2026 年 1 月 25 日 联合研究 光研集萃（2026 年 1 月第 3 期） ——光大周度观点一览 要点 策略周度观点 保持稳健，持股过节。本周市场继续小幅震荡上行，同时前期上涨行业出现了一 定程度的分化，市场交易热度也出现了一定程度的降温。参考之前的市场行情， 我们认为春节前市场将会保持震荡，难以保持稳定的趋势，这主要与春节之前投 资者交易热度有所下行，以及微观流动性短期趋紧有关。从历史情况来看，春节 前 20 个交易日，主要指数上涨概率不足 50%。预计春节之后市场将会迎来新一 轮上行动力，春节后 20 个交易日主要指数上行概率与平均涨幅均较高。因此建 议投资者近期以稳为主，但仍应持股过节。结构层面上，成长风格与小盘风格通 常在春季行情中明显占优，成长领域仍然将是中期主线，建议关注基本面更加扎 实的半导体产业链、AI 硬件、新能源等方向。 本周重点行业 【电新】1）光伏：行情跟随商业航天消息面催化，马斯克表示 SpaceX 与 Tesla 将在 3 年内在美国建设 100GW 光伏产能，并且未来将发射太阳能供电的 AI 卫 星；2）氢氨醇：在政策持续加码背景下"十五五"期间有望获得更多投资；3） ...

中经记者 吴清 北京报道 2026年1月24日，英伟达创始人兼CEO黄仁勋悄然现身上海办公室，2026年新年首度来华。 《中国经营报》多方获悉，这是他连续多年在春节前开启的中国行，行程包括上海、北京、深圳三站， 内容主要包括与员工座谈、出席供应商答谢会和分公司年会。值得关注的是，相比往年的高调亮相，今 年黄仁勋行程颇为低调，且在此次内部交流中，对当前最敏感的H200芯片话题选择了刻意回避。 这种"静默"恰如一面镜子，映照出英伟达在全球AI算力格局重塑与中美科技博弈深化背景下的复杂处 境。多位业内人士在接受记者采访时表示，这一边是美国政府刚为H200对华销售"开了一道口子"，另 一边是中国市场传出可能对其使用设限。英伟达这家独占全球AI芯片市场90%份额的巨头，正行走于一 根钢丝之上，试图在夹缝中维系其全球生态优势。 科技观察人士王如晨表示，跟过去几年尤其2025年大秀特秀不一样，这次全程未对媒体开放，官方也没 发布任何通稿。 业内人士多认为，这次访问的背景也早已不再是单纯的庆祝。2026年1月初，美国商务部在持续的压力 下，有条件批准了英伟达向中国销售其最新一代AI芯片H200。但几乎同时，有市场消息传出，中 ...

Core Viewpoint - Nvidia has missed the optimal opportunity to enter the Chinese market, particularly in the AI sector, with the recent approval to sell the H200 chip being a delayed response to changing geopolitical dynamics [2][46]. Group 1: Market Dynamics - The U.S. government has allowed Nvidia to sell the H200 AI chip to China under specific conditions, including limited customer approval and a 25% revenue share with the U.S. government [2][19]. - The release of the H200 coincides with the IPO cycles of domestic GPU companies, such as Moore Threads, which saw a significant stock price increase upon listing [3][4]. - The U.S. has defined China as its primary economic competitor, aiming to restrict China's industrial upgrade to high-end technologies [7]. Group 2: Competitive Landscape - The H200 chip, released in November 2023, boasts nearly double the inference performance of its predecessor, the H100, but is now considered outdated compared to upcoming Nvidia chips [15][16]. - Nvidia's market share in China has drastically declined from 20%-25% to single digits, and it has completely missed the rapid growth phase of China's AI market [16][36]. - Domestic AI chip companies, such as Huawei and Cambrian, have gained significant market share, with Huawei's AI chip market share reaching 40% [34][36]. Group 3: Regulatory Environment - There is significant internal debate within the U.S. regarding the sale of advanced chips to China, with proposed legislation aiming to prioritize U.S. customers over foreign sales [21][24]. - The approval of the H200 was expedited to counteract potential legislative restrictions that could limit chip exports to China [26][27]. Group 4: Market Acceptance and Future Outlook - Despite the H200's entry into the Chinese market, it faces challenges in market acceptance due to the rise of domestic alternatives that are increasingly integrated with local AI models [41][42]. - The existing domestic chips, while not matching Nvidia's top-tier performance, are sufficient for most inference tasks and are preferred for their security and reliability [41][42]. - The demand for Nvidia's chips may surge due to previously unmet orders, but the overall market landscape has shifted, making it less reliant on Nvidia than before [45][46].

Core Insights - The U.S. government has allowed NVIDIA to sell its H200 AI chips to "approved customers" in China and other regions, with a condition of a 25% revenue share to the U.S. government [1] - Jensen Huang, NVIDIA's CEO, expressed uncertainty about the future necessity of CPUs in an AI-driven era, suggesting that GPUs may eventually replace CPUs [1] - NVIDIA's revenue from data center GPUs is projected to surge from $15 billion in 2023 to $115.2 billion in the fiscal year 2025 [1] Industry Trends - The GPU market is experiencing a surge in interest, highlighted by the significant stock price increase of Chinese GPU company Moore Threads on its debut [3] - The demand for GPUs is rising due to the explosion of large model training, but the complete replacement of CPUs by GPUs is debated [4][6] - CPUs remain essential for complex task management, while GPUs excel in parallel computing tasks [4][6] Competitive Landscape - Major tech companies are accelerating the deployment of new GPU clusters, with Alibaba Cloud and Baidu developing their own chips to enhance AI capabilities [7][9] - Amazon and Google are also investing in self-developed chips to reduce dependency on NVIDIA, focusing on efficiency and cost control [9][10] - The shift towards GPU dominance in cloud computing is evident, but companies are also developing their own solutions to avoid being solely reliant on NVIDIA [9][10] Future Directions - The transition of AI tasks from cloud to local devices is reshaping the computing architecture, with GPUs becoming increasingly important in smartphones and PCs [10][11] - The rise of AI PCs emphasizes the importance of GPU performance over traditional CPU metrics [11] - The automotive industry is also leveraging GPUs for real-time data processing in autonomous driving applications [11] Ecosystem Development - CPU manufacturers like Intel and AMD are not retreating; they are adapting by enhancing their AI processing capabilities and developing competitive ecosystems [14][15] - NVIDIA's strength lies in its established ecosystem, particularly with CUDA, which poses challenges for competitors [15] - The competition in the computing sector is shifting towards who can build a comprehensive AI ecosystem, with companies like Huawei making significant strides [15][16]

Core Viewpoint - Nvidia is expected to maintain strong profitability and growth, with optimistic projections for future earnings and revenue growth, despite facing challenges in the Chinese market [4][8]. Financial Performance - Nvidia's adjusted EPS for the last quarter was $1.05, with total revenue of $46.74 billion, exceeding market expectations. For the upcoming quarter, adjusted EPS is projected to reach $1.25, a 19% quarter-over-quarter increase, while revenue is expected to be $54.8 billion, a 17% increase [1]. - Analysts predict a compound annual growth rate (CAGR) of 22.0% for EPS and 20.2% for revenue over the next five years, indicating more than double growth in both metrics [4]. - The projected profit margin is expected to expand, with the adjusted net profit margin increasing from 55.3% to 56.2% in the next quarter [1]. Growth and Profitability Projections - By January 2026, EPS is projected to be $4.54, with revenue at $207.33 billion and a profit margin of 53.98%. By January 2028, EPS is expected to reach $8.47, with revenue of $356.75 billion and a peak profit margin of 58.52% [5]. - The average net profit margin over the next five years is anticipated to be 57.0% [4]. Market Dynamics - Nvidia's market share in the Chinese AI chip market is projected to decline from 66% in 2024 to 54% in 2025, primarily due to the rise of domestic competitors and complexities in chip supply recovery [8]. - Domestic AI chip manufacturers, such as Huawei, are rapidly iterating their products, which may further pressure Nvidia's market position [9]. Competitive Landscape - The growth of domestic AI chips may impact Nvidia's pricing power, which has been strong due to its near-monopoly in high-performance AI chips. The increasing competition could lead to price pressures and a potential loss of market share [13]. - Analysts note that the Chinese market's scale provides local companies with opportunities to grow and develop their ecosystems, potentially leading to competition beyond China [13].

Core Viewpoint - The global competition in artificial intelligence (AI) is increasingly focused on the physical foundation of computing power, leading to a silent war over "Compute Sovereignty" [2][3][4]. Group 1: Understanding Compute Sovereignty - Compute sovereignty is a complex issue that must be deconstructed into three levels: the location of AI computing resources, the nationality of the companies owning these data centers, and the origin of the AI accelerators (chips) powering them [2][3]. - A survey of nine leading public cloud service providers reveals a highly uneven global distribution of computing power, with only 33 countries hosting critical AI infrastructure, indicating a significant gap between "compute-rich" and "compute-poor" nations [3][4]. Group 2: Territorial Illusions and Economic Considerations - The concept of territorial sovereignty in computing power is primarily about having physical AI data centers within a country's borders, which is seen as essential for ensuring supply security and regulatory oversight [4][5]. - The report highlights that while attracting foreign tech giants to build data centers can bring economic benefits, the environmental and resource costs may outweigh these advantages, especially for countries lacking competitive energy and climate conditions [5]. Group 3: Supplier Loyalty and Geopolitical Strategies - Merely having data centers does not equate to true sovereignty; the nationality of AI cloud service providers introduces a layer of complexity due to overlapping legal jurisdictions [6][7]. - Countries face strategic choices between two approaches: "Aligning" with a single foreign superpower's digital infrastructure or "Hedging" by diversifying suppliers to mitigate risks [8][9]. Group 4: The Chip Dependency - The report identifies a critical dependency on AI accelerators, with U.S. companies like NVIDIA dominating 80% to 95% of the global market, leading to a situation where most countries rely on U.S. technology for their AI capabilities [10][11]. - Countries like the EU and China are striving for "strategic autonomy" in chip production, but achieving this is a long-term and costly endeavor [12][13]. Group 5: Conclusion on Sovereignty - The report concludes that compute sovereignty is not a straightforward goal but a complex framework filled with trade-offs, where a nation may achieve sovereignty in one area while remaining dependent in another [13].

Investment Rating - The report maintains an investment rating of "Outperform the Market" for the computer industry [2] Core Views - The domestic software industry showed a revenue growth of 12.6% year-on-year for the first eight months of 2025, with a total revenue of 9.64 trillion yuan, indicating a continuous recovery trend [10][11] - The AI server market demand in China is increasingly concentrated in the internet sector, reflecting the impact of domestic chip localization trends on market competition [3] - The HarmonyOS 5 has surpassed 20 million terminal devices, with an upcoming significant upgrade to HarmonyOS 6 expected to enhance system capabilities [2][32] Summary by Sections Industry Data - The software industry experienced a profit growth of 13.0% year-on-year for the first eight months of 2025, with total profits reaching 13,186 billion yuan, outpacing revenue growth by 0.4 percentage points [11] - The IC design sector was the highest-performing sub-industry, with a year-on-year growth of 17.7% for the same period [15] - The domestic AI chip market is projected to grow from 21 billion USD to 38 billion USD by 2025, with a notable increase in the market share of domestic chip manufacturers [34] Localization - The localization of domestic chips is accelerating, with the AI chip localization rate increasing from 34% in the second half of 2024 to 35% in the first half of 2025 [33] - Huawei's HarmonyOS has become the second-largest mobile operating system in China and the third globally, with a market share of 17% in Q2 2025 [45][46] - The EDA sector is experiencing increased activity, with domestic companies like Huada Empyrean enhancing their product offerings to meet local demand [59][68] Computing Power - The demand for AI servers is expected to concentrate among internet companies, leading to a decrease in supply concentration [3] - OpenAI has signed significant supply agreements with major chip manufacturers, indicating a diversification in chip demand [3] - The report highlights the rapid growth of domestic AI chip companies, with notable revenue increases for companies like Cambricon and Haiguang [39][41] AI - DeepSeek has made significant advancements in optimizing AI models for domestic chips, addressing ecological barriers posed by CUDA [3] - The report emphasizes the importance of collaboration between software and hardware in the AI industry, particularly in the context of domestic chip development [3][5]

Core Insights - Nvidia, the dominant player in the global AI chip market, has seen its market share in China plummet from 95% to 0% due to tightening U.S. export controls [1] - The exit of Nvidia from the Chinese AI chip market has created significant opportunities for local companies, with Cambricon reporting a staggering 1332.52% year-over-year revenue increase in Q3 2025 [1] - The competitive landscape is shifting, with Huawei's Ascend series chips and AMD's MI300 series actively seeking to fill the void left by Nvidia's absence in the Chinese market [1] Market Dynamics - Nvidia's market share in China has effectively "returned to zero," highlighting the impact of U.S. export restrictions on its operations [1] - Cambricon's Q3 2025 revenue reached 1.727 billion yuan, a dramatic increase compared to a loss of 194 million yuan in the same period last year [1] - The exit of Nvidia has transitioned the market from a single-player dominance to a more competitive environment with multiple players vying for market share [1]

Core Viewpoint - Nvidia, the dominant player in the global AI chip market, has seen its market share in China drop from 95% to 0% due to tightening U.S. export controls, prompting the company to make various adjustments to its product offerings and strategies [2][5][10]. Group 1: Market Dynamics - Nvidia's market share in China's AI chip market was approximately 85% in 2022, with over 90% in the core area of large model training [5][6]. - The introduction of U.S. export controls in October 2022 marked a turning point, leading Nvidia to release adjusted versions of its chips, such as the A800 and H800, to maintain market presence [6][8]. - By 2023, further restrictions included the A800 and H800, forcing Nvidia to launch even lower-performance chips like the H20, which were referred to as "the most stripped-down version" [8][10]. Group 2: Financial Performance - Cambrian (寒武纪) reported a staggering 1332.52% year-on-year increase in revenue for Q3 2025, reaching 1.727 billion yuan, and a net profit of 567 million yuan, compared to a loss of 194 million yuan in the same period last year [3][17]. - For the first three quarters of 2025, Cambrian achieved a total revenue of 4.607 billion yuan, a 2386.38% increase from 185 million yuan in the same period last year [17][19]. Group 3: Competitive Landscape - The absence of Nvidia's high-end products has created a significant market vacuum, allowing domestic companies like Cambrian and Huawei's Ascend series chips to emerge as strong alternatives [17][20]. - AMD is also actively seeking opportunities in the Chinese market with its MI300 series AI chips, indicating a shift towards a more competitive landscape with multiple players [20]. Group 4: Strategic Shifts - Nvidia's global strategy is shifting, with a focus on domestic manufacturing in the U.S. and rapid iteration of its AI chip products, such as the new Blackwell platform, which boasts 208 billion transistors [14][12]. - Despite the advancements in product development, Nvidia's latest offerings appear to be increasingly isolated from the Chinese market due to ongoing export restrictions [14][15].

Core Insights - Nvidia's market share in China's AI chip sector has plummeted from 95% to 0% due to U.S. export controls [2][3] - The emergence of domestic competitors like Cambricon, which reported a staggering 1332.52% revenue increase in Q3 2025, highlights the significant market opportunity left by Nvidia's exit [3][20] - The shift in market dynamics indicates a transition from Nvidia's dominance to a more competitive landscape with multiple players [25] Nvidia's Market Position - Nvidia previously held approximately 85% of the AI accelerator card market in China, with over 90% in the large model training segment [6] - The introduction of adjusted performance chips like A800 and H800 initially helped Nvidia maintain some market presence, but these too were later restricted [9][10] - The company's latest chip, Blackwell, represents a shift towards U.S. domestic manufacturing, distancing itself from the Chinese market [18][19] Financial Performance of Competitors - Cambricon's Q3 2025 revenue reached 1.727 billion yuan, a 1332.52% increase year-over-year, with a net profit of 567 million yuan compared to a loss of 194 million yuan in the previous year [20][21] - For the first three quarters of 2025, Cambricon's total revenue was 4.607 billion yuan, a 2386.38% increase from 1.85 billion yuan in the same period last year [20] - Despite significant revenue growth, Cambricon reported a negative cash flow from operating activities of -29 million yuan [23] Market Dynamics and Future Outlook - The absence of Nvidia's high-end products has created a substantial market vacuum, allowing domestic companies to fill the gap [22][24] - The AI chip market is evolving, with a shift from training to inference, where domestic chips are gaining a competitive edge [23] - The competitive landscape is diversifying, with companies like Huawei and AMD also positioning themselves as viable alternatives in the market [24][25]