Zhuhai Allwinner Technology Co., Ltd. 2025 Annual Report Summary

Log in to the Sina Finance APP and search for [Information Disclosure] to view more evaluation levels.

Stock Code: 300458 Stock Abbreviation: Allwinner Technology Announcement No: 2026-0327-003

I. Important Notice

This annual report summary is derived from the full annual report. To fully understand the company’s operating results, financial status, and future development plans, investors should carefully read the full annual report in the media designated by the China Securities Regulatory Commission.

All directors have attended the board meeting to discuss this report.

Tianjian Certified Public Accountants (Special General Partnership) provided the audit opinion for this year’s financial report as: a standard unqualified opinion.

Non-standard audit opinion prompt

□ Applicable √ Not applicable

The company was not profitable at the time of listing and has not achieved profitability currently.

□ Applicable √ Not applicable

The profit distribution plan or reserve fund capitalization plan reviewed by the board during the reporting period.

√ Applicable □ Not applicable

The profit distribution plan approved by the board at this meeting is: based on 825,427,382 shares, a cash dividend of 2 yuan (tax included) will be distributed for every 10 shares to all shareholders, with 0 shares of bonus shares (tax included), and 2 shares will be distributed for every 10 shares from the capital reserve to all shareholders.

Board resolution on the preferred stock profit distribution plan for this reporting period.

□ Applicable √ Not applicable

II. Basic Information of the Company

1. Company Profile

■

2. Main Business or Product Overview during the Reporting Period

(1) Main Business

The company’s main business currently involves the research and design of smart application processors SoC, high-performance analog devices, and wireless interconnection chips. The main products are smart application processors SoC, high-performance analog devices, and wireless interconnection chips. The company’s products meet the application needs in industrial, automotive, and consumer fields and are widely used in smart hardware, smart robots, smart home appliances, smart IoT, smart automotive electronics, tablets, network set-top boxes, as well as power simulator devices, wireless communication modules, and other product markets.

(2) Main Operating Model

Procurement and production model: the company adopts a Fabless model, responsible for the design of integrated circuits, while the manufacturing, packaging, and testing of integrated circuits are completed through outsourcing. The company procures wafers from foundries and packaging and testing services from integrated circuit packaging and testing companies.

Sales model: after the qualified chips are delivered to the company, the company sells the chip products to solution providers and OEMs. Solution providers purchase finished chips and sell them to OEMs after secondary development, while OEMs produce various terminal electronic products.

R&D model: the company insists on independent research and development of key core technologies, selectively integrates mature IP resources in the industry, and provides distinctive competitive product combinations for target market customers in a timely manner. While ensuring the R&D of existing product technologies, the company is also preparing technologies for the next generation of products.

(3) Operating Situation

1. Categories of Main Chip Products

According to the definitions in the China Securities Regulatory Commission’s “Guidelines for the Classification of Listed Companies”, the industry in which the company operates belongs to “C Manufacturing Industry -> 39 Manufacturing of Computers, Communications, and Other Electronic Equipment”. According to the National Economic Industry Classification and Code (GB/T4754-2017), the products and services operated by the company belong to “65 Software and Information Technology Services -> 652 Integrated Circuit Design”. According to the definition of “National Key Supported High-tech Fields”, the technology field in which the company operates is classified as “I. Electronic Information -> (II) Microelectronics Information -> 2 Integrated Circuit Product Design Technology”.

2. Major Domestic and International Competitors

Major domestic and international competitors include MediaTek, Amlogic, Rockchip, and StarChip Technology.

3. Infrastructure of Main Chip Product Packages

The company has been committed to providing system-level SoC product packages for customers. In order to enhance R&D delivery capability and accelerate product iteration speed, the company has continuously built and improved various technology and product platforms, and has developed four platforms through years of accumulation:

(1) SoC Design Platform: includes process technology platforms (mature processes, advanced processes), mixed-signal IP, codec and display IP technology, SoC multi-core heterogeneous and bus, system low-power consumption technologies, etc.

(2) Hardware System Platform: has formed a series of supporting chips for SoC, including power management chips, wireless interconnection chips, audio processing chips, and complete hardware system design, including signal and power integrity, thermal design, manufacturability design of board-level design technologies.

(3) Software Development Platform: provides complete support for various operating system platforms such as RTOS/Linux/Android, as well as ecological adaptation of mainstream domestic operating systems; simultaneously, in conjunction with product applications, it has formed corresponding middleware and application software delivery.

(4) Ecological Service Platform: provides technical support services, open cooperation ecology, and quality system services for service empowerment, while providing efficient toolchain support for downstream customers.

The overall infrastructure diagram of the SoC product package is as follows:

■

4. Main Chip Product Downstream Application Fields and Application Examples

■

5. Development of New Technologies and Future Trends

(1) Rapid Development of Artificial Intelligence Technologies

Artificial intelligence, generally speaking, is the simulation of human consciousness and thinking processes, but with the advancement of technology, it has gradually transcended the concept of humanoid. For example, abstracting recognition of structures, identifying and matching them into various patterns of machine thinking, utilizing machine learning and data analysis methods to supplement and enhance human thinking capabilities. As a key productivity tool, artificial intelligence is accelerating its deep integration with various industries, driving industrial upgrades. AI technology has evolved from early auxiliary tools to intelligent agents with autonomous decision-making capabilities, playing a significant role in various industries such as autonomous driving, smart homes, security monitoring, medical devices, robotics, and smart education, becoming the core driving force behind the intelligent transformation of industries.

By 2025, the technical framework and application paradigm of AI agents will mature, promoting the evolution of artificial intelligence from tool-based assistance to autonomous task execution and decision collaboration. The continuous breakthroughs in large model technology are reflected in its core progress in understanding complex instructions, multi-step planning, and unified handling of cross-modal tasks, providing reliable support for high-value scenarios such as code generation and personalized education. The prosperity of open-source ecosystems accelerates technological inclusiveness, enabling high-performance models to be deployed in a wider range of industries at a lower cost. Simultaneously, multi-modal large models are deeply integrating text, images, voice, and video understanding, enhancing experiences in creative content generation, intelligent interactive assistants, and realizing plug-and-play deployments in vertical fields such as education, healthcare, and industrial quality inspection, demonstrating the tremendous potential of AI in empowering the real economy and improving operational efficiency.

With users’ increasing demands for response speed, cost, security, privacy, and personalization, applications related to large models are rapidly migrating to the edge and adapting through performance enhancement of edge computing capabilities and optimization of large model algorithms, forming a multi-layer application architecture of cloud, edge, and terminal. Edge products represented by AI smartphones and AI PCs have already spawned numerous emerging hardware products and diverse software applications. As related technologies evolve and innovate, the development of edge AI will cover various forms from industrial devices to consumer electronics, bringing new demands for hardware, chips, and challenges for technological innovation.

For the successful implementation of edge AI products, three factors must be present: computing power, algorithms, and data. The recent boom in AI has largely benefited from the accumulation of big data and the significant enhancement of AI-specific computing power. As the large-scale implementation of edge AI and large models accelerates, the requirements for end hardware in terms of computational performance, heterogeneous collaboration, and energy efficiency have become higher. Leading overseas enterprises are speeding up the deep integration of large models into their operating systems, core applications, and hardware ecosystems, striving to build a complete experience loop of cloud-edge collaboration. This trend brings a clear and strong demand pull to the upstream chip design industry, driving it to seek structural breakthroughs in terminal chips that support the new generation of lightweight large models. To meet the dynamic balance requirements of computing power, memory bandwidth, and energy efficiency for lightweight large models, enhance the ability to efficiently support complex models, and push advanced applications such as real-time multi-modal interaction into low-power, high-response edge computing scenarios. Meanwhile, from open-source models to toolchain ecological collaboration, the technical system of “cloud training - edge inference” is continuously being improved, collectively forming the core engine driving the deep development of the AI industrialization.

(2) Increased Demand for High-Performance Computing

The rapid proliferation of edge AI scenarios (such as smart terminals, autonomous driving, IoT devices) has led to an explosive growth in the demand for computing power in SoC chips, driving SoC design into a new stage of “advanced processes, computing power upgrades, architecture reconstruction, and energy consumption revolution.”

Advanced processes and packaging technologies: Utilizing advanced processes to enhance SoC computing power has become the choice for high-performance SoCs. 3D stacking, Chip to Chip, and Die to Die chip interconnection methods have also become effective solutions to address high costs and low yields in processes.

Continuous upgrades of general-purpose and dedicated computing power: As the complexity of application scenarios for edge products increases, the number of AI application scenarios grows, and the functional integration and density rise, the demand for CPU, GPU, NPU, and other computing power provided by SoC chips has dramatically increased. High computing power, multi-core, high frequency, ultra-high definition, and multi-channel DDR will become standard for high-performance SoCs.

Architectural challenges of heterogeneous collaboration: The computing units of edge SoC chips typically include CPU, GPU, NPU, VPU, and DSP acceleration units, which must achieve collaborative computation and data sharing under different application scenarios. The overall system architecture of SoC poses new challenges; it requires the continuous optimization of bus bandwidth and priority mechanisms, provision of multi-channel high bandwidth DDR solutions, unification of memory architecture, and memory consistency management between chips, dynamic task scheduling algorithms, and multi-chip interconnect technologies to provide more solutions for SoC architecture upgrades.

Effective management and control of energy consumption: Whether for cloud training or edge inference, the scale of required computations and the number of systems are unprecedented, and energy consumption is also rapidly increasing. The energy consumption optimization strategy of SoC will upgrade from “local energy consumption optimization” to “system-level energy efficiency optimization”: through dynamic voltage frequency scaling (DVFS) technology and AI-driven load balancing strategies to enhance multi-core utilization, such as AI-driven task scheduling algorithms and AI predictive power management; hybrid process designs will become a key strategy to balance performance and cost, such as Chiplet designs and advanced 3D packaging designs.

(3) Industrial Control Intelligentization

Against the backdrop of global manufacturing transformation and upgrading, industrial control intelligentization is advancing from “automation” to “autonomy.” This transformation is characterized by the deep integration of a new generation of information technologies (AI, edge computing, digital twins, etc.) with traditional industrial control, driving industrial systems towards capabilities of perception, decision-making, and execution in a closed loop, posing new challenges and opportunities for underlying chip platforms.

With the high-end, intelligent, and green transformation of domestic manufacturing, the current transformation presents the following characteristics:

Human-machine collaboration mode: the capabilities of industrial robots are improving, and human-machine collaboration requires chips to have high real-time performance and multi-task parallel processing capabilities.

AI technology integration: Generative AI, AR/VR + digital twins empower applications in design, operations, training, etc., requiring chips to support edge AI inference and achieve multi-modal data fusion such as visual and voice.

Deeply integrated systems: the requirements of smart factories for equipment interconnection and data communication drive chips to develop towards heterogeneous multi-core, hard real-time, and high-security directions, and to be compatible with various industrial bus protocols.

Hardware computing power upgrade: Emerging applications such as embodied intelligent robots show an order-of-magnitude increase in computing power demand; high-performance CPUs, AI acceleration units (NPU), PCIe, CAN, and Gigabit Ethernet and other industrial-grade high-speed connection interfaces have become standards for the next generation of industrial chips.

(4) Automotive Intelligence

With the booming development of the domestic automotive industry, to promote the localization of automotive intelligent technology, the government has successively introduced relevant policies to support local enterprises in developing core technologies such as chips and operating systems and guide the localization of the industrial chain by establishing domestic standards, reducing external reliance, and enhancing industry security and competitiveness. Meanwhile, the automotive industry is rapidly developing in terms of intelligence, electrification, and connectivity, with the electronic and electrical architecture of vehicles upgrading from traditional distributed to centralized domain controller architectures. This transformation demands higher performance from automotive-grade SoCs, with leading domestic automakers taking the lead in full-stack integration of hardware, algorithms, and operating systems, evolving from traditional cockpit domains, co-driver domains, and control domains to cross-domain integrated high-level architectures of “cockpit-parking integration, cockpit-driving integration.”

In terms of hardware: “advanced processes, high-end architectures, computing power stacking, bus-level cascading expansion” achieves “high-performance heterogeneous architecture + AI large computing power + functional safety + information security + high-speed bandwidth & expansion interfaces,” realizing single-chip integration of “cockpit, parking, driving.”

In terms of software: “hardware isolation + hardware virtualization” cross-domain integration architecture achieves multi-screen interaction in the cockpit + AI multi-modal interaction, perception, and decision-making for assisted driving, ensuring the functional safety and information security of the “driving, parking” domains.

Functional redundancy is embedded in software OTA activation, meeting commercial needs for price configuration segments and customized paid upgrades for OEMs.

In addition to cockpit intelligentization, vehicle-road collaboration technology is also being advanced simultaneously. Vehicle-road collaboration (V2X) technology, as a key to achieving intelligent transportation, is currently experiencing rapid development. Through information exchange between vehicles and vehicles (V2V), vehicles and infrastructure (V2I), and vehicles and people (V2P), functions such as traffic flow optimization, accident warnings, and intelligent parking can be achieved.

Based on the above trends, OEMs and Tier 1 suppliers are actively laying out these emerging technological paths to meet the diverse needs of different vehicle models and markets. With the improvement of process technology, the performance of hardware such as SoC chips continues to enhance, providing strong computing power support for intelligent cockpits; at the same time, SoC chips integrated with more functional modules will further enhance the integration and performance of intelligent cockpits.

(IV) Operational Situation during the Reporting Period

The company insists on continuous high-intensity investment in new technologies, new chips, and new applications, actively expanding new markets, new customers, and new applications in intelligent automotive electronics, industrial control, and consumer electronics. The company’s new products and solutions have successfully achieved mass production, driving the growth of the company’s performance. During the reporting period, the company achieved operating income of 283,795.39 million yuan, an increase of 24.04% compared to the same period last year, and a net profit attributable to listed company shareholders of 26,213.26 million yuan, an increase of 57.20% compared to the same period last year.

1. Enhancing Product Competitiveness through Technological Innovation

(1) Continuously Building a High-Performance General-Purpose Heterogeneous Computing Platform

With the rapid development of artificial intelligence technology, the demand for high computing power, heterogeneous integration, high bandwidth, and extreme energy efficiency is growing exponentially. The company is committed to building a serialized general-purpose heterogeneous computing platform and product matrix to comprehensively promote intelligent upgrades across various fields.

During the reporting period, the company successfully achieved large-scale production of complex heterogeneous chips covering CPU, GPU, NPU, DSP, and RISC-V coprocessors by deeply optimizing bus architecture, intelligent scheduling algorithms, and underlying operating systems, marking the company’s capability to provide flexible computing power combination strategies that can accurately match the differentiated needs of different application scenarios. On the ARM architecture, the company has completed multi-tier high-performance processor layouts, including octa-core A55, octa-core A73 + A53, octa-core A76 + A55, etc.; while actively exploring the application of RISC-V processors in coprocessors and main processors to build a controllable computing power foundation. In the audio processing field, leveraging DSP power supplements like HiFi4 and HiFi5, the diverse audio application needs have been fully met. Looking to the future, the company is actively positioning itself in cutting-edge technologies, continuously tackling high computing power platforms, exploring breakthroughs in 100K+ DMIPS CPU computing power, 512G+ Flops GPU computing power, and 10T+ Flops NPU computing power, conducting Chiplet technology research to enhance chip integration and performance flexibility; advancing 8K ultra-high-definition codec and display technology research integrated with AI to continuously empower ultra-high-definition multimedia; and delving into high-speed interconnect technologies, conducting research on high-speed SerDes to lay a solid foundation for the next generation of high-performance computing platforms.

(2) Improving AI Algorithms and Application Implementation

The company continues to delve into typical scenarios such as vision, speech, display, and human-machine interaction, actively reserving and adapting various AI algorithms to expand their application in various sub-fields; by promoting the collaborative upgrade of hardware, software, and algorithms, the company continuously optimizes scenario experiences, driving progress and innovation across various fields.

In visual technology, the company continues to deepen the integration of ISP and AI, aiming to enhance visual experience, reduce energy consumption, and broaden application scenarios. In terms of visual experience, the new generation AI-ISP optimizes algorithms for low-light environments, achieving a sensitivity increase of 2-4 times under the same signal-to-noise ratio conditions, significantly improving the performance of color reproduction in dark areas and noise control. In terms of energy efficiency, the new generation AI-ISP reduces memory usage and bandwidth requirements by nearly 30%-50%. In security application scenarios, the company continuously enriches its AI algorithm library, adding self-developed algorithms such as “AI Human-Vehicle-Pet Three-in-One Monitoring,” “AI Infant Crib Monitoring,” and “AI Human-Vehicle Perimeter Prevention,” in addition to existing algorithms for face/human detection and recognition, human tracking, vehicle/package/pet detection; simultaneously, AI SR algorithms are used for targeted optimization of infrared imaging technology, improving the quality of thermal infrared imaging to meet diverse needs in social management and family care. In photography scenarios, functions such as AI ISP smart imaging, AI night portrait, AI ultra-clear images, AI beauty, AI portrait blurring, AI cutout, and AI magical sky replacement have been launched, significantly enhancing picture quality and creative freedom.

In addition, the company will accelerate the application of AI vision on the edge, covering diverse scenarios such as photography, video communication, smart automotive, robot perception, and industrial inspection, fully promoting the commercialization process of algorithms like AI vision noise reduction, AI image detection, and AI visual recognition, building a new ecosystem of edge intelligence.

In ultra-clear display technology, the company continues to enhance the display experience through deep integration of AI and traditional algorithms. Addressing the pain points of low frame rates of massive internet video sources and high refresh rates of display devices causing playback stuttering and insufficient smoothness, the company utilizes heterogeneous computing power for parallel acceleration to achieve AI MEMC (AI Motion Estimation and Motion Compensation) technology, supporting input of video sources with any resolution and frame rate, and outputting real-time intelligent frame-interpolated 60fps videos, greatly improving viewing smoothness. Meanwhile, addressing playback needs for low-definition videos in cloud computers, mobile devices, and commercial displays, the company continues to optimize AI SR super-resolution technology, achieving upscaling of 480P videos to 4K resolution.

(3) Upgrading Core Technologies to Improve Product Series in Sub-fields

Based on the general-purpose computing platform, the company closely addresses the pain points of customers in sub-fields, relying on a unified, efficient, and high-quality R&D platform to rapidly iterate chip products and solutions, continuously promoting the upgrade of core technologies and the improvement of product matrices in various sub-fields.

In the smart tablet field, the company completed the verification of the new generation affordable smart tablet chip A333 and achieved large-scale production. This chip forms a more complete product matrix with the existing A1X series, A5X series, and A7X series tablet chips, and has passed Google Android 16 GMS Express certification. Meanwhile, the company has initiated the upgrade and iteration work for the new generation A7X series chips, aiming to further enhance product experience and competitiveness.

In the robotics and industrial control field, the company released the new generation control-type robot chip MR153, which has begun trial production in customer projects. This chip is equipped with a quad-core ARM processor and a dedicated RISC-V real-time processor, as well as rich interface resources such as UART, PWM, and GPIO, which can more accurately support the access of various sensors like infrared sensors, gyroscopes, ultrasonic, line lasers, and ToF, enhancing real-time computation and control performance. Currently, the company has completed a serialized layout of MR153, MR527, and MR536 in the robotics field.

In the industrial control and human-machine interaction field, the company released the new generation industrial control chip T153, which can be used for industrial PLCs, industrial HMIs, industrial gateways, and power applications. This chip is equipped with a quad-core ARM processor and a dedicated RISC-V real-time processor, with three Gigabit Ethernet ports, dual CAN-FD interfaces, and LocalBus, supporting high-throughput network connections, perfectly meeting the needs of complex data-driven applications. In addition, T153 provides rich peripherals such as 24-channel GPADC, 6-channel TWI interfaces, and 30-channel PWM, enhancing the integration flexibility and expandability of automation systems; the accompanying multi-core AMP architecture software development package fully meets the stringent requirements of industrial automation for high real-time performance.

In the smart vision field, the company completed the verification and trial production of the new generation smart security chip V861. V861 has fully upgraded the AI-ISP image processing unit and H.264/H.265 encoder, providing 4K video processing capabilities and supporting direct connections with three cameras. V861 also integrates the fully self-developed NPU core AWNN100, providing better support for AI algorithm implementation and scene landing. In addition, the company released the new generation AI smart glasses chip V881 and completed sample testing and verification. V881 has upgraded encoding capabilities to 4K30fps based on the V821 smart glasses chip, with a photography capability of 20 million pixels and better image processing capability, while upgrading wireless capabilities to WiFi6 to improve transmission efficiency and user experience. Currently, the company has formed a complete serialized layout of V821, V831, V851, V861, and V881 in the smart vision field, achieving comprehensive coverage of the 2M to 8M market.

In the intelligent decoding display field, the company completed the product release and large-scale production of the second generation smart projection chip H723 series and the H135 series for ultra-micro projections, as well as the tape out and sampling of the high-performance overseas video certified set-top box chip H626 and the smart projection chip H736. H723 is positioned as a smart projection chip, supporting decoding capabilities of 8K24fps and providing various display output interfaces; H135 is positioned as a micro-projection chip, with 1080P video decoding and display output capabilities; H736 is positioned as an overseas video certified projection chip, supporting AV1 hardware decoder, and upgraded to a quad-core A55 CPU and G310 GPU architecture, enhancing product performance and user experience; the entire H723, H736, and H135 series are equipped with built-in hardware trapezoidal correction engines, effectively improving UI refresh rates, reducing picture jaggedness and input latency, enhancing the projection experience. In the intelligent decoding set-top box field, the company released the overseas video certified and operator set-top box chip H626, using a quad-core A55 CPU and G310 GPU architecture, supporting AV1 decoder and a new generation of security architecture, supporting HDMI 2.1 output interface and USB3.1 interface, filling the gap in the overseas certification set-top box market.

(4) Supporting SoC Peripheral Chips to Enhance Solution Competitiveness

As the company’s product downstream application landscape continues to expand, the demand for supporting chips and overall solutions in various application scenarios is becoming increasingly urgent. Therefore, the company continues to increase R&D investment and accelerate the launch of high-performance supporting products to build a more competitive “SoC + supporting chips” product ecosystem.

In the wireless product field, the company successfully completed the tape out and verification of the first chip supporting dual-band 2.4G/5.8G and 80M bandwidth of WiFi6 and dual-mode Bluetooth 5.3 Combo, and has entered the customer trial production stage.

In the power management and power monitoring field, the fast-charge chip AXP517 supporting PD 3.0 protocol and the high-precision power meter chip AXP2602 have achieved large-scale production, while a dedicated power management chip AXP333 for security products has also entered the promotion stage.

2. Deepening the Application Market and Perfecting the Industrial Layout

During the reporting period, the company’s business development status in major application markets is as follows:

(1) Robotics and Industrial Control

With the continuous development of artificial intelligence technology, AI technology is also helping to drive the intelligent upgrade of industrial and robotic systems, with upgrades in perception, cognition, and control becoming the main direction for future product development.

In the robotics field, the company’s AI robot chip MR536 has been successfully introduced to several leading industry clients and core solution providers. Several products such as floor-cleaning robots and mowing robots built on this chip, with superior performance combining perception, visual obstacle avoidance, high-precision map positioning, and mixed cleaning capabilities, have been launched and achieved large-scale shipment. Meanwhile, leveraging the new generation control-type robot chip MR153, the company has developed entry-level service robot products and dedicated control modules in collaboration with several leading clients, with related products successfully entering mass production, further enriching the company’s product tier in the robotics field.

In the industrial control field, the company is promoting the application of high-performance chip T536 and control chip T153 among leading industry clients, with related product forms covering diversified scenarios such as power equipment, PLCs, industrial gateways, 3D printers, industrial HMIs, and industrial edge computing devices. Currently, industrial development boards equipped with T536 and T153 have been launched for sale. With excellent performance, high reliability, and wide scene adaptability, T536 won the “New Quality Award of Industrial Chips” at the “China Automation + Digital Industry Annual Conference” organized by China Industrial Control Network, fully demonstrating the company’s technical strength and product innovation capability in the smart industrial field. Meanwhile, the company is actively embracing the industrial open-source ecosystem, participating in open-source community activities, adapting to the domestic open-source Harmony operating system, and has received the ecological product compatibility certificate awarded by the Open Atom Open Source Foundation. At the 25th China International Industrial Expo, T536 won the “Integrated Circuit Innovation Achievement Award.”

(2) Leveraging Computing Power and Cost Performance to Lay Out Automotive Cockpits

In the smart automotive electronics market, the company is deepening its cooperation with mainstream automakers, focusing on promoting the landing of front-mounted designated project proposals based on the T527V platform, consolidating existing cooperation results while successfully expanding multiple new clients and new projects. The T736 smart cockpit solution, aimed at higher performance demands, has begun delivery, and discussions on a new round of designated projects are underway with several leading automakers.

The company will continue to promote chip solutions in the front-mounted smart cockpit field around these two platforms. To date, through active R&D collaborations with leading domestic automakers, the company has accumulated various intelligent module solutions such as smart cockpits, full-digital dashboards, AR-HUDs, smart laser headlights, and intelligent warning systems. With the gradual maturity of large model technology, the company will actively explore opportunities for large model applications in automotive intelligence and invest in R&D of related technologies and products, seizing industry opportunities for overall vehicle intelligence.

(3) Deepening the Tablet and Industry Applications around Platform Chips

In the general-purpose smart terminal market, the continuous evolution of the Android ecosystem has become the core engine driving industry development.

On the one hand, the company continuously improves the product matrix of high-performance smart terminal chips, expanding its application landscape. During the reporting period, the A537 chip, equipped with an octa-core A73 + A53 processor, has achieved large-scale production in the smart tablet market; with its excellent performance and power balance, this chip has received a good market response in the mid-range market and has further extended to smart education and smart home fields. Meanwhile, the A733 chip, equipped with an octa-core A76 + A55 processor, is actively engaged in joint development of educational tablets, commercial display devices, cash register terminals, and cloud computers with core clients, with related products officially entering mass production and progressing steadily.

On the other hand, the company deeply integrates AI technology, actively promoting the iteration and upgrade of smart terminals towards “+AI” to meet the demand for transition from traditional computing to intelligent computing across various industries. Closely collaborating with ecological partners, leveraging the powerful CPU + GPU + NPU heterogeneous computing power of A733, the company has actively developed and adapted AI super-resolution, AI image enhancement, intelligent voice interaction, and AI-assisted learning technologies in applications such as smart tablets and educational tablets, successfully producing multiple differentiated competitive “+AI” products, significantly enhancing user experience and product added value. In the future, the company will continue to promote traditional smart terminals to upgrade to AI smart terminals with proactive perception, deep learning, and intelligent decision-making capabilities.

(4) Decoding and Home Entertainment

In the smart set-top box market, benefiting from the growth in overseas video playback demand, the company’s H313 and H618 series products have received good market feedback for their excellent multimedia playback capabilities and compatibility. The octa-core smart media processor H728, equipped with 2T NPU, has achieved large-scale production, further enriching the high-end product line. The third-generation set-top box chip H626 for certification of overseas high-definition content has also successfully completed tape out and sampling, laying a solid foundation for expanding into global markets.

In the smart projection market, as single-chip LCD projection technology matures, the sales of home projectors continue to grow rapidly. During the reporting period, the company launched the ultra-micro projection chip H13X series and the second-generation smart projection chip H72X series. With excellent picture quality performance and the company’s self-developed hardware trapezoidal correction engine, these products have gained customer recognition and achieved large-scale production, becoming mainstream solutions in the smart projection market. Additionally, the third-generation smart projection chip H736, supporting overseas content certification, has also completed tape out and sampling verification, further enhancing the company’s influence in the smart projection field.

In the smart TV market, after successfully achieving mass production of the first-generation TV303 chip, the company has completed sample verification of the second-generation smart TV chip TV323 to meet the upgrade demands of the smart TV market and has begun promoting it to downstream customers.

(5) Smart Vision and Security Applications

In the smart vision and security market, during the reporting period, V821, with good performance and high integration, covered dozens of core security customers and quickly achieved large-scale production and continuous growth; at the same time, due to its complete solutions and product experiences, the V821 smart glasses solution has become the mainstream solution among glasses mass production schemes, gaining favor and recognition in the eyewear market, and has completed production at the million-level scale; new products, the security chip V861 and imaging chip V881, have been fully upgraded in terms of performance, picture quality, and power consumption, while benefiting from high inheritability of their product packages, clients are also rapidly adopting and developing them.

In response to security and imaging vision products and scenarios, the company has achieved new upgrades and application implementations in image noise reduction technology, stabilization, focusing technology, and image super-resolution technology, accelerating the expansion of market fields. In the smart wearables field, visual glasses and wearable products have achieved million-level production scale; in the smart travel field, several popular products of driving recorders have been created; in the smart education field, a series of AI camera products for learning have been launched; and in the smart security field, a comprehensive monitoring product system covering indoor to outdoor has been established. Thus, the company has successfully built a comprehensive visual product layout covering “people, vehicles, and homes,” forming a rich shelf of visual products across all segments and multiple fields.

3. Major Accounting Data and Financial Indicators

(1) Major Accounting Data and Financial Indicators for the Past Three Years

Does the company need to restate or adjust previous year’s accounting data?

□ Yes √ No

Yuan

■

(2) Major Accounting Data by Quarter

Unit: Yuan

■

Are there any significant discrepancies between the above financial indicators or their totals and those disclosed in the company’s quarterly report or semi-annual report?

□ Yes √ No

4. Share Capital and Shareholder Information

(1) Number of Common Shareholders and Preferred Shareholders with Voting Rights Restoration and Shareholding Table of the Top 10 Shareholders

Unit: Shares

■

Shareholders holding more than 5%, the top 10 shareholders, and the participation of the top 10 unrestricted circulating shareholders in the securities lending business.

□ Applicable √ Not applicable

Changes in the top 10 shareholders and the top 10 unrestricted circulating shareholders due to lending/returning in the securities lending business compared to the previous period.

□ Applicable √ Not applicable

Does the company have voting right differential arrangements?

□ Applicable √ Not applicable

(2) Total Number of Preferred Shareholders and Shareholding Table of the Top 10 Preferred Shareholders

There are no preferred shareholders in the company during the reporting period.

(3) Disclosure of the Property and Control Relationship between the Company and Actual Controllers in Flowchart Form

■

5. Bond Status Existing on the Date of Approval of the Annual Report

□ Applicable √ Not applicable

III. Important Matters

During the reporting period, there was no significant change in the company’s operating situation. For detailed matters during the reporting period, please refer to the “2025 Annual Report.”

Zhuhai Allwinner Technology Co., Ltd.

Legal Representative: Zhang Jianhui

March 26, 2026

Stock Code: 300458 Stock Abbreviation: Allwinner Technology Announcement No: 2026-0327-002

Zhuhai Allwinner Technology Co., Ltd.

2025 Annual Report Disclosure Notice

The company and all members of the board of directors guarantee that the content of the information disclosure is true, accurate, and complete, without false records, misleading statements, or major omissions.

The full text of Zhuhai Allwinner Technology Co., Ltd.'s “2025 Annual Report” was disclosed on March 27, 2026, on the information disclosure website of the Shenzhen Stock Exchange designated by the China Securities Regulatory Commission, www.cninfo.com.cn. Investors are advised to review it carefully.

This announcement is made.

Board of Directors of Zhuhai Allwinner Technology Co., Ltd.

March 27, 2026

A wealth of information and precise interpretations, all available on the Sina Finance APP.