本文档仅供上海信鱼私募基金管理有限公司18860455898研究使用，请勿外传 ab 7 January 2025 Powered by YES UBS Evidence Lab Global Research U.S. Software Latest Enterprise AI Survey Summary In this third iteration of our enterprise AI survey, we again leveraged the UBS Evidence Lab data team to survey IT execs at 125 organizations about their AI plans at the chip, cloud infrastructure, model, application software, data and security layers, with a focus on incremental changes in feedback from our previous May 2024 survey (see report here). T ...

Summary of Conference Call Notes Industry Overview - The PCB industry has shown strong performance recently, particularly driven by the demand for high-density interconnect (HDI) boards in AI servers. The compound annual growth rate (CAGR) for HDI boards is 16.3% [3][4][8] - China holds a significant share of the global PCB market, accounting for approximately 40%-60% [3][4][8] Key Insights - The upcoming release of NVIDIA's 300 series products in Q1 2024 is expected to boost the demand for related PCB manufacturers [3][4] - Current market sentiment is weak, with limited available capital and a trend of hot money flowing overseas, particularly into ETFs like the Nikkei ETF, rather than returning to the A-share market [4][5] - The domestic AI computing infrastructure is continuously advancing, with HDI boards becoming a crucial increment in AI servers, significantly increasing their value [4][6] - New storage modules, such as RPCAMM, are gradually replacing traditional memory modules, reducing energy consumption by approximately 60% [4][6] Stock Analysis - Stock performance analysis should focus on moving averages, volatility, and top-bottom signal patterns, combined with changes in trading volume to identify turning points [4][7] - Flagship stocks exhibit clear upward trends along specific volatility lines, which can serve as a reference for analysis [7] Market Dynamics - The PCB industry has several promising sub-sectors, including HDI boards, PCB products related to AI servers, and the optical communication field, all of which have high technical difficulty and added value [4][8] - Observing the flow of funds between sectors is crucial for understanding market rotation, with recent trends showing a shift from high-valued stocks to undervalued or potential growth opportunities [4][9] Future Opportunities - The edge and endpoint chip market is expected to grow rapidly by 2025, driven by increased demand for edge computing-specific chips and endpoint control chips [4][10] - The use of camera chips in emerging applications, such as robotics, is also anticipated to rise [10][11] Additional Insights - The pet economy sector is experiencing steady growth, although at a slower pace, with stable prospects in pet feed and veterinary medicine [4][12] - Investment strategies should adhere to the principle of buying low and selling high, ensuring that all research and strategies align with this goal [4][13] - There are potential recovery opportunities for pharmaceutical stocks this year, particularly if driven by significant products [4][14] - The lithium battery industry has shown strong recent performance, with signs of stabilization and an overall positive outlook [4][15] - The optical communication sector is expected to benefit from AI-driven demand, particularly from cloud computing, although high valuations and uncertain performance outlooks present challenges [4][16] - The Beijing Stock Exchange has shown positive sentiment and potential for further upward movement, with a solid structure and capital returning from the main board and STAR Market [4][17]

Summary of Key Points from the Conference Call Industry Overview - The conference call primarily discusses the **computer industry** in China, focusing on market performance, challenges, and opportunities for growth in 2024 and beyond [3][4][5]. Market Performance - The **computer sector** experienced a volatile market in 2024, characterized by initial declines followed by a strong rebound in the fourth quarter, reflecting a mismatch between expectations and reality [4][3]. - The sector was ranked last among 30 industries during the downward phase, indicating weak performance [4]. Core Insights - The **computer industry** is in a phase of weak improvement, with policy effects expected to materialize more significantly by the second quarter of 2025 [4][5]. - Ongoing **U.S. sanctions** against Chinese tech companies, including a combination of entity lists and tariffs, continue to impact procurement and sales [4][6]. - **Tencent** has been added to the U.S. military-related entity list, raising concerns about foreign capital withdrawal, although there are examples of companies successfully appealing such designations [4][7]. - The progress of **domestic substitution** varies across sectors, with hardware advancements being quicker than in core areas like chips and operating systems [4][9]. - The **信创 (Xinchuang)** software sector is expected to see growth due to debt restructuring and subsidy injections in late 2024 [4][14]. Challenges and Risks - The **U.S. sanctions** pose a multi-dimensional threat to Chinese tech firms, affecting both procurement and sales channels [6][8]. - The **domestic tech industry** faces challenges in achieving self-sufficiency, particularly in critical areas like chips and operating systems [9][10]. - The **local debt crisis** has impacted government spending on信创 software, but improvements are anticipated with upcoming financial support [12][14]. Future Trends - The **AI large model** sector is seeing rapid performance realization in upstream areas like computing power and optical modules, with significant potential in B2B applications, especially in finance and transportation [4][20]. - The **industrial software** market has substantial growth potential, with a significant gap in investment compared to global standards [15]. - The **B2B application** landscape is viewed as more promising than B2C in the Chinese market due to the complexities and barriers faced by large corporations in the B2B space [21]. Investment Opportunities - Key areas for investment include **信创 software** and **industrial software**, which are expected to see increased productization and application [30]. - The **semiconductor-related industrial software** is highlighted as having strong potential due to high downstream demand and ongoing upgrades [16]. - The **outbound revenue** from domestic companies is primarily derived from the U.S. and Europe, with potential growth in the Belt and Road countries [19]. Conclusion - The **computer industry** in China is navigating a complex landscape of challenges and opportunities, with significant implications for investment strategies moving forward. The focus on domestic substitution, AI applications, and industrial software development presents potential avenues for growth and resilience in the face of external pressures [4][30].

Summary of Conference Call Notes Industry or Company Involved - The discussion primarily revolves around the **technology sector**, particularly focusing on **consumer electronics**, **AI glasses**, and **humanoid robots**. The context also includes insights into the **stock market** performance and trends. Core Points and Arguments 1. The overall market has seen a slight decline of approximately **0.6%**, but over **2,700 stocks** have risen, indicating a structural recovery in the market despite some weak performances in certain sectors like **dividend stocks** and **energy** [1][2][3] 2. The **dividend-style assets** have been leading the decline, suggesting a shift in market sentiment and a potential end to the current downtrend cycle, which may not be entirely negative [2][3] 3. **Technology sectors** such as **semiconductors**, **AR glasses**, and **information security** are showing strong performance, indicating a rotation towards more elastic sectors [3][4] 4. The **CES 2025** event is highlighted as a significant catalyst for the consumer electronics industry, showcasing innovations and trends that could impact stock performance in the tech sector [4][5][16] 5. The **AI glasses** market is gaining traction, with many tech giants investing heavily, although the final product form is still under exploration [7][8][35] 6. The **humanoid robot** sector is experiencing significant interest, with companies like **Tesla** and **Huawei** making notable advancements, indicating a growing market potential [31][34][35] 7. The **spring festival effect** is anticipated to influence market activity, with expectations of increased trading volume as the holiday approaches [9][10][37] 8. The **consumer electronics** sector is expected to benefit from recent innovations and product launches, particularly in AI integration, which is seen as a major growth area [5][11][12] 9. The **PPR (Producer Price Index)** data indicates a stable inflation environment, with a slight increase in prices, suggesting a potential shift towards a more favorable economic outlook [36][37] Other Important but Possibly Overlooked Content 1. The **dividend stocks** and **energy sectors** are currently underperforming, which may signal a broader market correction or a shift in investor focus [1][2] 2. The **performance of the PCB (Printed Circuit Board)** sector is closely tied to advancements in technology, particularly in 5G and consumer electronics, indicating its critical role in the tech supply chain [12][19] 3. The **humanoid robot** market is not only driven by tech companies but also by traditional industries exploring automation, highlighting a cross-sector interest in robotics [31][34] 4. The **CES** event is described as a "spring festival" for the tech industry, emphasizing its importance in setting trends and expectations for the year ahead [16][17] 5. The **AI glasses** market is characterized by high investment activity, with products already entering the consumer market, indicating a rapid development cycle [8][35] This summary encapsulates the key insights and trends discussed in the conference call, providing a comprehensive overview of the current state and future outlook of the technology sector and related industries.

Summary of US Enterprise Hardware and Networking Conference Call Industry Overview - The focus is on the US Enterprise Hardware and Networking sector, particularly in relation to AI infrastructure investments for 2025 and beyond [2][3][11]. Key Companies Mentioned - **Dell Technologies (DELL)**: Rated as a "Buy" with a price target of $158, expected to benefit significantly from the AI server market and projected to achieve at least 10% revenue growth in ISG for 2025 [6][16]. - **TD SYNNEX (SNX)**: Rated as a "Buy" with a price target of $150, anticipated to see 5% growth in billings and return 50% of free cash flow to shareholders [6][16]. - **Arista Networks**: Neutral rating, expected to see conservative revenue growth due to its exposure to hyperscalers [4][8]. - **Celestica**: Neutral rating, also expected to have conservative estimates due to significant exposure to major tech companies [4][8]. - **Cisco**: Mentioned as a vendor benefiting from AI-related investments [3][8]. Core Insights and Arguments - **AI Infrastructure Investments**: Strong investments in AI infrastructure are expected to continue into 2025 and the first half of 2026, with data center capital expenditures projected to increase by approximately 33% in 2025 and 23% in 2026 [3][8]. - **Ethernet Switch Market Growth**: The market for AI back-end ethernet switches is expected to grow from approximately $1 billion in 2024 to $3.4 billion in 2025, driven by increased demand for AI workloads [3][39]. - **Revenue and Earnings Revisions**: Unlike 2024, where AI stocks benefited from both revenue growth and multiple expansions, 2025 is expected to focus more on "beat and raise" quarters due to elevated valuations [4][12]. - **Market Sentiment**: The market is currently pricing in strong investments in AI infrastructure, particularly among hyperscalers, while traditional tech companies are experiencing more modest valuations [8][12]. Financial Projections - **Dell**: Projected free cash flow growth of 15% in 2025 and low double digits in 2026, with shares trading at 12% of CY26E free cash flow [6][16]. - **TD SYNNEX**: Expected to achieve roughly 10% EPS growth, with its AI infrastructure business projected to grow to several billion dollars in revenue [6][16]. - **CDW**: Revenue and EPS estimates have been trimmed due to macroeconomic headwinds, with a new price target of $220 [5][7]. Potential Risks - **Valuation Concerns**: Elevated valuations of AI-exposed companies may limit further multiple expansions, making stock selection critical for 2025 [12][30]. - **Consumer Demand for AI Products**: Limited consumer interest in generative AI features for smartphones, particularly for Apple, may hinder growth expectations [8][46]. Other Important Insights - **AI Back-End vs. Front-End Investments**: The distinction between back-end and front-end AI investments is crucial, with back-end spending expected to grow significantly while front-end investments are also anticipated to increase [41][42]. - **Market Dynamics**: The market has shifted focus to secondary AI-related companies, with significant price appreciation observed in companies like Nvidia and Dell, while traditional tech companies face de-rating risks [22][23]. This summary encapsulates the key points from the conference call, highlighting the expected trends, financial projections, and potential risks within the US Enterprise Hardware and Networking sector as it relates to AI investments.

Summary of Key Points from the Conference Call Industry Overview - The focus is on the **AI industry**, particularly the evolution and adoption of AI technologies since the launch of ChatGPT over two years ago, indicating that the industry is still in its early stages of diffusion [1][2][4]. Core Insights and Arguments 1. **Rate of Change in AI**: The rate of change in AI capabilities continues to surprise analysts, with significant adjustments made to the AI exposure and materiality of over 585 stocks, representing a market cap of $13 trillion [2][27]. 2. **Outperformance of AI Stocks**: Stocks that have seen increased AI exposure and materiality have outperformed the market, with a projected 29% upside to price targets for overweight-rated stocks identified in the latest survey [3][27]. 3. **Agentic AI**: The year 2025 is anticipated to be pivotal for **Agentic AI**, where software will transition from reactive to proactive functionalities, leading to enhanced productivity and broader enterprise adoption [4][52]. 4. **Sector Performance**: The semiconductor sector's outperformance is expected to give way to software, with a notable shift in market dynamics as AI technologies become more integrated into enterprise solutions [4][53]. Important but Overlooked Content 1. **Financial Sector's AI Adoption**: The financial sector has seen the highest net increase in AI materiality, with 17% of financial stocks now considering AI as significantly important to their investment thesis [27][69]. 2. **Pricing Power**: Companies classified as Adopters with high pricing power have outperformed those with low pricing power by approximately 30% since the launch of ChatGPT, indicating the importance of pricing strategy in leveraging AI benefits [65][66]. 3. **Global AI Adoption Trends**: The U.S. leads in generative AI adoption, but there is potential for other regions, particularly Europe and Asia, to catch up, especially among moderate adopters [40][41][42]. 4. **Future Projections**: Gartner predicts that by 2028, 33% of enterprise software applications will incorporate agentic AI, highlighting a significant shift in how software interacts with users and processes tasks [52]. Conclusion - The AI industry is poised for substantial growth and transformation, particularly with the rise of Agentic AI. Investors are encouraged to focus on stocks with high materiality and pricing power, as these are likely to yield significant returns in the evolving landscape of AI technologies [63][65][86].

Summary of Key Points from the Conference Call Industry Overview - The focus for 2025 is on the growth of AI infrastructure software, with a noted rotation from AI hardware to AI software infrastructure already in progress. The transition to AI applications is expected to occur in 2026 [1][3] - The technology sector is experiencing a shift, with early cycle winners being replaced by software companies. Notably, NVDA has remained flat since June, while MSFT has decreased by approximately 9%, contrasting with the IGV index which has increased by about 30% since summer lows [3] Company-Specific Insights Top Picks for 2025 - **Snowflake (SNOW)**: Expected to benefit from AI as a significant tailwind, with a projected durable growth rate of over 20% due to the ongoing shift to cloud services [3] - **Datadog (DDOG)**: Positioned to sustain over 20% revenue growth, leveraging its observability software as enterprises migrate to the cloud and adopt new AI tools [3] - **JFrog (FROG)**: Anticipated to see growth from its security add-on and GitHub integration, with a favorable setup for 2025 as AI tools boost code creation [6] - **Commvault (CVLT)**: Despite outperforming the IGV by approximately 66% in 2024, it trades at a 20-30% discount compared to similar SaaS peers. Expected to achieve mid-to-high teens organic ARR growth through FY27 [3] - **Zeta Global (ZETA)**: Currently trading at less than 4x 2025 revenue, despite showing over 20% growth for 13 consecutive quarters. Positioned to benefit from upselling opportunities and the legacy marketing cloud replacement cycle [6] Growth Projections and Financial Metrics - **CRM's 2025 organic growth**: Any growth exceeding the current 9% consensus is considered AI-driven, requiring an incremental $1 billion to achieve over 11% organic growth [3] - **Datadog's AI-native customers**: Approximately 6% of ARR comes from AI-native customers, which may present near-term optimization risks but is expected to provide long-term growth support [6] - **Phreesia**: Expected to achieve 14% revenue growth in 2025, with a market valuation significantly below its peers [6] Market Dynamics - The transition to AI applications is viewed as premature due to challenges such as data structure, governance, privacy, and talent shortages [3] - The overall sentiment is that AI infrastructure providers will perform best in 2025, as they address the data needs essential for successful AI tool deployment [3] Additional Insights - The report includes a recap of coverage performance in 2024 and aims to expand coverage from 65 to 75 software companies in the upcoming year [1] - The analysis highlights the importance of company-specific growth stories that may be misunderstood or underestimated by the market [3] This summary encapsulates the critical insights and projections from the conference call, focusing on the technology sector's dynamics and specific company performances expected in 2025.

Summary of Key Points from the Conference Call Industry Overview - The conference call is part of the "Data Centers & AI: Powering the Future Series 7" focusing on the utilities and clean energy sector, particularly the evolving power demand for AI and data centers [1][3]. Core Insights and Arguments - **Power Demand Evolution**: Discussion on how AI and data centers' power demand has evolved in 2024 and expectations for 2025/26 [9]. - **Nuclear Power Viability**: Examination of the role of nuclear power in supporting AI and data centers, including potential risks for big tech companies in expanding nuclear capacity [9]. - **Gas Power Purchase Agreements (PPAs)**: Consideration of behind-the-meter or virtual PPA deals for gas power plants with data centers, including who would bear commodity risks [9]. - **Net Zero Targets**: Analysis of how rising power consumption and emissions could impact big tech's net zero targets [9]. - **Renewable Energy Contribution**: Exploration of the potential for wind and solar energy to play a significant role in powering AI [9]. - **Emerging Power Solutions**: Discussion on other emerging powering and energy efficiency solutions for AI and data centers [9]. - **Political Debate**: Overview of the political debate in the US regarding the impact of AI and data center demand on grids and electricity bills [9]. Additional Important Content - **Speaker Background**: Brian Janous, Chief Strategy Officer and Co-Founder of Cloverleaf Infrastructure, has extensive experience in energy and decarbonization, previously serving as VP of Energy at Microsoft [5][10]. - **Consultancy Role**: Cloverleaf Infrastructure focuses on grid expansion to support large electric customer loads, which is critical for economic growth and investment [5]. - **Call Registration**: The call is hosted by a team of equity analysts specializing in utilities and clean energy, indicating a collaborative approach to research and analysis [2]. This summary encapsulates the key points discussed in the conference call, highlighting the critical issues facing the utilities and clean energy sector in relation to AI and data centers.

Summary of Key Points from Conference Call Industry Overview - The conference call primarily discusses the **diesel generator market** and **electric equipment industry** with a focus on supply chain challenges and technological advancements [2][4]. Core Insights and Arguments - **Supply Chain Bottlenecks**: The diesel generator market is experiencing supply chain bottlenecks, particularly in engines and generators. Companies like Cummins and Caterpillar are facing tight business conditions and increasing demand, as evidenced by Cummins raising revenue guidance for two engine-related business segments over the past four quarters, with their 95 series engines sold out for 2025 [2][4]. - **Technological Innovations**: Key areas of technological change include APP-related applications, system integration, and modularization. The impact of lithium batteries replacing lead-acid batteries is noted, although the overall effect is considered minor. Energy storage applications are currently less significant domestically but are gaining attention overseas, especially in the context of green computing centers [3][4]. - **Emerging Market Trends**: In early December, new trends emerged, including an increase in hyper-converged BB segments and heightened cooling demands in air conditioning systems due to increased equipment density. Additionally, overseas large data centers are experiencing unexpected growth in gas energy demand [4][5]. - **Power Supply Challenges**: The electric power supply sector is facing challenges as the density of base stations and the scale of data centers increase. The existing 10 kV power supply systems are becoming bottlenecks, necessitating upgrades to higher voltage levels, which will impact the performance of related equipment suppliers like Wanda [4][9]. - **High Voltage Direct Current (HTC) Technology**: HTC technology has advantages in data center power supply, but domestic penetration rates are declining, and overseas markets remain low. However, future development potential is significant, with companies like Nokia and Ericsson actively promoting related solutions [4][11][13]. - **Lithium Battery Advantages**: Lithium batteries are gradually replacing lead-acid batteries due to their reliability, compact size, and fast response times. Companies like Yiwei Lithium Energy and Panda Group are expected to benefit from this trend [4][17]. - **Panama Power Solution**: The introduction of the Panama power solution has increased the average transaction value for system integrators, although it may lead to a decline in profit margins. System integrators need to focus on cost control and technological upgrades [4][12][25]. Other Important Insights - **Company Developments**: Companies like He Wang and Sheng Hong are highlighted for their positive growth prospects, with He Wang entering the overseas wind energy market and Sheng Hong achieving significant profitability in charging and energy storage sectors [19][20]. - **Market Positioning**: Zhongheng is noted for its leading position in the domestic market, with a market share that once approached 50%. Despite not releasing typical reports recently, it remains dominant in HUDC research [24]. - **Challenges in Power Line Construction**: The construction of power lines involves complex processes that require professional teams and can be time-consuming, especially overseas. Standardizing and modularizing these processes could significantly reduce costs and improve efficiency [27]. This summary encapsulates the key points discussed in the conference call, providing insights into the current state and future outlook of the diesel generator and electric equipment industries.

Summary of Conference Call Industry Overview - The conference discussed the developments in the computing power industry, particularly in the context of AI and large models, highlighting the recent CES event where new products were launched by companies like Dongyida [1][2]. Key Points and Arguments - **Investment in Computing Power**: The capital expenditure in computing power is not a sudden increase but a gradual process influenced by advancements in technology and infrastructure, particularly in large model training [2]. - **Model Parameter Growth**: There is a consensus that the ideal parameter size for models has increased significantly, with current expectations suggesting that models should exceed 50 trillion parameters, with OpenAI aiming for 100 trillion parameters by 2025 or 2026 [3][5][24]. - **Data and Model Relationship**: The relationship between model parameters and data volume is complex, with ongoing research indicating that even synthetic data can contribute to model performance [4]. - **Market Predictions**: The market for AI applications is expected to grow significantly, with optimistic estimates suggesting a doubling of user engagement and revenue in the coming year [6]. - **Domestic vs. Foreign Technology**: There is a notable shift towards domestic computing solutions, with companies like Huawei making strides in performance comparable to leading foreign products [7][31]. - **AI and Data Security**: Ensuring data security and privacy remains a challenge, with both technological and regulatory measures being necessary to protect user data [11][12]. - **Environmental Concerns**: The increase in computing power and AI applications raises concerns about energy consumption and environmental impact, although advancements in efficiency are being made [13][14]. - **Cost of AI Implementation**: The cost of AI inference has significantly decreased, making it more accessible for businesses, with predictions of further reductions in the near future [15][16]. - **Investment Distribution**: Currently, around 60-70% of AI-related capital expenditure is allocated to hardware, particularly chips and servers, with the remainder going to software and other related technologies [16][17]. Additional Important Insights - **Future of AI Models**: The future of AI models is expected to focus on the integration of AI with various applications, including robotics and smart devices, which could open new markets [26][27]. - **Market Dynamics**: The competition in the AI space is intensifying, with companies needing to adapt quickly to maintain relevance and market share [29][30]. - **Regulatory Landscape**: The regulatory environment surrounding AI and data privacy is evolving, with ongoing debates about the implications of AI on traditional industries [12][13]. This summary encapsulates the key discussions and insights from the conference call, providing a comprehensive overview of the current state and future outlook of the computing power and AI industry.