Dongguan Mingpu Optoelectronics Co., Ltd.
(Hereinafter referred to as "Mingpu Optoelectronics" or "the Company") is a company listed on the Main Board of the Shenzhen Stock Exchange (hereinafter referred to as "SZSE"). To meet the Company's business development needs, enhance its profitability and market competitiveness, in accordance with the "Company Law of the People's Republic of China," "Securities Law of the People's Republic of China," "Administrative Measures for the Registration of Issuance of Securities by Listed Companies" (hereinafter referred to as the "Registration Measures") and other relevant laws, regulations, and normative documents, the Company has compiled an analysis report on the proposed issuance of A shares to specific objects.
(Unless otherwise specified in this report, the terms used herein have the same meaning as in the "Dongguan Mingpu Optoelectronics Co., Ltd. 2026 Plan for Issuance of A Shares to Specific Objects.")
I. Background and Objectives of the Proposed Issuance to Specific Objects
(I) Background of the Proposed Issuance to Specific Objects
- With the rapid development of the AI industry, the global optical module market continues to grow.
Since 2024, global leading cloud vendors have continued to increase their capital expenditure on AI computing power clusters, leading to a surge in demand for high-density intelligent computing interconnection. Data communication high-speed optical devices are entering a period of rising prices and volumes. According to the latest statistics from Omdia in June 2026, the global optical device market revenue reached US$17.8 billion in 2024, a year-on-year increase of 40%. In the fourth quarter of 2024, quarterly revenue was US$5.36 billion, an increase of over 50% year-on-year, reaching a new historical high for quarterly revenue. Optical modules, as the core carriers for photoelectric signal conversion in computing power networks, are indispensable basic components for data centers and communication networks.
According to Light Counting's updated forecast report in May 2026, AI large model training and inference are driving exponential bandwidth demand, continuously expanding the global optical module market. The global optical module market is projected to reach US$26.8 billion in 2025, a year-on-year increase of 74%. By 2029, the global optical module market is expected to exceed US$51 billion, with a compound annual growth rate of 20% from 2024 to 2029. The data center Ethernet optical module segment is a key growth area, with a market size expected to exceed US$20 billion in 2029. 800G and 1.6T high-speed products are expected to contribute over 64% of the market growth. In the first quarter of 2026, global optical module quarterly sales reached US$10 billion, a year-on-year surge of 90%, with industry prosperity continuing to exceed expectations.
[Chart: Quarterly Sales of Optical Transceivers]
In the domestic market, domestic general-purpose large models and multimodal models are continuously improving, coupled with the implementation of the "East Data West Computing" national integrated computing power network, leading to the large-scale construction of domestic intelligent computing centers and ten-thousand-GPU clusters. According to the "Blue Book of Advanced Computing and Computing Power Development (2025)" by the China Academy of Information and Communications Technology, as of June 2026, China's total computing power reached 962 EFLOPS, with a compound annual growth rate exceeding 70% from 2020 to 2025. Intelligent computing power accounts for 782 EFLOPS, representing 81%, and has become the core pillar of computing power growth. Frost & Sullivan's 2026 forecast indicates that China's computing power will grow from 962 EFLOPS in 2025 to over 680 EFLOPS in 2029, with a compound annual growth rate of 17.9% from 2025 to 2029. From 2025 to 2026, China's AI industry investment landscape has become diversified, with internet giants like ByteDance, Alibaba, and Tencent, as well as the three major telecom operators and local state-owned platforms, increasing investment in intelligent computing infrastructure. Domestic large models such as DeepSeek R1, Wenxin Yiyan, and Tongyi Qianwen have achieved performance comparable to overseas leading products, accelerating the commercialization of AI. The demand for computing power on the inference side is rapidly increasing, further boosting the demand for high-speed optical modules.