科技行业 | 行业追踪 比亚迪智能化战略发布会：开启 全民智驾时代 2 月 10 日晚间，比亚迪举办智能化战略发布会，发布天神之眼 C 高阶智驾以及 21 款首批搭载车型。 比亚迪发布"天神之眼"高阶智驾，包含天神之眼 A（高阶智驾三 激光版 DiPilot 600，主要搭载品牌仰望）、天神之眼 B（高阶智驾 激光版 DiPilot 300，主要搭载品牌腾势、比亚迪）、天神之眼 C（高 阶智驾三目版 DiPilot 100，主要搭载品牌比亚迪）。 比亚迪将全系搭载高阶智驾。本次发布会重点是天神之眼 C，是比 亚迪用来推动智驾向大众市场快速普及的重要技术系统，涵盖长 途、通勤、泊车等刚需场景。 天神之眼 C，在架构（璇玑架构）、传感器（5R12V12U，12 颗摄像 头、5 颗毫米波雷达、12 颗超声波雷达）、算法（BAS 3.0 + 行泊算 法）、数据（预计 2025 年年底训练里程为 1.5 亿公里/天）等四个 方面具备优势。同时，比亚迪璇玑架构全面接入 DeepSeek，赋能 车端 AI、云端 AI 双循环。这符合我们对于 DeepSeek 利于应用端 企业的基本判断。 搭载天神之眼 C 的比亚迪首批 2 ...

浦银国际研究 宏观洞察 | 宏观经济 月度中国宏观洞察：消费动能尚可， 关税战提前增加经济不确定性 中美关系：关税战提前到来，4 月 1 日或是下一个关键时点。特朗普 在上任不到月余就开始加征关税早于预期，尽管循序渐进式的加征方 式符合预期。我们估算加征关税将拖累中国 GDP 0.6 个百分点。关税战 带来的高度不确定性将继续影响中国，接下来需关注在 4 月 1 日贸易 调查截止日到来时，关税战是否会进一步升级。 1-2 月数据回顾：消费表现尚可。从需求端来看，春节消费数据——不 论是耐用品消费还是服务消费（旅游、电影票房、餐饮）均表现不俗。 房产销售在春节前大致和去年同期持平。从供给端来看，数据部分因 为春节停工因素回落。 政策展望：2 月传统政策真空期或更重政策落实，重点聚焦 3 月初两 会制定的经济目标和财政预算。尽管关税战开始时间早于我们和市场 的预期，但在关税战前景未明、且 1-2 月缺乏硬数据发布的情况下，我 们认为政府并不急于加码政策支持。近期财政政策或侧重在升级的设 备更新和消费品以旧换新政策的落实上。"适度宽松"的货币政策短 期内或继续受制于汇债两市的压力。然而，如果金融市场条件允许， 那么 ...

Investment Rating - The report maintains a "Buy" rating for Yum China (YUMC.US/9987.HK) with a target price of $61.5, representing a potential upside of 27.1% from the current price of $48.4 [2][24]. Core Insights - Same-store sales performance is identified as a key factor influencing the company's 2025 earnings and stock price. The company anticipates a slight decline in same-store sales for 2024 due to adjustments in customer pricing to maintain steady traffic. However, operational efficiency improvements are expected to offset the negative leverage impact, allowing for slight margin expansion [1][4]. - The management's cautious outlook for 2025 reflects uncertainties in same-store performance, which is expected to significantly impact restaurant profit margins. The company aims for systematic revenue growth in the mid-single digits for 2025, with a target of opening 1,600-1,800 new stores [4][6]. - The competitive landscape is observed to be stabilizing, with some industry players reducing discounts and slightly increasing menu prices. Yum China plans to maintain stable pricing for KFC while lowering prices for Pizza Hut to attract more customers [4][11]. Financial Projections - Revenue projections for Yum China are as follows: - 2023: $10,978 million - 2024: $11,303 million (3.0% growth) - 2025: $11,720 million (3.7% growth) - 2026: $12,509 million (6.7% growth) - 2027: $13,281 million (6.2% growth) [6][9]. - Net profit estimates are projected as: - 2023: $827 million - 2024: $911 million (10.2% growth) - 2025: $924 million (1.4% growth) - 2026: $996 million (7.8% growth) - 2027: $1,099 million (10.4% growth) [6][9]. Market Performance - The current market capitalization of Yum China is approximately $18,253 million, with an average daily trading volume of $102.9 million over the past three months [2][4]. - The stock has traded within a 52-week range of $28.5 to $52.0, indicating significant volatility [2][4].

浦银国际研究 公司研究 | 科技行业 业绩会要点及展望：1）2024 年，联发科旗舰手机 SoC 贡献约 20 亿 美元，超出管理层预期。2）天玑 9400 获得 OPPO 和 vivo 采用，公 司预期今年更多机型采用天玑旗舰芯片。3）今年汽车收入将逐季提 升。4）AI 趋势将为公司端侧 AI SoC 系列产品带来更多机会。 估值：我们采用 DCF（Discounted Cash Flow，现金流量贴现法）估 值方法。我们假设联发科 2030-2034 的成长率为 18%，永久增长率为 3%，WACC（Weighted Average Cost of Capital，加权平均资金成本） 是 13.1%，上调联发科目标价至 1,820.8 新台币，潜在升幅 19%，对 应 2025 年市盈率为 24.1x。 投资风险：全球经济持续疲软，智能手机、新能源车等需求不及预 期；地缘冲突导致关税上升，提前备货透支未来出货量；中国手机国 补带动需求持续性较弱；AI 智能手机渗透率上扬速度较慢，渗透中 低价格段手机速度较慢；半导体周期下行，行业竞争加剧，拖累公司 利润；研发等费用增速过快；AI 芯片迭代慢于预期。 图表 ...

Investment Rating - The report maintains a "Buy" rating for the company with a target price of $55 for US shares and HK$43 for Hong Kong shares, indicating a potential upside of 100% from the current price [5][14]. Core Insights - The company is expected to achieve product net revenues of $104 million in Q4 2024 and $394 million for the full year 2024, primarily driven by the continued growth of Aigamod [2][3]. - Aigamod is projected to generate approximately $91 million in sales for the year 2024, surpassing the previous guidance of $80 million [3]. - The company aims to achieve profitability by the end of 2025, with significant catalysts expected in 2025, including the launch of Aigamod, DLL3 ADC data readouts, and potential overseas licensing agreements [1][4]. Financial Projections - The company forecasts a net loss of $250 million for 2024, with a Q4 net loss of $63 million, while the gross margin is expected to be around 63.6% for the year [2][14]. - Revenue is projected to grow significantly from $394 million in 2024 to $538 million in 2025, with a further increase to $801 million in 2026 [15][17]. - The company anticipates a decrease in R&D expenses in 2024 compared to 2023, while sales and administrative expenses are expected to rise slightly [2][14]. Product Pipeline and Catalysts - DLL3 ADC (ZL-1310) is on track for multiple Phase 1 data readouts in 2025, which could lead to potential overseas licensing agreements [4][5]. - The company plans to submit NDAs for six additional drugs in China in 2025, including KarXT and TIVDAK, with significant clinical data readouts expected [5][13]. - Aigamod's commercialization team is expected to remain stable at around 170 personnel, supporting its growth trajectory [3][5].

浦银国际研究 公司研究 | 科技行业 高通（QCOM.US）：FY1Q25 业绩好于 预期，端侧 AI 拉动公司长期业务增长 我们略微上调 FY2025 和 FY2026 年盈利预测，调整目标价至 205.7 美 元。潜在升幅 17%，重申"买入"评级。 重申高通的"买入"评级：高通 FY1Q25 业绩及 FY2Q25 指引均略高 于市场一致预期，其手机、汽车、IoT 业务均有较好增长支撑。首先， 从今年来看，中国出台智能手机补贴政策拉动智能手机销售，尤其是 高端旗舰机型，出货量增长明显。这为高通 2025 财年提供较好的基 本盘支撑。其次，我们预期高通智能座舱以及舱驾一体的解决方案将 在今明两年持续放量。最后，高通受益于端侧 AI 在智能手机快速渗 透，估值有望提升。高通当前远期市盈率 15.1x，仍然具备吸引力。 FY1Q25 业绩强劲增长，高于指引上限；公司指引 FY2Q25 业绩稳定 成长：高通 FY1Q25 收入为 116.7 亿美元，同比增长 17%，环比增长 14%，高于指引区间上限，高于市场一致预期。其中，QCT 业务收入 达到 101 亿美元，创历史新高，主要由于安卓手机出货量、IoT 及汽 ...

Investment Rating - The report maintains a "Buy" rating for Spotify (SPOT.US) with a target price adjusted to $700, indicating a potential upside of 12% from the current price of $626 [5][6]. Core Insights - The company's overall performance exceeded expectations, with FY24 Q4 revenue reaching €4.242 billion, a year-on-year increase of 15.6%, and a gross margin improvement of 5.6 percentage points to 32.2% [3][4]. - User growth remains strong, with FY24 Q4 Monthly Active Users (MAU) increasing by 12% to 675 million, driven by emerging markets [4]. - The company has demonstrated strong pricing power, with Average Revenue Per User (ARPU) rising by 5% to €4.85, contributing to a 16.9% increase in Premium revenue [4]. Financial Performance Summary - FY24 Q4 revenue: €4.242 billion, up 15.6% YoY, exceeding market expectations by 1.9% [3]. - FY24 Q4 operating profit: €477 million, with a net profit of €367 million, achieving a record net profit margin of 8.7% [3]. - FY25E/FY26E revenue projections: €18.1 billion and €20.8 billion respectively, with adjusted EBITDA expected to be €2.736 billion and €3.552 billion [5][6]. User and Revenue Growth - FY24 Q4 MAU reached 675 million, with subscription users at 263 million, primarily driven by growth in emerging markets [4]. - Premium revenue for FY24 Q4 was €3.705 billion, reflecting a 16.9% increase YoY, supported by both user growth and ARPU increase [4]. Advertising Business Outlook - The advertising segment showed modest growth, with Ad-Supported revenue increasing by 7.2% to €537 million, although this was partially offset by pricing weakness [5]. - The company is enhancing its programmatic advertising technology and seeking more partnerships to boost ad inventory sales, with expectations for significant growth in this area by 2026 [5].

Investment Rating - The report suggests a positive outlook for the AI industry, particularly highlighting DeepSeek's potential to disrupt the market and challenge existing players like OpenAI [2]. Core Insights - DeepSeek has achieved significant cost innovation, with overall costs being only 5%-10% of OpenAI's, including training, inference, and commercialization pricing [2][12]. - The shift in the AI industry is moving from a "compute race" to "algorithm optimization," emphasizing the importance of algorithm density and ecosystem collaboration over mere resource monopolization [2][27]. Summary by Sections Cost Innovation - DeepSeek's training cost is $5.58 million, significantly lower than OpenAI's $1 billion, achieved through an efficient MoE architecture and sparse activation strategy [7][8]. - Inference costs for DeepSeek are $0.14 per million tokens for input and $0.28 for output, compared to OpenAI's $2.5 and $10 respectively, showcasing a cost advantage of 1/10 to 1/20 [9][10]. - Commercialization pricing for DeepSeek is $0.48 per million tokens, far below OpenAI's $18, making it more accessible for budget-sensitive enterprises [10][11]. Impact on AI Industry Chain - The low-cost advantage of DeepSeek is reshaping the AI industry chain, leading to a transformation in cloud computing, large models, and application layers [22][27]. - In cloud computing, the demand for inference is expected to exceed 70% of total GenAI demand by 2026, prompting hardware manufacturers to innovate [22]. - The rise of edge computing and hybrid cloud solutions is driven by DeepSeek's support for consumer-grade hardware, indicating a shift from centralized cloud to edge deployment [23]. Large Models and Open Source Ecosystem - DeepSeek's approach is breaking the "compute monopoly" and promoting technological democratization, with its open-source model achieving 1.1 million downloads within six days [24][25]. - The focus is shifting from pre-training to reinforcement learning and inference optimization, fostering the development of vertical models in sectors like finance and healthcare [25][26]. Application Layer Transformation - DeepSeek's low-cost API enables small developers to create AI applications affordably, unlocking potential in underserved markets [26]. - Traditional cloud providers are transitioning to high-margin "AI + industry solutions" models, creating new growth opportunities [26][27].

Investment Rating - The report suggests a positive outlook for the consumption industry, particularly in the context of the Spring Festival, indicating an improvement in consumer sentiment and spending [6]. Core Insights - The Spring Festival data shows a record high in travel, with over 2.3 billion people traveling, a 6% increase year-on-year, reflecting a growing consumer willingness to spend [1][5]. - The overall consumption scale during the Spring Festival continues to rise, with payment transaction numbers and amounts increasing by 25.7% and 17.1% respectively during the first seven days of the holiday [1][2]. - Domestic travel demand remains strong, with 501 million domestic travelers during the Spring Festival, a 5.9% increase compared to last year [5][6]. - The introduction of government subsidies and consumption vouchers has positively impacted consumer spending during the holiday period [2][6]. - The performance of the film industry during the Spring Festival exceeded expectations, with total box office revenue surpassing 10 billion RMB, setting a new record [1][2]. Summary by Sections Travel and Tourism - The report highlights a significant increase in travel intentions, with the Spring Festival holiday extending to eight days, leading to a rise in multi-segment travel [2][5]. - The implementation of visa-free policies and the recovery of international flight capacity have contributed to a 6.3% increase in cross-border travel compared to last year [6]. Consumer Spending - Consumer spending in retail and dining sectors showed resilience, with key retail and dining enterprises reporting a 5.4% increase in sales during the first four days of the Spring Festival [2]. - The average ticket price for movies during the Spring Festival increased to 50.8 RMB, up from 49.2 RMB in 2024, indicating a willingness to spend more on entertainment [1][2]. Investment Recommendations - The report recommends focusing on the progress of the Online Travel Agency (OTA) industry driven by cross-border travel demand and the recovery of hotel prices post-Spring Festival [6].

Core Insights - The report indicates that the U.S. has imposed a 10% tariff on imports from China, with China retaliating by imposing tariffs of 10%-15% on certain U.S. goods, which aligns with market expectations. This is not expected to cause significant market fluctuations in the short term, but the uncertainty surrounding future tariff policies suggests a structural market trend [2][5][26] - The release of DeepSeek is anticipated to narrow the valuation gap in the AI industry between China and the U.S., prompting investors to focus on stocks within the AI supply chain [2][5][26] Tariff Impact Analysis - The report highlights that the sectors most affected by the new tariffs include electronics, machinery, toys, and textiles, which are heavily reliant on imports from China. Conversely, sectors with lower dependency, such as fuel and minerals, will experience less impact [12][16] - The report also notes that the short-term market sentiment is likely to be cautious, with investors advised to avoid sectors negatively impacted by the tariffs and currency fluctuations [5][7][19] AI Industry Insights - DeepSeek's introduction is expected to accelerate the development of AI technologies by shifting the focus from merely increasing computational power to optimizing algorithms. This change is projected to foster the emergence of more efficient and cost-effective AI models [23][24] - The report suggests that the AI industry in China is likely to benefit from increased domestic production capabilities in response to U.S. restrictions on advanced technologies, which may enhance the growth of the AI supply chain in China [25][26] Market Sentiment and Investment Strategy - The report anticipates that the market will exhibit structural trends rather than significant volatility in the short term, with a recommendation for investors to be cautious and to focus on sectors that are less sensitive to currency fluctuations, such as information technology and healthcare [5][19][26] - It is suggested that the AI sector, particularly companies involved in AI chips, cloud computing, and core technologies, will attract investment interest due to the narrowing valuation gap between Chinese and U.S. AI companies [24][26]