在去年高增长、高基数的基础上，2024年生成式AI业务依然保持了极高的增速。这不仅彰显了以大模 型技术为基础的业务所具有的强劲生命力，也验证了商汤"AI基础设施(大装置)-大模型(日日新)-应用"三 位一体战略的可行性。该战略构建了完整的商业闭环：AI大装置为应用落地提供算力支撑，应用反馈 又指导基础设施升级；大模型能力与基础设施相互促进，实现算力效率的持续优化；同时，多元化的技 术路径加速了应用场景渗透，而应用实践又推动大模型向精细化方向发展。 特别值得关注的是，2024年生成式AI业务在总收入中的占比从34.8%提升至63.7%，是商汤集团最大的 业务收入。这标志着公司彻底完成从视觉AI向生成式AI战略转型的全面胜利，其收入结构变革已形成 不可逆的良性发展态势。 一、"三位一体"战略持续巩固技术领先地位，训练效率优于DeepSeek公开报告 商汤科技依托"大装置-大模型-应用"的深度协同创新，形成核心技术差异化优势，并持续巩固技术领先 地位。 2024年，商汤生成式AI业务收入突破24亿元，同比增长103.1%，连续两年保持三位数增长。这一强劲 表现推动商汤2024年全年收入同比增长10.8%，达到37. ...

Core Insights - The rapid advancement of AI models is accelerating, with improvements in the last six months surpassing those of the previous six months, driven by three overlapping expansion laws: pre-training expansion, post-training expansion, and inference time expansion [1][3]. Group 1: AI Model Developments - Claude 3.7 showcases remarkable performance in software engineering, while Deepseek v3 indicates a significant reduction in costs associated with the previous generation of models, promoting further adoption [3]. - OpenAI's o1 and o3 models demonstrate that longer inference times and searches yield better answers, suggesting that adding more computation post-training is virtually limitless [3]. - Nvidia aims to increase inference efficiency by 35 times to facilitate model training and deployment, emphasizing a shift in strategy from "buy more, save more" to "save more, buy more" [3][4]. Group 2: Market Concerns and Demand - There are concerns in the market regarding the rising costs due to software optimizations and hardware improvements driven by Nvidia, potentially leading to a decrease in demand for AI hardware and a symbolic oversupply situation [4]. - As the cost of intelligence decreases, net consumption is expected to increase, similar to the impact of fiber optics on internet connection costs [4]. - Current AI capabilities are limited by cost, but as inference costs decline, demand for intelligence is anticipated to grow exponentially [4]. Group 3: Nvidia's Roadmap and Innovations - Nvidia's roadmap includes the introduction of Blackwell Ultra B300, which will not be sold as a motherboard but as a GPU with enhanced performance and memory capacity [11][12]. - The B300 NVL16 will replace the B200 HGX form factor, featuring 16 packages and improved communication capabilities [12]. - The introduction of CX-8 NIC will double network speed compared to the previous generation, enhancing overall system performance [13]. Group 4: Jensen's Mathematical Rules - Jensen's new mathematical rules complicate the understanding of Nvidia's performance metrics, including how GPU counts are calculated based on chip numbers rather than package counts [6][7]. - The first two rules involve representing Nvidia's overall FLOP performance and bandwidth in a more complex manner, impacting how specifications are interpreted [6]. Group 5: Future Architecture and Performance - The Rubin architecture is expected to deliver over 50 PFLOPs of dense FP4 computing power, significantly enhancing performance compared to previous generations [16]. - Nvidia's focus on larger tensor core arrays in each generation aims to improve data reuse and reduce control complexity, although programming challenges remain [18]. - The introduction of the Kyber rack architecture aims to increase density and scalability, allowing for a more efficient deployment of GPU resources [27][28]. Group 6: Inference Stack and Dynamo - Nvidia's new inference stack and Dynamo aim to enhance throughput and interactivity in AI applications, with features like intelligent routing and GPU scheduling to optimize resource utilization [39][40]. - The improvements in the NCCL collective inference library are expected to reduce latency and enhance overall throughput for smaller message sizes [44]. - The NVMe KV-Cache unload manager will improve efficiency in pre-filling operations by retaining previous conversation data, thus reducing the need for recalculation [48][49]. Group 7: Cost Reduction and Competitive Edge - Nvidia's advancements are projected to significantly lower the total cost of ownership for AI systems, with predictions of rental price declines for H100 chips starting in mid-2024 [55]. - The introduction of co-packaged optics (CPO) solutions is expected to reduce power consumption and enhance network efficiency, allowing for larger-scale deployments [57][58]. - Nvidia continues to lead the market with innovative technologies, maintaining a competitive edge over rivals by consistently advancing its architecture and algorithms [61].

Core Viewpoint - Nvidia's innovations in AI inference technologies, including the introduction of inference Token expansion, inference stack, Dynamo technology, and Co-Packaged Optics (CPO), are expected to significantly reduce the total cost of ownership for AI systems, thereby solidifying Nvidia's leading position in the global AI ecosystem [2][4][68]. Group 1: Inference Token Expansion - The rapid advancement of AI models has accelerated, with improvements in the last six months surpassing those of the previous six months. This trend is driven by three expansion laws: pre-training, post-training, and inference-time expansion [8]. - Nvidia aims to achieve a 35-fold improvement in inference cost efficiency, supporting model training and deployment [10]. - As AI costs decrease, the demand for AI capabilities is expected to increase, demonstrating the classic example of Jevons Paradox [10][11]. Group 2: Innovations in Hardware and Software - Nvidia's new mathematical rules introduced by CEO Jensen Huang include metrics for FLOPs sparsity, bidirectional bandwidth measurement, and a new method for counting GPU chips based on the number of chips in a package [15][16]. - The Blackwell Ultra B300 and Rubin series showcase significant performance improvements, with the B300 achieving over 50% enhancement in FP4 FLOPs density and maintaining an 8 TB/s bandwidth [20][26]. - The introduction of the inference stack and Dynamo technology is expected to greatly enhance inference throughput and efficiency, with improvements in smart routing, GPU planning, and communication algorithms [53][56]. Group 3: Co-Packaged Optics (CPO) Technology - CPO technology is anticipated to significantly lower power consumption and improve network scalability by allowing for a flatter network structure, which can lead to up to 12% power savings in large deployments [75][76]. - Nvidia's CPO solutions are expected to enhance the number of GPUs that can be interconnected, paving the way for networks exceeding 576 GPUs [77]. Group 4: Cost Reduction and Market Position - Nvidia's advancements have led to a performance increase of 68 times and a cost reduction of 87% compared to previous generations, with the Rubin series projected to achieve a 900-fold performance increase and a 99.97% cost reduction [69]. - The overall trend indicates that as Nvidia continues to innovate, it will maintain a competitive edge over rivals, reinforcing its position as a leader in the AI hardware market [80].

Group 1 - The article highlights a new wave of capital expenditure expansion globally driven by generative AI technology, with significant attention on China's AI industry following the launch of the DeepSeek model [1][2] - Major foreign institutions believe that innovations like DeepSeek demonstrate China's breakthroughs in AI algorithms and its cost advantages, which are accelerating AI applications and creating vast opportunities across the AI industry chain [1][3] - The semiconductor, robotics, and computing infrastructure sectors are identified as new investment hotspots, with expectations that capital will continue to focus on areas related to the AI industry chain [1][3][9] Group 2 - The article notes that the capital expenditure of major US cloud and AI companies is projected to double from $150 billion in 2023 to $300 billion, indicating a strong commitment to AI infrastructure [3] - Nvidia's data center revenue is expected to grow from $50 billion in 2023 to approximately $180 billion, reflecting a more than threefold increase [3] - The article discusses the potential impact of DeepSeek on the market's expectations for high-performance AI chip demand, emphasizing that lower training costs could lead to increased overall demand for AI applications [4][6] Group 3 - The launch of DeepSeek is seen as a pivotal moment for reshaping the valuation logic of Chinese tech companies, with expectations for rapid AI application development in China due to low inference costs and a robust manufacturing supply chain [6][7] - The article mentions that the DeepSeek event has led to a swift return of capital to China's AI sector, with significant interest in related A-share tech companies [7] - The article emphasizes that the current AI-driven market dynamics may position China as a leader in the global technology revolution, supported by an improving policy environment [7][10] Group 4 - Experts predict that sectors such as semiconductors, robotics, and computing infrastructure will continue to see growth opportunities in the context of the AI revolution and global manufacturing upgrades [9][10] - The article highlights that while the semiconductor industry faces cyclical challenges, its diverse application scenarios remain attractive for long-term investment [9] - The article concludes that as AI technology integrates deeper into the economy, Chinese companies with manufacturing advantages and innovative capabilities are likely to gain a higher position in the global supply chain [10]

Investment Rating - The report does not explicitly provide an investment rating for the industry or specific companies involved. Core Insights - DeepSeek's significant cost advantages in training and inference have led to substantial market impacts, including a notable drop in Nvidia's stock price and market capitalization [2][11][12] - The introduction of DeepSeek's models has the potential to disrupt existing AI companies by lowering the barriers to entry for smaller firms and individuals, thereby increasing overall demand for computational resources [15][16] Summary by Sections Section on DeepSeek's Market Impact - DeepSeek achieved the top position in download rankings on both the Chinese and US App Store, coinciding with a major drop in the semiconductor index and Nvidia's stock [2] - Nvidia's market value decreased by nearly $600 billion, marking one of the largest single-day market cap losses in history [2] Section on Cost Structure - DeepSeek's training costs for their V3 model were reported to be under $6 million, utilizing approximately 2000 H800 GPUs [6][7] - The inference cost for DeepSeek's models is significantly lower than that of OpenAI, with DeepSeek charging only 3% of OpenAI's rates for similar token inputs and outputs [7][9] Section on Training Innovations - DeepSeek implemented innovative training strategies that reduced costs, particularly by optimizing the supervised fine-tuning (SFT) process [9][10] - The team utilized pure reinforcement learning (RL) without human feedback, achieving performance comparable to OpenAI's models [9][10] Section on Future Implications for AI Industry - DeepSeek's advancements may lead to increased competition among AI firms, particularly those relying on self-developed large models [12][13] - The report suggests that while Nvidia's stock may have been negatively impacted in the short term, the overall demand for their chips could increase as AI commercialization accelerates [14][16]

Core Viewpoint - The storage chip market is experiencing a resurgence driven by the increasing demand for AI infrastructure and applications, particularly in enterprise storage, while consumer-grade chip prices are declining due to oversupply [1][2][8]. Group 1: Market Dynamics - The enterprise storage market is facing a supply-demand imbalance, with strong demand leading to price increases of nearly 15% for enterprise-grade storage chips, while consumer-grade chips are seeing price declines [10][12]. - The overall storage chip market has seen a cumulative increase of over 30% since January 7, 2023, indicating a recovery phase [1][2]. - Major players like SK Hynix and Samsung are adjusting their production strategies, focusing on high-value storage chips to alleviate inventory pressures [12][17]. Group 2: Technological Advancements - High-performance storage chips, such as HBM (High Bandwidth Memory) and SSDs (Solid State Drives), are critical for modern data centers, fulfilling the needs for high efficiency and low latency [3][4]. - Domestic companies like Changxin Storage and Yangtze Memory Technologies are making significant advancements in high-density storage technology and low-power solutions, aiming to catch up with global leaders [6][7]. - The introduction of innovative architectures, such as Changjiang Storage's Xtacking technology, is expected to enhance the performance of NAND flash memory [7]. Group 3: Future Outlook - The demand for AI applications is anticipated to drive further investments in AI infrastructure, with companies like Alibaba planning to invest more in cloud and AI capabilities over the next three years [2][5]. - The market for storage chips is expected to stabilize as manufacturers implement production cuts and inventory management strategies, which will help balance supply and demand [13][17]. - Predictions indicate that NAND Flash prices may see a rebound in the second half of the year, while enterprise storage prices are expected to remain stable [15][16].

Core Insights - The introduction of DeepSeek-R1 marks a pivotal moment in the AI landscape, bridging the gap between open-source and proprietary models, with significant implications for AI infrastructure and generative AI economics [1][2][8] Open Source vs. Proprietary Models - DeepSeek-R1 has significantly narrowed the performance gap with proprietary models like OpenAI, achieving parity in key reasoning benchmarks despite being smaller in scale [2] - The emergence of DeepSeek is seen as a watershed moment for open-source AI, with models like Llama, Qwen, and Mistral expected to catch up quickly [2][3] - The competitive landscape is shifting, with a vibrant and competitive LLM market anticipated, driven by the open-source model's advancements [2][3] AI Infrastructure and Developer Utilization - DeepSeek-R1 utilizes reinforcement learning (RL) to enhance reasoning capabilities, marking the first successful large-scale implementation of this approach in an open-source model [3][4] - The model's success is expected to democratize access to high-performance AI, allowing enterprises to customize solutions based on their specific needs [3][4] - The shift in AI infrastructure is characterized by a move away from closed models, enabling more control and flexibility for developers [4] New Applications: Large-Scale AI Reasoning - Enhanced reasoning capabilities of DeepSeek open up new application possibilities, including autonomous AI agents and specialized planning systems across various industries [5][6] - The demand for GPU computing is expected to increase due to the accelerated adoption of agent applications driven by DeepSeek [6] - Companies in highly regulated industries will benefit from the ability to experiment and innovate while maintaining control over data usage [6] Generative AI Economics: Changing Cost Dynamics - DeepSeek is driving a trend towards lower costs and higher efficiency in reasoning and training, fundamentally altering the economics of generative AI deployment [7][8] - Models like R1 can be up to seven times cheaper than using proprietary APIs, unlocking previously unfeasible use cases for many enterprises [7] - The economic advantages of open-source models are expected to lead to a broader adoption of AI technologies across various sectors [7][8] Conclusion - DeepSeek represents a significant milestone in the AI industry, enabling open-source models to compete effectively with proprietary alternatives, while emphasizing the importance of high-quality, domain-specific data and labeling for future advancements [8]

Core Viewpoint - The article discusses the emergence of the "computing power equality movement," which is reshaping the underlying logic of artificial intelligence development, driven by significant reductions in AI model training costs and the democratization of technology through open-source collaboration [1][2]. Group 1: Computing Power Equality Movement - The training cost of the DeepSeek-V3 model is $5.576 million, which is significantly lower than the hundreds of millions spent by Silicon Valley giants, marking the start of the computing power equality movement [1]. - The CEO of ASML highlighted that as the training cost of AI models decreases, the demand for computing power may paradoxically increase, leading to exponential market expansion [2]. Group 2: Decentralization and Innovation - The article emphasizes a dual spiral of algorithmic innovation and open-source ecosystem collaboration that is dismantling computing power monopolies, allowing innovations to flow from tech giants to SMEs and individuals [4]. - Cloud service providers are restructuring the computing power landscape by creating decentralized networks and optimizing scheduling algorithms, with Chinese cloud providers playing a crucial role in this transformation [5]. Group 3: Challenges in Cloud Services - The article identifies a "trilemma" faced by cloud service providers: achieving model performance, stability, and accessibility simultaneously is nearly impossible, yet some players are approaching this ideal [7]. - Fire Volcano Engine's DeepSeek+ model has achieved high alignment with official models, providing full capabilities without compromising performance [8]. Group 4: Performance Metrics - Fire Volcano Engine's DeepSeek models have demonstrated superior performance in terms of response speed, with inference delays reduced to around 30ms, and achieving a 100% response rate in third-party evaluations [11][12]. - The platform can handle a throughput of 5 million tokens per minute, significantly enhancing the capacity for complex reasoning requests compared to traditional APIs [14]. Group 5: Application in Financial Sector - Fire Volcano Engine has integrated DeepSeek models into over 60 financial institutions, addressing key pain points such as data security, computing power shortages, and innovation constraints [15][16]. - The AI one-stop machine developed by Fire Volcano Engine is tailored for the financial sector, ensuring data security while meeting the high computing demands of the industry [17][19]. Group 6: Full-Stack AI Services - Fire Volcano Engine aims to build a prosperous AI ecosystem by offering a full-stack AI service that includes various models and platforms, facilitating intelligent transformation for enterprises [22][24]. - The integration of multiple capabilities, such as language processing and image generation, allows businesses to enhance efficiency and competitiveness [24][25]. Group 7: Future Outlook - The launch of DeepSeek-R1 serves as a test of cloud providers' technical capabilities, with Fire Volcano Engine demonstrating its leadership in high-demand scenarios [26]. - The company is positioned to lead the AI industry into a new era of ecological prosperity, leveraging its full-stack services to reshape the value ecosystem [26].

Core Viewpoint - Dell's stock has underperformed since November due to market concerns about a slowdown in AI data center construction, but the company is positioned to benefit from the shift towards inference computing, suggesting potential upside for its stock price [1][10]. Group 1: Market Concerns and Opportunities - The market is worried about the efficiency of AI chips leading to a slowdown in GPU demand, which could impact sales growth expectations for companies like Dell [1]. - Despite concerns, key factors are shifting favorably for Dell, particularly in the inference computing space, which is expected to perform well [1][10]. - The transition from pre-training to inference computing is anticipated to happen faster than expected, with more cost-effective data centers supporting AI inference [3][10]. Group 2: Strategic Partnerships - Dell has partnered with AMD to integrate Ryzen AI PRO processors into new Dell Pro devices, marking a significant milestone in their strategic collaboration [4]. - AMD's CEO highlighted that the total cost of ownership (TCO) for AMD's inference computing solutions is significantly lower than Nvidia's, which could benefit Dell in both PC and server markets [4][9]. Group 3: Financial Performance Expectations - Dell is expected to report solid earnings and revenue growth in its upcoming Q4 financial results, with analysts predicting a 14.46% year-over-year increase in earnings per share (EPS) to $2.52 [5]. - Revenue forecasts for Q4 are set at $24.57 billion, indicating a 10.09% year-over-year growth, with a consensus among analysts on the earnings estimates [5][6]. Group 4: Valuation Metrics - Dell's non-GAAP expected price-to-earnings (P/E) ratio is 14.50, significantly lower than the industry median of 23.87, indicating a 39.26% discount [9]. - The expected price-to-sales (P/S) ratio for Dell is 0.83, which is 73.43% lower than the industry median of 3.11, suggesting strong valuation metrics [9]. Group 5: Future Growth Catalysts - Dell is projected to benefit from a $5 billion deal with Elon Musk's xAI and an anticipated $4 billion increase in AI server shipments from FY 2024 to FY 2025 [8][9]. - The shift towards inference computing is expected to catalyze Dell's next growth phase, supported by recent strategic agreements [11].

作者：MD 出品：明亮公司 2月19日，微软宣布， 全球首款拓 扑 量子芯片Major ana 1发布， 据相关报道，该芯片由微软公司 历时近20年研发，有望于2030年之前上市。而微软的目标是未来在量子芯片上实现百万个 量子比特 的相 干操纵。 据第一财经报道，Majorana 1是基于全新的物质状态——"拓扑"构建而成的全球首款拓扑量子芯片， 采用了半导体砷化铟和超导体铝材料。 微软在2月19日发布的一篇博客中称，开发合适的材料来构建量子比特，并理解量子比特相关的物质 拓扑状态的难度极大，这也是大多数量子研究都集中在其他类型量子比特的原因。 同日，微软CEO萨提亚·纳德拉与主播Dwarkesh Patel的播客访谈也对此进行了讨论。在1小时17分钟 的访谈中，纳德拉分享了他对于微软在量子计算领域取得突破的感受、过程（" 这对我们来说是一个 30年的旅程。"） 和未来潜在的应用场景。此外，纳德拉还着重分享了他对于AI在法律和社会治理层 面的思考，以及AGI的认知，目前AI领域的技术突飞猛进，但 纳德拉认为AGI来临的真正标志是世 界经济增长10%。 关于DeepSeek带来的成本变化，此前纳德拉在X上提到的 ...