Company code: 688088 Company abbreviation: Arcsoft Technology

ArcSoft Technology Co., Ltd

Summary of the 2024 Annual Report

Section 1 Important Notice

1. The summary of this annual report is from the full text of the annual report, in order to fully understand the company's operating results, financial status and future development plan, investors should go to the www.sse.com.cn website to read the full text of the annual report carefully.

2. Major risk warning

The Company has set out the risk factors that the Company may face in the course of its operations in the 2024 Annual Report, please refer to the "Section 3 Management Discussion and Analysis/IV. Risk Factors" section of the 2024 Annual Report. Investors are advised to pay attention to the investment risks.

3. The Board of Directors, the Board of Supervisors, the directors, supervisors and senior management of the Company guarantee the authenticity, accuracy and completeness of the contents of the annual report, and that there are no false records, misleading statements or material omissions, and bear individual and joint legal liabilities.

4. All directors of the company attend the meeting of the board of directors.

5. Lixin Certified Public Accountants (Special General Partnership) issued a standard unqualified audit report for the company.

6. The company is not profitable and has not yet achieved profitability at the time of listing

□ Yes √ No

7. The profit distribution plan or the plan for the conversion of provident fund into share capital for the reporting period approved by the board of directors

After deliberation and approval at the third meeting of the third board of directors of the company, the company intends to register on the record date of the implementation of equity distribution in 2024

The total share capital is deducted from the shares in the company's special securities account for repurchase as the basis for distributing profits. The profit distribution plan is as follows:

The company intends to distribute a cash dividend of 2.50 yuan (tax included) to all shareholders for every 10 shares, and will not convert capital reserve into share capital, and will not give bonuses

Share. As of April 11, 2025, the company's total share capital is 401,170,400 shares, deducting 439,000 shares in the repurchase special securities account

shares, based on this calculation, the total proposed cash dividend is 100,182,850.00 yuan (tax included).

In 2024, the company's cash dividends (including the interim distributed cash dividend of 60,109,710.00 yuan) totaled 160,292,560.00

yuan, accounting for 90.72% of the net profit attributable to shareholders of listed companies in 2024; In 2024, the amount of share repurchase that has been implemented in cash and by means of centralized bidding and offer is RMB 13,199,788.71 (including transaction costs such as stamp duty and transaction commission), cash

The total amount of dividends and repurchases was 173,492,348.71 yuan, accounting for 98.19% of the net profit attributable to shareholders of listed companies in 2024.

The company holds 439,000 shares of the company through the repurchase special account and does not participate in this profit distribution.

If there is a change in the total share capital of the company due to the conversion of convertible bonds into shares/repurchase of shares/repurchase of shares granted by equity incentives/repurchase and cancellation of shares of major asset restructuring during the period from the date of disclosure of this report to the registration date of the implementation of equity distribution, the company intends to maintain the distribution ratio per share unchanged and adjust the total distribution amount accordingly. If there is a subsequent change in the total share capital, the specific adjustment will be announced separately.

The profit distribution plan still needs to be submitted to the company's 2024 annual general meeting of shareholders for consideration.

8. Whether there are special arrangements for corporate governance and other important matters

□ Applicable √ Not applicable

Section 2 Basic information of the company

1. Company profile

1.1 Company Shares

√ Applicable □ Not applicable

Company stock profile

Stock Type Stock Listed Exchange and Sector Stock Abbreviation Stock Code Stock Abbreviation Before the Change

A-shares, Shanghai Stock Exchange's Science and Technology Innovation Board, Arcsoft Technology, 688088 Not applicable

1.2 Brief description of the Company's Depositary Receipts

□ Applicable √ Not applicable

1.3 Contacts and Contact Details

Secretary of the Board of Directors and Representative of Securities Affairs

Name: Hao Huimei, Liao Juanjuan

Address: 277 Longlan Road, Xuhui District, Shanghai Donghangbin, No. 277 Longlan Road, Xuhui District, Shanghai

15th Floor, T2, Jiang Center, 15th Floor, T2, Jiang Center

Tel: 021-52980418 021-52980418

Fax: 021-52980248 021-52980248

E-mail invest@arcsoft.com invest@arcsoft.com

2. Brief introduction of the company's main business during the reporting period

2.1 Main business, main products or services

The company focuses on the field of computer vision, provides algorithm authorization and system solutions for the industry, and is the world's leading visual artificial intelligence company, which has always been committed to the research and development and application of visual artificial intelligence technology, adheres to technological innovation as the core driving force, and provides one-stop visual artificial intelligence solutions for intelligent devices on a global scale.

The company has a rich product line of visual algorithms for mobile intelligent terminals such as smart phones and smart cars, and its main business income comes from the licensing of independent research and development of core technologies. At present, the main customers include Samsung, Xiaomi, OPPO, vivo, Honor, Moto and other world-renowned mobile phone manufacturers, as well as major domestic independent brands, some joint ventures and foreign brand automobile OEMs.

For smartphones, based on years of research and development, the company can provide most of the mainstream smartphone vision artificial intelligence algorithm products on the market, including single/dual/multi-camera camera shooting and high-quality imaging in various scenes, depth camera shooting and high-quality imaging in various scenes, periscope telephoto camera stepless zoom, 3D modeling, panorama, SLAM, AR/VR, face unlocking, superpixel lossless zoom, human drive and other important core functions. On laptops, the company improves the user's video conferencing experience from several aspects such as picture quality, privacy, and beauty. Improve the image quality of meetings by removing video noise and increasing dynamic range; User privacy is protected by blurring and changing backgrounds; Improve the effect of video conferencing through technologies such as portrait beautification, automatic centering of the human body, face detail restoration and enhancement, and glasses to remove reflection.

In the field of AI-XR, the company quickly responded to the market demand for AI glasses, and actively deployed AI glasses image algorithms in the original VR/MR/AR

On the basis of the algorithm solution, the rapid expansion of AI glasses image algorithm can not only provide VR/MR/AR one-stop multi-sensor calibration solution, but also headset 6DoF tracking, plane detection, dense reconstruction, semantic analysis, depth estimation, real-time human/hand segmentation,

Spatial awareness solutions such as real-time 3D mesh reconstruction, 6DoF tracking of the controller, 3D gesture interaction with bare hands, gaze tracking, and virtual digital

Interactive solutions such as human expression-driven and human-driven, as well as visual presentation solutions such as asynchronous time warping, asynchronous space warping, and video perspective (VST), can also provide solutions such as AI glasses single camera shooting and high-quality imaging in various scenes, intelligent face recognition, and intelligent gesture recognition.

In the field of intelligent vehicles, for the outside of the cabin, the company reserves visual perception algorithms based on front-view, circumferential-view, surround-view, rear-view, and night vision camera systems in the driving assistance function, which can provide advanced intelligent driving assistance functions such as ACC, LCC, AEB, ILC, etc.; On the automatic parking function (APA), the visual perception algorithm, the ultrasonic vision fusion algorithm and the regulation algorithm are reserved. It provides a 360° surround view subsystem, supports 2D/3D AVM panoramic image function, and supports transparent chassis and sentry monitoring functions. For the cabin, the company has accumulated a series of visual perception algorithms in the form of DMS and OMS, which can provide cabin monitoring and interactive system functions such as fatigue detection, distraction detection, health monitoring, identity recognition (Face ID), gesture recognition, and residual reminders.

In the field of intelligent commercial photography, the company launched the commercial photography image and video generation based on the ArcMuse computing technology engine

PhotoStudio？ The AI intelligent commercial auction cloud studio (PSAI) is divided into clothing version and merchandise version, providing services for different customer groups.

PSAI supports AI model graphics, AI model scene graphics, AI product graphics, intelligent fitting image generation, and AI model video generation

Yes. Through the organic combination of controllability and expressiveness of image and video content generation, it provides merchants with the perfect display of clothing and other products. At the same time, PSAI also provides AI toolkits such as intelligent cutout and intelligent lighting, and provides intelligent post-processing tools for image and video generation content to meet the differentiated needs of users.

2.2 Main business model

1. Profit model: The company's main profit model is to deeply integrate computer vision algorithm technology with customer-specific equipment, and authorize it to customers through contracts, allowing customers to use relevant algorithm software or software packaging on smart devices of the agreed model, so as to charge technology and software license fees. At the same time, the company also sells software and hardware integrated vision solutions to customers.

2. R&D model: The company mainly adopts the mode of independent research and development. The R&D process is broadly divided into the following 9 steps: (1) obtaining requirements information; (2) Management decision-making on R&D direction; (3) Build a R&D project team; (4) Verify the algorithm of the R&D project and conduct project testing; (5) Collective discussion and decision-making project algorithm; (6) Combine and optimize the underlying algorithm with the actual environment; (7) Conduct actual product combination tests; (8) Roadshow after the product matures, select the right customer for testing and cooperation; (9) Large-scale promotion after passing the test.

3. Sales model: The company adopts direct sales, mainly selling computer vision algorithm software and related solutions to manufacturers of smart phones, smart cars, laptops, smart homes, smart retail, and various AIoT equipment with cameras.

4. Charging mode: According to the different types of business contracts, the main charging mode of the company's computer vision algorithm software can be divided into two modes: fixed fee mode and piecework mode. (1) Fixed fee model: A fixed amount of software license fee will be charged according to the software license period agreed in the contract. During the software license period, a specific customer can legally load an unlimited amount of specific software containing ArcSoft algorithm technology on a smart device limited by the contract for a certain type or series of specific devices. (2) Piecework mode: During the software license period agreed in the contract, the customer will be charged according to the number of intelligent devices equipped with Arcsoft algorithm technology. Normally, the company will agree on a tiered price with the customer for different production quantity ranges to protect the interests of both parties. For the integrated software and hardware solutions, the company currently adopts the piecework model.

5. Procurement mode: The company's main procurement content includes all kinds of hardware equipment, software, services required for R&D, testing and operation, as well as materials required for product solutions. According to the requisition application of the demand department, the procurement department implements the process of supplier selection, procurement contract signing, contract execution tracking, and purchase payment application in accordance with the requirements of the "Procurement Management System". For the integrated software and hardware solution, the company designs the hardware and purchases the corresponding core components, and then entrusts a third party to assemble and produce.

2.3 The situation of the industry

(1). The development stage, basic characteristics and main technical thresholds of the industry

According to the "Industrial Classification of the National Economy" (GB/T4754-2017), the company belongs to the industry of "I65 software and information technology services"

I6513 Application Software Development. According to the Guidelines for the Statistical Classification of Listed Companies of the China Association of Listed Companies issued by the China Association of Listed Companies, the industry to which the company belongs is "Information Transmission, Software and Information Technology Service - Software and Information Technology Service Industry", and the industry code is "I65".

The company is engaged in the research and development and application of computer vision technology algorithms, and its core products include mobile intelligent terminal vision solutions, intelligent driving vision solutions, intelligent commercial photography solutions and other AIoT intelligent device vision solutions.

According to the 2016 edition of the Guiding Catalogue of Key Products and Services for Strategic Emerging Industries published by the National Development and Reform Commission, the company belongs to the "new generation

Information Technology Industry".

(1) The development stage and basic characteristics of the industry

(1) The development stage, basic characteristics and industrial policy of the industry

In the context of the new scientific and technological revolution and industrial transformation, artificial intelligence has accelerated its penetration into various industries, and has increasingly become an important driving force for scientific and technological innovation, industrial upgrading and productivity improvement. The government has actively introduced policies to promote the development and application of artificial intelligence technology, deepened the implementation of relevant policies on artificial intelligence, intelligent manufacturing, informatization and industrialization, which are closely related to visual artificial intelligence, and provided policy and supporting resource support for the development of visual artificial intelligence. Since 2024, the main new AI industry policies have been introduced as follows:

In January 2024, the executive meeting of the State Council will study and deploy the work related to promoting artificial intelligence to empower new industrialization. It was emphasized that

It is necessary to coordinate high-quality development and high-level security, take the deep integration of artificial intelligence and manufacturing as the main line, take intelligent manufacturing as the main direction, and take scene application as the traction, accelerate the intelligent upgrading of key industries, vigorously develop intelligent products, empower the industrial manufacturing system at a high level, accelerate the formation of new quality productivity, and provide strong support for the construction of a manufacturing power, a network power and a digital China.

In March 2024, in the "2024 Government Work Report of the State Council", it was proposed to deepen the research and development of big data and artificial intelligence

Carry out the "artificial intelligence +" action to build a digital industrial cluster with international competitiveness.

In April 2024, the General Office of the National Development and Reform Commission and the General Department of the National Data Bureau issued the "Key Points of the Digital Economy in 2024".

It is proposed to appropriately advance the layout of digital infrastructure, further promote the digital transformation of industries, accelerate the promotion of digital technology innovation and breakthroughs, continuously improve the level of public services, promote the improvement of the digital economy governance system, comprehensively build a solid digital security barrier, actively expand international cooperation in the digital economy, and strengthen cross-departmental coordination and linkage.

In July 2024, the Ministry of Industry and Information Technology and other four departments jointly issued the "Guidelines for the Construction of a Comprehensive Standardization System for the National Artificial Intelligence Industry".

South (2024 Edition)" proposes to seize the development opportunities of the artificial intelligence industry, improve the top-level design of artificial intelligence standards, strengthen the coordination of standard work of the whole industry chain, and coordinate the promotion of the research, formulation, implementation and internationalization of standards, so as to provide solid technical support for promoting the high-quality development of China's artificial intelligence industry.

In December 2024, the Ministry of Industry and Information Technology and other three departments jointly issued the "Implementation Guide for Digital Transformation of Manufacturing Enterprises", proposing:

Encourage enterprises to explore new applications of intelligent R&D, develop "artificial intelligence +" R&D and design software, build design models, simulation models and other data sets, carry out model training, develop innovative applications such as generative design and real-time simulation, and accelerate the research and development of new products.

Entering 2025, policies will continue to focus on the AI industry. In March, in the "2025 Government Work Report", it was proposed to continue

Promote the "artificial intelligence +" action, support the wide application of large models, and vigorously develop a new generation of intelligent terminals and intelligent manufacturing equipment such as intelligent networked new energy vehicles, artificial intelligence mobile phones and computers, and intelligent robots.

(2) The development status of subdivisions

The visual artificial intelligence industry provides basic support technologies for various artificial intelligence applications, which are widely used in various artificial intelligence subdivisions. Since 2024, the development of new AI technologies, including large models, will continue to empower a new generation of intelligent terminals such as smart cars, AI mobile phones and computers, and smart robots. The development status of the main segments related to the Company is as follows:

Mobile intelligent terminal segmentation

In 2024, despite global macroeconomic challenges, global smartphone shipments will usher in positive changes, benefiting from the recovery of some emerging markets, the promotion of AI technology, and the aggressive marketing strategies of manufacturers. According to Counterpoint statistics, global smartphone shipments in 2024 will achieve a positive growth of 4% year-on-year. AI technology is becoming a core selling point, and many Android phone manufacturers such as Samsung, Honor, Xiaomi, OPPO, vivo, and Transsion have successively launched mobile phone products with AI functions. Image has always been the main track of smartphone competition, and the application of AI technology has further improved the imaging capabilities of mobile phones, and the image algorithm with large model as the computing base is positively driven

Moving image technology innovation. Counterpoint predicts that the global penetration rate of AI phones will be about 40% in 2027, and shipments are expected to reach 522 million

Division. The industry believes that factors such as the innovation of interaction methods, functional upgrades, and the emergence of new forms such as folding screens complement each other, which is expected to drive the demand for replacement to a certain extent.

In the convergence of AI and XR technologies, smart terminals such as AI glasses are opening up a new mode of human-computer interaction. AI intelligent assistant system integration view

Multimodal data such as sensory recognition, voice interaction, and environmental perception build a technical closed-loop of "holographic perception-real-time analysis-intelligent response" to reshape the user's experience in work and life scenarios. The AI eyewear industry is in a critical period of vigorous development, and technology companies are pouring in

According to reports, more than 40 domestic and foreign manufacturers have entered the AI glasses, including Internet manufacturers, mobile phone giants, and AR Ming

Star enterprises, the number of products involved is expected to exceed 50. IDC estimates that the global smart glasses market is expected to ship 12.05 million units in 2025

units, a year-on-year increase of 18.3%, of which audio glasses and audio shooting glasses without display functions are expected to ship 5.47 million units, a year-on-year increase of 101.9%.

Intelligent car segment

According to data released by GlobalData, global vehicle sales in 2024 will increase by 2.1% year-on-year to 89 million units. from the domestic market

Although facing challenges such as changes in the external environment, intensified competition in the industry, and poor endogenous power of automobile consumption, due to the continuous effectiveness of a series of policies, the effective implementation of subsidy policies in various places, and the simultaneous efforts to stimulate the vitality of terminal consumption in the automobile market. According to the medium

4.5%。 Among them, the annual export volume was 5.859 million units, a year-on-year increase of 19.3%, reaching a new level.

China's intelligent automobile industry has ushered in a golden period of development, and high computing power platforms and intelligent configuration have become the core elements of differentiated competition among car companies. The National Development and Reform Commission predicts that by 2025, the number of smart cars in China is expected to reach 28 million

Farm penetration is expected to be 82%; By 2030, the number of smart cars will increase to 38 million, and the penetration rate is expected to reach 95%.

In the field of intelligent driving, from traditional automakers, traditional Tier 1 and emerging car companies, to technology companies, Internet companies and consumer electronics companies, they are actively promoting the commercialization of intelligent driving technology. Urban NOA, high-definition maps, and cabin-driver integration have become the main development trends in the field of intelligent driving, and the progress of these technologies will promote intelligent driving to a higher level of automation and intelligence.

In the field of intelligent cockpit, according to information released by China Automotive Data, the penetration rate of intelligent cockpit-related functions in China's passenger car market exceeds 70%, and China has become the largest market for intelligent cockpit. With the breakthrough of key technologies such as high-performance computing platforms, multi-modal interactions, and operating systems, intelligent cockpits are accelerating the development of software and hardware integration. It is estimated that by 2025, the domestic intelligent cockpit market space will reach the level of 100 billion yuan.

Commercial shooting segments

In the field of e-commerce, product display pictures and videos are the core touchpoints of the online shopping decision-making chain, which can stimulate consumers' desire to buy, and are the digital window to build product competitiveness and brand trust. With the rapid development of a new generation of artificial intelligence with large visual models as the core, intelligent commercial photography is expected to replace traditional commercial photography. Apparel stores have a stronger demand for commercial auctions, followed by cosmetics, daily necessities and other categories. According to the research content of Northeast Securities,

From the perspective of GMV, the annual cost of an online retail platform for commercial auctions is about 1%-2% of its GMV, which can be calculated as an e-commerce zero

The market space for commercial auctions is about 25.6-51.1 billion yuan. From the perspective of the number of stores, the average annual expenditure of stores on commercial auctions is more than 20,000 yuan, and assuming that the penetration rate of commercial auctions is 50%, the market space for commercial auctions is about 34 billion yuan per year.

With the maturity of visual AI large model technology, some AI commercial shooting solutions have emerged on the market, such as PhotoStudio?

AI, Weshop, PhotoMagic, etc., these solutions have brought significant cost reduction and efficiency improvement changes to the traditional commercial shooting field with the help of visual large-scale model application products. As a technology-intensive application in a vertical industry, the intelligent commercial photography technology stack is different from the application of general visual large models, which needs to take into account the creativity of large models and at the same time generate images or videos around the merchant's product content. The verticalization of scenarios, the demand for controllable generation and service refinement, differentiated customer needs and huge potential market capacity will provide broad market opportunities for track players with technology accumulation and innovation capabilities.

(2) Main technical thresholds

The company mainly provides visual artificial intelligence solutions for mobile intelligent terminals, intelligent cars and other intelligent devices and commercial shooting, and the main technical thresholds in the aforementioned fields include:

(1) Accumulation of end-to-end computing and edge computing technologies

Edge computing greatly solves the two major bottlenecks of bandwidth and latency in the centralized computing architecture in the context of the Internet of Things, and the main difficulty lies in the development ability of the low-resource embedded platform environment.

Since 2003, the company has clarified the development direction of visual artificial intelligence technology related to the research and development of embedded devices, and has accumulated deep experience in the field of edge computing technology. The company's accumulated algorithms are highly compact, stable, and easy to recall, and can achieve high-precision operation with high performance and effectively and significantly reduce resource consumption.

The company has been focusing on the research and development of embedded device algorithms for a long time, and has accumulated a lot of experience in the development of visual artificial intelligence algorithms based on terminal devices over the years. At present, the company's visual artificial intelligence algorithm based on end devices has high applicability and stable operation, which can initiate efficient computing on the edge side, and realize efficient image optimization, recognition and detection functions through devices such as smart phones, laptops, and smart wearables. The company's mobile intelligent terminal vision solutions, intelligent driving vision solutions, and other AIoT smart device vision solutions are all initiated from the edge side, and various visual artificial intelligence functions are realized in smartphones, laptops, smart wearable devices, automobiles, and various AIoT devices.

(2) Layered accumulation of visual artificial intelligence technology

In the era of digital cameras and mobile phone feature phones, the company began to focus on the research and development and application of visual artificial intelligence technology, the company's main algorithm technology has been tempered for a long time, from the basic black and white small resolution image acquisition, enhancement, editing, detection and recognition to high-definition large images, real-time processing of video have laid a solid foundation, creating favorable and leading conditions.

The visual artificial intelligence algorithm technology mastered by the company is versatile and scalable. Over the years, the company has been actively working to bring visual people

capability, the company quickly implemented the visual artificial intelligence algorithm technology into a mature solution, and further expanded the application field from smart phones, smart wearables, smart homes, smart retail and other industries to promote industry upgrading. In addition, based on its own profound technology accumulation, the company can quickly provide solutions with high performance, high efficiency, strong hardware platform adaptability and excellent power consumption control for the above-mentioned industries, greatly reducing the threshold for various customers to use advanced technology in their products, helping customers improve product competitiveness, and helping the popularization of visual artificial intelligence and artificial intelligence-related applications.

(3) Project landing capacity

Since its establishment, ArcSoft has been committed to long-term and extensive cooperation with mainstream companies in the core industry chain, in addition to continuously accumulating and developing its own technology, mastering the ability to continuously develop and iterate algorithms that better match hardware to hardware. The company has established long-term and stable cooperative relations with Qualcomm, MediaTek and other mainstream chip companies, and continues to cooperate and exchange in research and development, deeply understand the hardware characteristics of the platform and optimize it for its targeting, jointly develop core functions, and continuously improve the adaptability of visual artificial intelligence technology algorithm products and mobile chips. The company has also established business exchanges or cooperative relations with core device partners in the industry such as Sony Sensor, Samsung Semiconductor, Geke Micro, OmniVision, Sunny Optics, Xinli, etc., and has done algorithm adaptation and tuning for specific cameras or hardware in the early stage of the project. For the chip platform of smart terminals, the company has strong optimization capabilities for various computing power units such as CPU, GPU, DSP and NPU. Combined with the capabilities of each hardware computing unit and the computing power requirements of algorithm modules, the company's heterogeneous computing optimization capabilities can more effectively optimize performance and reduce power consumption at the system level. Thanks to this, in addition to the outstanding core technical capabilities, the company also has obvious advantages in the ability to implement the project, after the customer puts forward the technical needs, it can better combine and give full play to the advantages of multi-party resource cooperation in camera modules, software and hardware platforms, production lines, algorithms, etc., so as to provide solutions with good effect, low energy consumption, high efficiency, wide adaptability of hardware platform and rapid landing.

(2). Analysis of the company's position in the industry and its changes

At present, the visual artificial intelligence market has initially formed a competitive pattern of "head agglomeration + vertical deep cultivation", and factors such as accelerated technology iteration, deepening industry penetration, and policy guidance are still driving market dynamic changes. The accumulation of core technologies, productization capabilities, and industrial ecological chain cooperation all constitute the core barriers of various vertical industries. The company is a leading algorithm service provider and solution provider in the computer vision industry, and a leading computer vision artificial intelligence company in the world.

The company adheres to the development strategy of deepening the layout of computer vision algorithm technology, deepening the layout of various industries, and gradually expanding the application of computer vision algorithm technology to more intelligent terminal equipment fields. In addition to the company, the main domestic enterprises in the industry are SenseTime and Megvii Technology, and the foreign companies are mainly Seeing Machines, Mobileye, Cipia and Smarteye.

In the field of smart phones, the company provides customers with full-stack visual artificial intelligence solutions, and is one of the world's largest suppliers of smart phone visual artificial intelligence algorithms. The company's main customers include Samsung, Xiaomi, OPPO, vivo, Honor, Moto and other world-renowned mobile phones

Manufacturers, according to IDC's statistics, among the top five mobile phone brands in terms of global shipments from 2019 to 2024, except for Apple, which is completely adopted

In addition to the self-developed visual artificial intelligence algorithm, the mainstream models of other Android mobile phones are equipped with the company's visual artificial intelligence solutions. At the same time, the company has in-depth cooperation with mobile chip manufacturers such as Qualcomm and MediaTek to optimize algorithms for different hardware platforms, improve the adaptability with mobile chips, and establish technical exchanges with sensor manufacturers such as OmniVision, Sony, and Samsung Semiconductor, forming a complete ecological chain from the underlying hardware to terminal applications.

In the field of AR/VR/XR, the company has carried out forward-looking layout, built a complete set of Arcsoft space computing technology system from the four dimensions of calibration, perception, interaction and visual presentation, and created a one-stop and all-round product solution for customers. While vigorously updating technology, the company also maintains close communication and exchanges with a number of leading manufacturers in the industry, and continues to expand the technical cooperation ecosystem. In the process of deep integration of AI-XR technology, intelligent terminal devices represented by AI glasses are opening a new paradigm of human-computer interaction. The company is fast

In response to the market demand for AI glasses, we have actively deployed AI glasses imaging algorithms and established in-depth cooperation with many well-known AI glasses brands

relationship, according to the product positioning and market demand of the brand, customized development of image algorithm solutions. At present, it has successfully helped Thunderbird V3 become the first AI glasses to be sold in China.

In the field of smart cars, the company provides customers with VisDrive? One-stop vehicle vision software solution, is a mainstream visual algorithm supplier based on Qualcomm intelligent cockpit platform in China. The company's customers cover major domestic independent brands, some joint venture brands and foreign brand automobile OEMs, and have enabled many customers to successfully achieve mass production of products in relevant countries and regions, greatly accelerating the marketization process of new vehicles.

Since its establishment, the company has experienced a multi-stage evolution from shallow pattern recognition to deep learning to intelligent computing driven by large models. At present, artificial intelligence technology is undergoing a paradigm change with large models as the core, and under the triple drive of policy guidance, industrial upgrading and technological leap, large models of vertical industries are reconstructing the ecology of various industries. Relying on the deep accumulation of visual AI technology and the self-developed ArcMuse computing technology engine as the core, the company continues to promote the technological progress and industry implementation of visual large models. The company will launch PSAI in 2023, and has completed Taobao Qianniu, 1688, Douyin Doudian,

Platforms such as TikTok, Jingjingmai, Pinduoduo and Shein have settled in, becoming the first service provider in China to complete the full coverage of mainstream e-commerce platforms, serving hundreds of thousands of merchant customers. As the core base of the company's vision AI, the next-generation ArcMuse computing technology engine will also serve the company's smartphones, smart cars, AI glasses, XR, robotics and other diverse scenarios.

(3) The development and future development trend of new technologies, new industries, new forms of business and new models during the reporting period

In terms of core technology, artificial intelligence technology is undergoing a paradigm change with large models as the core. Powered by DeepSeek V3 and R1

The open-source large language model of the table has set off a wave of AI equality, allowing the large model track to return from heaping computing power infrastructure to engineering innovation and algorithm innovation itself. The success of the DeepSeek model on large language models has greatly promoted the dissemination and acceptance of large model application products in China. At the same time, it has become a new trend in the industry to replicate and migrate DeepSeek engineering and algorithm innovation methods in the field of visual large models.

In terms of industrial applications, vertical industry visual models are reconstructing the ecology of various industries. In the field of consumer electronics, image algorithms based on large models are driving the innovation of imaging technology. Based on multi-modal data fusion and dynamic adaptive modeling technology, a cross-platform ecology of algorithm-hardware-application collaboration is built to form high-precision semantic understanding and real-time rendering capabilities, accelerate the reconstruction of underlying computing power and the standardization process of computational photography, and meet consumers' core needs for professional-level image experience and creative freedom integration. In the field of intelligent vehicles, large models will drive the intelligent vehicle industry to a new stage of technological innovation and user experience upgrading in terms of realizing scene perception and personalized services, improving environmental perception capabilities, optimizing decision-making and planning, and improving the adaptability and decision-making ability of autonomous driving systems in complex environments. In the field of commercial shooting, traditional commercial shooting has problems such as long cycle, high cost, complicated process and serious homogeneity, and the emergence of vertical industry visual models will combine professional technology with scene depth, effectively solve the pain points of commercial shooting, balance standardization and personalized needs, and accelerate the trend of reducing costs and increasing efficiency in the commercial shooting industry.

3. The company's main accounting data and financial indicators

3.1 Key accounting data and financial indicators for the past 3 years

Unit: Yuan Currency: RMB

2024 2023 This year is compared to 2022

Increase/Decrease (%) After Adjustment Before Adjustment

Total assets 3,021,180,411.41 3,120,573,359.75 -3.19 2,989,603,256.60 2,982,674,093.89

Net assets attributable to shareholders of listed companies 2,717,799,736.52 2,672,606,161.32 1.69 2,579,022,814.44 2,579,177,049.30

Operating income 815,173,516.81 670,254,360.04 21.62 531,648,526.82 531,648,526.82

Net profit attributable to shareholders of listed companies 176,685,765.99 88,487,716.83 99.67 57,822,806.53 57,797,302.49

Deductions attributable to shareholders of listed companies are not 155,831,765.47 68,069,398.65 128.93 6,814,662.44 6,789,158.40

Net profit from recurring gains and losses

Net cash flow from operating activities 10,887,649.04 255,528,317.66 -95.74 44,630,176.92 44,630,176.92

Weighted average return on equity (%) 6.56 3.38 increased by 3.18 percentage points 2.26 2.26

Basic earnings per share (RMB/share) 0.44 0.22 100.00 0.14 0.14

Diluted earnings per share (RMB/share) 0.44 0.22 100.00 0.14 0.14

Ratio of R&D investment to operating income (%) 48.81 54.11 decreased by 5.30 percentage points 54.15 54.15

3.2 Key accounting data by quarter in the reporting period

Unit: Yuan Currency: RMB

Q1 Q2 Q3 Q4

(January-March) (April-June) (July-September) (October-December)

Operating income 183,929,058.24 196,987,586.65 192,764,996.57 241,491,875.35

Net profit attributable to shareholders of listed companies 34,165,336.48 27,293,589.35 26,831,794.62 88,395,045.54

Deduction of non-recurring losses attributable to shareholders of listed companies 30,284,626.80 23,509,652.96 24,111,387.64 77,926,098.07

Net profit after profit

Net cash flow from operating activities -16,554,162.85 -48,292,167.53 66,649,559.89 9,084,419.53

Explanation of the discrepancies between the quarterly data and the data disclosed in the periodic report

□ Applicable √ Not applicable

4. Shareholders

4.1 The total number of shareholders of common shares, the total number of shareholders of preferred shares whose voting rights have been restored, the total number of shareholders holding special voting shares, and the top 10 shareholders

Unit: Shares

The total number of ordinary shareholders (households) at the end of the reporting period was 18,676

The total number of common shareholders (households) at the end of the previous month prior to the disclosure date of the annual report was 21,971

The total number of preferred shareholders whose voting rights were restored as of the end of the reporting period (households) was 0

The total number of preferred shareholders whose voting rights were restored at the end of the previous month before the disclosure date of the annual report (households) is 0

The total number of shareholders holding special voting shares as of the end of the reporting period (households) 0

The total number of shareholders holding special voting shares at the end of the previous month before the disclosure date of the annual report (households) 0

Shareholdings of the top 10 shareholders (excluding shares lent through refinancing)

Name of Shareholder (Full Name) Increase or Decrease in Reporting Period Number of Shares Held at the End of the Period Proportion (%) Holding Limited Sales Pledges, Marks or Freezes Nature of Shareholders

Number of shares Status of shares Quantity

Hongrun Capital Management Co., Ltd. 0 118,698,800 29.59 0 None 0 Overseas legal person

Nanjing UBP Emerging Industry Investment Fund Partnership (Limited -3,567,732 46,070,332 11.48 0 None 0 Other

Gang)

Rainbow Global Limited 0 19,595,020 4.88 0 None 0 Overseas legal person

Hangzhou Hongli Investment Management Partnership (Limited Partnership) 0 14,902,950 3.71 0 None 0 Other

Hongyu Co., Ltd. -186,170 10,804,800 2.69 0 None 0 Overseas legal person

Hangzhou Hongli Investment Management Partnership (Limited Partnership) 0 6,114,213 1.52 0 None 0 Others

Ruizhong Life Insurance Co., Ltd.-Own funds 4,108,693 4,108,693 1.02 0 None 0 Other

Lin Shiyi 45,000 3,531,124 0.88 0 No 0 Nature

person

Hangzhou Hongren Investment Management Partnership (Limited Partnership) 0 3,135,231 0.78 0 None 0 Other

Zhonghui Life Insurance Co., Ltd. - Traditional Products 3,000,000 3,000,000 0.75 0 N/A 0 Others

Hongrun Capital Management Co., Ltd., controlled by Mr. Hui Deng, and Mr. Hui Deng

Description of the above-mentioned shareholder affiliation or concerted action Rainbow Global Co., Ltd., controlled by Ms. Liuhong Yang, is a person acting in concert. In addition to this, the company

It is not known whether the other shareholders are related or acting in concert.

Explanation of preferred shareholders whose voting rights have been restored and the number of shares held None

Holders of depositary receipts

□ Applicable √ Not applicable

The top 10 shareholders in terms of the number of voting rights as of the end of the reporting period

□ Applicable √ Not applicable

4.2 Block diagram of the property rights and control relationship between the company and the controlling shareholder

√ Applicable □ Not applicable

4.3 Block diagram of the property rights and control relationship between the company and the actual controller

√ Applicable □ Not applicable

4.4 The total number of preferred shareholders and the top 10 shareholders of the company at the end of the reporting period

□ Applicable √ Not applicable

5. Corporate bonds

□ Applicable √ Not applicable

Section 3: Important Matters

1. The company shall, in accordance with the principle of materiality, disclose the major changes in the company's operation during the reporting period, as well as the events that have a significant impact on the company's operation during the reporting period and are expected to have a significant impact in the future.

During the reporting period, the company achieved operating income of 815.1735 million yuan, an increase of 21.62% over the same period of last year; Realization attributable to listed companies

The net profit of shareholders was 176.6858 million yuan, an increase of 99.67% over the same period of last year.

2. If there is a delisting risk warning or termination of listing after the disclosure of the company's annual report, the reasons for the delisting risk warning or termination of listing shall be disclosed.

□ Applicable √ Not applicable