways.

In the field of machine vision, upstream components and software algorithms account for a staggering 80% of total value. The maturation of the camera and sensor industries has broadened the application scenarios for machine vision. Integrating AI technologies—such as deep learning, machine learning, and large language models—into 3D visual perception has led to enhanced efficiency in data processing, feature extraction, and recognition. This application of AI technologies has endowed intelligent systems with greater adaptability and intelligence, resulting in symbiotic growth between machine vision and upstream industries.

Moreover, the innovation within smartphone camera technology remains pivotal. Newly designed smartphones are relying on a synergy of two to three fixed focal length lenses to achieve optical zoom capabilities. Here, long focal length lenses are central to dual or multi-camera setups. The trend is that the higher the zoom factor, the greater the dimension of the long focal length camera. The periscope camera, utilizing a horizontally oriented design within the phone, employs specialized light direction techniques through micro prismatic systems to facilitate the refraction of light into the lens group, thereby extending the spatial choices available within the lens configuration. This crucially resolves issues concerning optical path length, preventing increased thickness in the smartphone body that typically accompanies zoom lenses. Ten years ago, companies like Sunny Optical Technology, Huawei, and others began exploring the periscope camera technology. The primary components of periscope lenses include prisms, lenses, image sensors, and voice-coil motors—all of which overlap with traditional camera parts except for the prism itself. Notably, the micro prism is essential for achieving high optical zoom.

The iPhone 15 series has been identified as the first smartphone to incorporate a periscope long focal length lens, which has further advanced the miniaturization and slimming technologies associated with camera modules. This popularity has opened up new market opportunities for the development of micro prisms. Research institutions forecast that the global market for periscope cameras could reach $4.4 billion by 2025. In the first quarter of 2024, global smartphone shipments totaled 300 million units, representing a 12% year-on-year increase, which has buoyed the performance of upstream optical suppliers.

For instance, in the first quarter, Sunny Optical’s optical components, film optical panels, and reflective materials each experienced substantial growth of approximately 85%, 45%, and 50% respectively. The AR segment of automotive electronics remained steady, while semiconductor optics saw a decline of approximately 15%. The profit margins for optical components and film optical panels increased, due in large part to a market revival in smartphones, alongside recent strategic adjustments that enhanced competitive strength through innovation and high-end product development. Products like micro prism modules and composite light-absorbing filters have arisen as novelties within the scope of high-end smartphone optics.

Significant investments have been made in the development of micro prism modules, with companies now targeting wearable devices, laptops, and tablets, diversifying their potential revenue streams. In 2023, Sunny Optical reported high capital expenditures linked to new project developments, culminating in a net fixed asset value increase of approximately 900 million RMB. For 2024, the company aims to focus capital investments primarily on new projects while attempting to limit overall budgets by about 30% compared to 2023.

Furthermore, the market sentiment has markedly improved. In 2023, Sunny Optical achieved revenues of 5.076 billion RMB, a 16% rise compared to the previous year, with optical components accounting for 47% of total revenues—a 21% increase in revenue year-on-year and a 4.5 percentage point rise in profit margins. Meanwhile, Yutong Optical posted income of 2.145 billion RMB, 96.26% of which stemmed from its optical lens manufacturing. This segment grew by 15.24%, though profit margins faced a 5.07 percentage point drop, with security revenue constituting 66.98% of total business. By the first quarter of 2024, revenues for both Sunny Optical and Yutong Optical surged by 53.41% and 38.47% respectively, with comparatively stable overall profit margins.

Companies like Lidin Optics are employing a diversified strategy, focusing on various sectors while remaining competitively aligned with established international optical firms. In 2023, driven by increases in sales for security zoom lenses and drone cameras, production, sales, and inventory levels for Lidin’s zoom lenses improved by 75.59%, 53.79%, and 106.23% respectively. Examining patent activity, Sunny Optical’s invention patents rose from 28 in 2020 to 83 in 2023, while Yutong’s increased from 17 to 79, and Lidin’s grew from 11 to 20.

The growing demands of downstream manufacturers for optical lenses, particularly in zoom capabilities, high resolution, optical stabilization, and reliability, reflect an ongoing competitive landscape in the mobile and security industries. Likewise, consumer pressure has led to price reductions in automotive optical products. Companies must sustain their competitive edge across research, manufacturing processes, scale, client relationships, and profit margins. Sunny Optical and Yutong Optical maintain a concentrated client sales strategy, with Sunny reportedly achieving 57.56% of its total sales from its top three clients in 2023, whereas Yutong sourced 51.93% from its top five clients, with its primary and secondary clients respectively accounting for 24.75% and 16.69% of total sales. Both companies are exploring shifts towards automotive optics and the metaverse for a second growth curve, necessitating timely optimizations in client and business structures to mitigate risks associated with client dependency.

The global VR terminal market saw shipments of 7.65 million units in 2023, with projections indicating that shipments will surpass 8.1 million units in 2024. VR devices have evolved through three phases, starting from aspheric lens solutions, moving to Fresnel optics, and now Pancake-based designs, all aimed at minimizing the weight and size of optical modules. For AR devices to approach the form of modern glasses, glass lens waveguide technology is essential, necessitating the use of high-quality glass wafers as the substrate. To maximize the near-eye display's field of view, glass wafers should exhibit high refractivity. To address dispersion issues, manufacturers often employ multi-layered glass wafers that transmit varying light spectra and maintain a thickness of just 0.3mm, presenting significant manufacturing challenges. It is crucial that domestic manufacturers accelerate their placements in the VR and AR optical lens sectors.

The market for automotive optical products shows great potential. In vehicles, there are various types of cameras, including interior, rear, front, side, and surround cameras. A complete ADAS in new energy vehicles typically requires six cameras, while high-end models can accommodate up to eight, assisting drivers with parking and emergency braking scenarios. The advent of L3 and higher automated driving models has recently expanded, leading to research predictions that by 2026, the market for automotive cameras will reach $35.5 billion, with long-term potential nearing $50 billion.

Initially, the automotive lens market was dominated by multinational corporations with robust global supply chains and stringent evaluations of suppliers based on proficiency and production capabilities; this has established a significant entry barrier. However, as domestic new energy vehicles gain traction, newer manufacturers are now incorporating an average of over ten cameras per vehicle. To meet advanced driving assistance demands, car manufacturers are continuously integrating cameras across all viewpoints to enhance the accuracy and safety of assistance systems.

HUD (Heads-Up Display) technology represents a transformative step, projecting vital information—such as speed, road conditions, and navigation—onto the vehicle’s windshield or dashboard. This allows drivers to access critical information without diverting their gaze from the road, thereby enhancing safety. As the trend towards smart and connected vehicles grows, the integration of HUDs into ADAS becomes a pathway for navigating the global smart automobile wave. Forecasts suggest that the AR HUD market could expand from RMB 3 billion to RMB 30 billion between 2021 and 2025.

The rise of domestic manufacturers has significantly lowered the prices of HUDs, spurring industry development. In 2023, Yutong Optical’s automotive product revenues reached 220 million RMB, marking a staggering surge of 192.92%, encompassing automotive lenses, HUD optical components, and LIDAR optical elements. Additionally, Yuanjong Optical's intelligent driving sector saw revenue growth of 166.82%, aided by multiple mass production products like automotive lenses and AR HUDs. Sunny Optical reported delivering over 200,000 HUD units in 2023, acquiring numerous dedicated projects from various automakers, while also acquiring several new mass production clients for its glass laser radar windows.

By the first quarter of 2024, the AR HUD segment captured 21.1% of the overall HUD market, reflecting a 10.5% year-on-year increase. As the architectural framework and real-world capabilities of AR engines continue to mature, AR HUDs equipped with navigation functions are expected to become standard features in many vehicle models, potentially accelerating their market share growth. Domestic AR HUD products have also achieved full coverage of mainstream technology routes and have been mass-produced. Recent observations indicate that prices for optional AR HUD installations have pushed down from previously high levels of 10,000 RMB to around 5,000 RMB, indicating a movement towards affordability and a promising trajectory for sustained growth.

In terms of visual sensing, 3D visual perception technology was first utilized in industrial applications aimed at high-precision three-dimensional measurements and micromovement detection of objects and materials. Applications have extended to three-dimensional scanning, precise measurement, robotic navigation, and positioning tasks within the industrial realm. By the end of 2023, Orbbec Lumens technology leveraged voice, language, and visual-language models to underpin the development of a robotic arm capable of understanding and executing voice directives.

The pervasive influence of Apple’s flagship smartphones has propelled the mass application of 3D visual sensors into the mobile domain, indicating a significant consumer-facing adoption of 3D visual perception technologies. Beyond facial recognition payments, the scope of 3D visual sensors is rapidly expanding into sectors like smart locks and virtual real estate viewing. Industries leveraging 3D visual perception—including consumer electronics, finance, retail, dining, automotive, and AIoT—are poised for expansive growth, with global market predictions projecting the 3D visual sensing sector to reach $17.2 billion by 2028.

Robotic vision demands heightened capabilities in 3D perception and integration, catering to complex and variable environments. The robust demand for these solutions is prompting swift evolution in mainstream 3D visual perception technologies, further accelerating advancements within the robotics sector.

In the field of autonomous driving, sensors bifurcate into visual and laser radar categories. Visual solutions generally rely on cameras complemented by radar and other sensing components, which ultimately lowers costs but places high demands on algorithmic sophistication. Conversely, laser radar solutions emphasize a strong perceptive capability with a lower algorithmic need through the integration of various sensors, including cameras and radar.

As the penetration rate of L2, L3, and beyond technologies continues to rise, the number of sensors per vehicle is expected to increase significantly. For instance, the anticipated count for L3-configured vehicles could reach between 17 and 34 sensors, substantially surpassing the 6 to 14 sensors typical for L1 vehicles. Based on projections from Dongguan Securities, by 2025, the domestic market for autonomous driving sensors is expected to reach 52.77 billion RMB, with automotive cameras, millimeter-wave radars, and ultrasonic radars predicted to achieve substantial market sizes of approximately 13.29 billion RMB, 14.22 billion RMB, and 17.24 billion RMB, respectively.

In 2023, Yongxin Optical achieved sales revenues of over 90 million RMB in its automotive and laser radar divisions, reflecting an increase exceeding 50%, while transitioning laser radar from small-scale shipments to full scale production. In Q1 2024, the company secured alignment contracts for Hesai's standard laser radar components and for the 905nm laser radar elements provided by Tudatong.

For 2023, the global machine vision market was valued at approximately 92.52 billion RMB, presenting a year-on-year growth of approximately 5.80%, and anticipated to surpass 110 billion RMB by 2025. The majority market share of the machine vision industry remains dominated by foreign enterprises, with players like Keyence and Cognex together commanding over 50%. The second half of 2023 marked a noticeable recovery in consumer electronics, presenting a prime opportunity for upstream suppliers in the machine vision value chain. With the advancing shift towards high-end domestic consumer electronics and the gradual establishment of an accompanying industry infrastructure, more opportunities for the adoption of machine vision products will emerge. The machine vision market in China alone reached 18.51 billion RMB in 2023, reflecting a year-on-year increase of 8.49%. As 3D visual technology and deep learning continue to evolve, the industrial machine vision sector is set to permeate increasingly intricate industrial inspection arenas.

The convergence of 3D visual perception technology with VR and AR applications is set to facilitate more authentic and immersive interaction experiences, opening new development avenues within sectors such as education, training, and entertainment. The proven application of new technologies in expanding the boundaries of industrial machine vision underscores their significance in this rapidly evolving landscape.

Looking ahead, the performance of AI models is expected to become increasingly diverse. Every form of interaction—whether through text, sound, or image—will further optimize and proliferate. The implementation and fine-tuning of large-scale AI and industrial AI technologies in the machine vision domain promise to broaden the industry's horizons.