Company Overview: DingTong Technology Industry and Company - The company primarily operates in two segments: communication connectors and automotive connectors [1][2] - It supplies individual components of connectors rather than assembling them [1] - Major clients include Amphenol, Molex, TE Connectivity, and China Aviation Optical-Electrical Technology [1] Core Business Insights - **Communication Connectors**: Comprises backplane connectors and IO connectors, with a significant focus on high-speed optical modules [2][8] - **Automotive Connectors**: Includes control system connectors and high-voltage interlock connectors, with a shift towards direct supply to end customers since 2021 [1][2] Revenue Trends - The revenue composition has shifted, with communication connectors dominating until 2023, when automotive connectors saw a rise due to a decline in communication connector demand [2][3] - By the end of 2023, communication connectors accounted for approximately 80% of revenue [3] Growth Path - The company has expanded its customer base from primarily Tier 1 and Tier 2 connector manufacturers to include end automotive manufacturers and battery pack manufacturers [3][4] - R&D efforts have led to the development of high-current and high-voltage connectors since 2020, with recent advancements in 112G and 224G products [5][9] Capacity Expansion - The company has established subsidiaries in various locations, including Henan, Dongguan, Malaysia, and plans for Vietnam [6] - The management team has a strong technical background, supporting the company's R&D and production capabilities [6] Profitability and Financial Performance - The company experienced negative net profit growth in 2023 due to a decline in communication connector demand [7] - However, profitability is expected to rebound significantly starting in 2024 as demand for communication connectors increases [7] Industry Dynamics - The connector industry is seeing a trend towards concentration, with major players capturing a larger market share [7] - The communication and automotive sectors are the primary application areas for connectors, with a notable increase in competition in the automotive sector [7] Future Market Outlook - The optical module market is projected to grow rapidly, driven by AI infrastructure and increased demand for high-speed transmission [8][9] - The overall market for high-speed optical modules is expected to exceed $22 billion by 2030 [9] Profit Forecast - Projections for 2025 to 2027 indicate significant revenue growth, with expected net profits of approximately $2.45 billion, $7.84 billion, and $20.04 billion respectively [10] - The anticipated demand for 112G and 224G connectors is expected to drive this growth [10]

Summary of Conference Call on Optical Interconnect Technologies Industry Overview - The discussion focused on optical interconnect technologies, specifically AOC (Active Optical Cable), LPO (Linear Passive Optical), NPO (New Passive Optical), and CPO (Coherent Passive Optical) [1][2][17]. Key Points and Arguments AOC Technology - AOC is defined as a fixed optical fiber module suitable for short-distance multimode transmission, particularly within 50 meters, as used by Google [2][19]. - AOC's cost structure was analyzed, with 800G AOC priced over a thousand dollars and 400G AOC costing several hundred dollars [2][19]. - AOC is expected to support 1.6T transmission capabilities in the future, although current limitations exist [20]. LPO Technology - LPO modules differ from traditional optical modules by omitting DSP (Digital Signal Processor), utilizing linear drive technology, and achieving a transmission distance of up to 500 meters [4][19]. - LPO is projected to see shipments of three to four million units in 2023, potentially doubling by 2027 due to demand from major clients like Google, Microsoft, and Meta [5][19]. - The combination of AOC and LPO is anticipated for different application scenarios, with AOC for short-distance interconnects and LPO for longer distances [19]. NPO Technology - NPO optical engines are compact, low-power, and do not require an external shell, making them suitable for high-density applications like AI accelerators [7][10]. - NPO technology is expected to mature and be sold to cloud providers and equipment manufacturers, with significant market potential [9][23]. - The cost of NPO components, such as the optical engine, is estimated to be around $35 to $40 per unit [26]. Market Dynamics - The market for 800G optical modules is currently in mass production, while 1.6T modules are still in early development, expected to mature in two to three years [5][6][22]. - The supplier landscape is concentrated, with limited opportunities for new entrants, particularly in the TIA (Transimpedance Amplifier) and driver chip sectors dominated by companies like Marvell [6][22]. Technical Challenges - The discussion highlighted challenges in the 200G optical module market regarding transmission distance and industry maturity [3]. - The transition from NPO to CPO technology was noted as technically feasible but with potential challenges in product form factor [9][23]. Future Trends - The integration of optical communication technologies in data centers is expected to grow, with a focus on reducing costs and power consumption while improving system latency [10][14]. - The potential for NPO to replace existing GB300 systems was discussed, contingent on its maturity and performance improvements [25]. Additional Important Content - The role of silicon photonics in NPO technology was emphasized, with key suppliers identified as Group, Tog, and TSMC [15][28]. - The discussion also touched on the advantages of glass substrates in CPU applications, highlighting their cost-effectiveness compared to ceramic and PCB substrates [27]. - The collaboration between Google and various suppliers for LPO technology was noted, with initial production expected from companies like Xuchuang [28]. This summary encapsulates the critical insights and developments in the optical interconnect technology sector as discussed in the conference call.

中际旭创 20260115 摘要 光模块市场受益于速率和可拓展性等物理优势，预计 2026 年和 2027 年将持续扩张，尤其是在超节点元年开启后，"scale up"市场将成为 主要增长点，流量带宽呈 9 倍增长。 2026 年是光模块行业放量大年，800G 产品持续放量，1.6T 产品超预 期增长，推动行业整体规模提升。OCS、MPO 等新技术及 CPO 迭代加 速推出，提升利润率，为迎接更大市场需求做准备。 旭创作为光模块龙头企业，通过自研核心芯片器件和垂直一体化策略提 高利润率，并提前锁定供应链，应对供不应求局面。公司正向平台型公 司转型，估值有望随业绩释放而提升。 苏州旭创定位于高端产品，是 Google 认证供应商，业绩受海外云厂商 资本开支驱动。筹备港股上市可能带来估值溢价，公司在中短期和长期 均具备增长空间。 AI 算力建设推动 1.6T 光模块需求激增，预计 2026 年需求量将达到 2000~3,000 万只，同比增长超过 10 倍，表明产业存在合理性和需求。 Q&A 今年（2026 年）对光模块行业有何重要意义？ 今年（2026 年）是一个放量的大年，不仅 800G 产品进一步大放量，1 ...

Core Viewpoint - Marvell Technology announced the acquisition of semiconductor startup Celestial AI for $3.25 billion in cash and stock, potentially rising to $5.5 billion if revenue milestones are met, with a positive earnings report boosting its stock by approximately 13% in after-hours trading [1][2]. Group 1: Acquisition Details - Celestial AI will receive $1 billion in cash and $2.25 billion worth of 27.2 million shares of Marvell common stock, with a revenue milestone of $2 billion by the end of fiscal 2029 triggering the maximum payout [2]. - The acquisition is a strategic move for Marvell to enhance its position in the AI sector, particularly in light of the increasing demand for high-speed, energy-efficient chips driven by the generative AI trend [2]. Group 2: Financial Performance - Marvell's Q3 revenue reached $2.07 billion, a year-over-year increase of 36.8%, aligning with market expectations [3]. - The company forecasts Q4 revenue of approximately $2.2 billion, slightly above LSEG's expectation of $2.18 billion, and anticipates total revenue of around $10 billion for the next fiscal year [3]. Group 3: Future Outlook - Marvell expects a 25% growth in data center revenue and a 20% increase in custom chip business revenue, with no significant fluctuations anticipated in quarterly revenue [3]. - The technology from Celestial AI is expected to be integrated into "large XPU" systems and contribute significantly to revenue starting in the second half of fiscal 2028, with projected annualized revenue of $500 million in Q4 of fiscal 2028 and doubling to $1 billion by Q4 of fiscal 2029 [3].

Group 1 - The communication equipment sector experienced a significant rebound, with the Guozheng Communication Index rising by 0.45% as of 10:39 on December 2, 2025, driven by stocks such as Changjiang Communication (+6.11%) and Galaxy Electronics (+4.78%) [1] - The release of the mobile assistant technology preview by Doubao, an AI assistant software developed in collaboration with mobile manufacturers, has contributed to the positive performance of the communication sector, influenced by recent AI developments and increased global investment in communication infrastructure [1] - According to CITIC Securities, the rapid development of AI large models and applications relies on continuous investment in computing power infrastructure, with high-performance, high-bandwidth, and low-latency networks being crucial for the performance upgrade of computing clusters [1] Group 2 - As of November 28, 2025, the top ten weighted stocks in the Guozheng Communication Index accounted for 67.58% of the index, including major players like Zhongji Xuchuang and ZTE [2] - The Communication ETF (159695) closely tracks the Guozheng Communication Index, providing a convenient way to invest in the optical communication industry chain [3] - Investors can access the AI-driven optical communication investment opportunities through the Communication ETF linked fund (019072) [4]

Group 1 - The ChiNext 50 Index (399673) has shown strong growth, with notable increases in constituent stocks such as Beijing Junzheng (300223) up by 11.78%, Runze Technology (300442) up by 8.09%, and Zhongji Xuchuang (300308) up by 5.47% [1] - The AI industry chain is experiencing a resurgence, driven by a surge in global AI computing power demand, leading to a severe supply-demand imbalance in global memory semiconductors, with memory chip prices rapidly increasing [1] - The price of 16G+512G memory is expected to rise by nearly 500 RMB by the end of this year [1] - CITIC Securities reports that the rapid development of AI large models and applications relies on continuous investment in computing power infrastructure, emphasizing the importance of high-performance, high-bandwidth, and low-latency networks for performance upgrades [1] - The growth of optical interconnect technology is driven by increased penetration of Scale up networks and the value upgrade of Scale out networks, presenting significant development opportunities [1] - The advantages of leading manufacturers are becoming more pronounced due to the increasing demands on R&D capabilities, production capacity, and material delivery capabilities in AI networks [1] Group 2 - The ChiNext 50 ETF (159681) closely tracks the ChiNext 50 Index, which consists of the 50 stocks with the highest average daily trading volume in the ChiNext market, reflecting the overall performance of well-known, large-cap, and liquid companies [2] - As of November 28, 2025, the top ten weighted stocks in the ChiNext 50 Index include Ningde Times (300750), Zhongji Xuchuang (300308), and others, accounting for a total of 71.14% of the index [2]

Core Insights - The AI sector is experiencing significant growth, with the Zhongzheng AI Theme Index (930713) rising by 0.79% and key stocks like Beijing Junzheng (300223) and Zhongke Xingtou (688568) showing notable increases in their share prices [1][2] - The rapid development of AI large models and applications is supported by continuous investment in computing infrastructure, emphasizing the importance of high-performance, low-latency networks [2] Group 1: Market Performance - The Zhongzheng AI Theme Index has seen a rise of 0.79%, with component stocks such as Beijing Junzheng increasing by 5.10% and Zhongke Xingtou by 4.63% [1] - The AI Artificial Intelligence ETF (512930) has also increased by 0.78%, currently priced at 2.07 yuan [1] Group 2: Technological Advancements - AI large models are achieving breakthroughs in both theoretical and practical applications, with Alibaba's "Qianwen" app surpassing 10 million downloads in just one week, marking it as the fastest-growing AI application globally [1] - The DeepSeekMath-V2 model has introduced a self-verification mechanism, achieving gold medal status in the International Mathematical Olympiad, showcasing China's innovation in foundational algorithms [1] Group 3: Industry Outlook - The report from CITIC Securities highlights that the growth of AI large models is closely tied to ongoing investments in computing infrastructure, with a focus on optical interconnection technology [2] - The top ten weighted stocks in the Zhongzheng AI Theme Index account for 63.92% of the index, indicating a concentration of investment in leading companies within the AI sector [2]

Core Viewpoint - The rapid development of AI large models and applications relies on continuous investment in computing power infrastructure, with high-performance, high-bandwidth, and low-latency networks being key to upgrading computing cluster performance [1] Group 1: Industry Insights - Optical interconnection technology is expected to experience significant growth opportunities driven by the dual factors of increased penetration of Scale up networks and the value enhancement of Scale out networks [1] - The increasing demands for manufacturers' R&D capabilities, production capacity, material delivery capabilities, and new technology layout requirements are highlighting the advantages of leading manufacturers in the market [1] Group 2: Company Outlook - There is optimism regarding the development potential of domestic optical communication leading companies [1] - The communication industry maintains a "stronger than market" rating [1]

Core Insights - The AI computing power industry chain is experiencing significant fluctuations, with notable movements in related ETFs and stocks [1] - Google is at the center of a new investment chain, launching multiple products that create a positive feedback loop, enhancing AI revenue through increased token consumption [1] - The vertical integration of computing power, models, and applications by Google is expected to clarify the business model of AI applications and establish a sustainable profit model [1] Group 1: Market Performance - The ChiNext AI ETF (159381) decreased by 0.61%, while the 5G Communication ETF (515050) fell by 0.35% [1] - Stocks such as Putian Technology hit the daily limit, while Jinxin Nuo and Changxin Bochuang rose over 4% [1] Group 2: Google and Alibaba's Strategies - Dongwu Securities highlights that Google and Alibaba are enhancing their full-stack capabilities in the AI industry, leading to increased demand for TPU and AI ASIC chips [2] - Customized computing clusters are raising networking requirements, with Google's TPU introducing new optical interconnection solutions [2] - The optimistic capital expenditure outlook from Google and Alibaba is expected to generate substantial incremental demand for the optical interconnection supply chain [2] Group 3: Related ETFs - The ChiNext AI ETF (159381) tracks the ChiNext AI Index and has a significant allocation in optical modules, with the top three holdings being Zhongji Xuchuang (24.78%), Xinyi Sheng (19.40%), and Tianfu Communication (5.11%) [3] - The 5G Communication ETF (515050) focuses on the 5G communication theme index and has a scale exceeding 9 billion, with major holdings including Zhongji Xuchuang, Xinyi Sheng, Luxshare Precision, Industrial Fulian, and Zhaoyi Innovation [3]

Core Viewpoint - Lixun Precision (002475) reported strong performance in the first half of 2025, with revenue of 124.5 billion yuan, a year-on-year increase of 20.18%, and a net profit of 6.644 billion yuan, up 23.13% year-on-year, reflecting the company's successful transformation and growth strategy in the AI-driven market [1][3][4]. Financial Performance - The company achieved a revenue of 1245.03 billion yuan in the first half of 2025, marking a 20.18% increase compared to the previous year [1]. - Net profit attributable to shareholders was 66.44 billion yuan, representing a 23.13% year-on-year growth [1]. - Revenue from automotive and communication businesses surged, with automotive interconnect products and precision components generating 8.658 billion yuan, up 82.07%, and communication interconnect products and precision components reaching 11.098 billion yuan, a 48.65% increase [3]. Business Segments - Consumer electronics revenue remained stable at 97.799 billion yuan, a 14.32% increase, accounting for 78.55% of total revenue [3]. - The company has expanded its global production capacity, with overseas bases in Vietnam and Mexico mitigating geopolitical and tariff risks, leading to an export revenue of 1057.26 billion yuan, up 17.38% [3]. Strategic Initiatives - Lixun Precision is focusing on next-generation AI optical interconnect technology, developing high-speed optical module products to support large-scale data center construction and enhance AI computing power [1][6]. - The company has made significant R&D investments, totaling 45.94 billion yuan in the first half of 2025, with approximately 30% directed towards cutting-edge technologies [6]. Market Position and Future Outlook - The global market for optical modules is expected to grow significantly, with a projected CAGR of 22%, potentially exceeding 37 billion USD by 2029 [8]. - Lixun Precision is positioned as a leader in the precision manufacturing solutions industry, ranking fourth globally and first in mainland China [11]. - The company anticipates a net profit of 10.890 billion to 11.344 billion yuan for the first three quarters of 2025, reflecting a year-on-year growth of 20% to 25% [4]. International Expansion - The company is pursuing international capital market strategies, having submitted an application for H-share listing on the Hong Kong Stock Exchange, which could enhance its competitive advantage [2][10]. - Recent acquisitions, such as the 50.1% stake in German automotive wiring manufacturer Leoni, are part of Lixun's strategy to strengthen its position in the automotive electronics sector [12].