我与 Fanatic Research、史蒂文团队及固定收益研究部门的同事合作。去年夏天我们发 布了一份报告，核心就是探讨数据中心的融资缺口问题。Steven 及其团队测算，未来四 年全球数据中心资本支出需求达 3 万亿美元。为此，我们与跨资产类别的固定收益团队合 作，思考这部分资金将如何筹集。 我们得出的结论是，优质超大规模企业能产生可观现金流，其运营现金流可覆盖约一半的 资金需求。但剩余资金缺口需依赖固定收益市场填补，其中私募信贷是主力，企业信贷和 证券化信贷也将提供支持。 发言人 1: 欢迎回到《市场观察》，也欢迎收看我们在科技、媒体与电信大会现场对话的第二部分。 我是摩根士丹利美国主题与股票策略师米歇尔·韦弗。今天我们继续与史蒂文·伯德、乔希· 贝尔和林赛·泰勒对话，本次聚焦 AI 融资及相关风险。现在是 3 月 6 日周五上午 11 点，地 点在旧金山。 昨天我们讨论了 AI 应用，尽管市场对此主题热情高涨，但也出现了一些隐忧。林赛，我想 先从融资话题开始问你。这是 AI 建设的另一个关键环节。关于数据中心融资缺口的规模， 以及信贷市场在此扮演的角色，你有哪些最新观点？ 发言人 2: 我们观察到私募 ...

Group 1: Market Sentiment and Focus - The European software sector is experiencing a sharp sell-off, leading to a search for stabilizing factors, with SAP's upcoming earnings report being a focal point for market sentiment [1] - Morgan Stanley analysts expect core software companies to report solid Q4 data despite ongoing concerns about generative AI, which has weakened investor sentiment [1][3] - Barclays previously indicated that the earnings season may not serve as a positive turning point for market sentiment, citing fears over AI's disruptive potential and unsustainable capital expenditures by large cloud service providers [1] Group 2: SAP's Performance Expectations - Morgan Stanley's data highlights that the market is focused on whether SAP can deliver on its high growth promises for cloud business, with expected cloud revenue growth of 26% to 28% year-over-year [3] - SAP's total revenue is projected to be between €33.1 billion and €33.6 billion, with non-IFRS EBIT expected to reach €10.3 billion to €10.6 billion, reflecting a year-over-year growth of 26% to 30% [3] - Morgan Stanley maintains a positive outlook on SAP, emphasizing that achieving its mid-term goals could instill much-needed confidence in the software sector [3] Group 3: Payment and Fintech Sector Dynamics - The payment and fintech sectors are showing increasing divergence, with companies like Wise demonstrating acceleration in growth, while others like Worldline face transformation risks [2][4] - Morgan Stanley is optimistic about Adyen, viewing it as a structural winner with growth acceleration potential, while Worldline is rated underweight due to execution risks in its transformation plan [4] Group 4: Performance Disparities in Tech Stocks - In the IT services and hardware sectors, significant performance disparities are noted, with Computacenter showing strong growth and exceeding profit expectations [5] - Morgan Stanley adopts a cautious stance on Oracle, citing potential revenue opportunities from GPUaaS but also highlighting risks related to infrastructure costs and financing needs [6] - The firm advises investors to focus on companies demonstrating resilient growth and not adversely affected by structural technological changes like generative AI [6]

Core Insights - The hybrid cloud strategy is expected to drive interest in specialized AI cloud services and emerging cloud providers that can support the expanding demands of AI workloads [2] - Emerging cloud providers' revenue is projected to grow by 205% year-over-year in Q2 2025, with annual revenue expected to exceed $23 billion [2] - By 2030, the revenue for emerging cloud providers is forecasted to reach nearly $180 billion, with an annual growth rate of 69% [2][7] Market Dynamics - The specialized AI cloud services segment will face intense competition by 2026, requiring emerging providers to intensify efforts to penetrate the enterprise buyer market [2] - Over 80% of enterprise cloud buyers are looking to modernize their cloud strategies, indicating a shift in purchasing preferences [2] - Emerging cloud providers can compete with large cloud service providers through aggressive pricing strategies and simplified services [2] Competitive Landscape - CoreWeave is identified as a leading competitor among emerging cloud providers, with significant contracts from major clients like Microsoft and OpenAI [4] - Microsoft has invested billions in CoreWeave's computing resources, while OpenAI has signed contracts exceeding $22 billion with the provider [4] - Other emerging cloud providers such as Vultr, FluidStack, and DataCrunch are also meeting the growing demand for AI infrastructure [4] Client Preferences - Clients engaged in AI-intensive tasks are increasingly willing to explore alternatives beyond existing cloud providers due to the high importance and cost of their work [5] - Emerging cloud providers are seen as attractive alternatives due to their value propositions, which sometimes include pricing advantages [4][5] Future Challenges - The emerging cloud market faces survival challenges as it approaches 2026, with growth dependent on enterprise market adoption [6] - To achieve broader adoption, emerging cloud providers need to adjust their marketing and distribution strategies, create more out-of-the-box functionalities, and customize service level agreements for critical business applications [6]

Financial Performance - Lenovo Group reported a revenue of 498.5 billion yuan for the fiscal year ending March 31, 2025, representing a year-on-year growth of 21.5% and a net profit increase of 36% to 10.4 billion yuan [2] - The revenue from non-PC businesses increased by nearly 5 percentage points, reaching 47% of total revenue [2] - The Intelligent Devices Group (IDG) achieved double-digit revenue growth, with a 13% year-on-year increase in the fourth quarter [2] Market Position and Trends - Lenovo's PC business accounted for 23.7% of the market share, widening the gap with the second competitor by 3.6 percentage points [2] - AI PCs represented 16% of total notebook sales in the Chinese market, with IDC forecasting a 4% year-on-year growth in PC shipments for 2025, driven by AI PC demand [2] Infrastructure and Services Growth - The Infrastructure Solutions Group (ISG) generated 104.8 billion yuan in revenue, with a 63% year-on-year increase in the fourth quarter, marking the second consecutive quarter of profitability [3] - The Solutions and Services Group (SSG) also saw double-digit revenue growth, with a nearly 20% increase in the fourth quarter and an operating profit margin exceeding 21% [3] Global Operations and Supply Chain - Lenovo has established an end-to-end global operational system, with over 30 manufacturing bases across 11 markets, enhancing its supply chain capabilities [4] - The company has adopted a "China + N" manufacturing strategy to leverage cost advantages while meeting local customer demands and mitigating high tariff barriers [4] Future Outlook - Lenovo's CEO expressed confidence in the potential for improved US-China trade relations, which could provide greater certainty for global businesses [4]

Core Insights - Microsoft has canceled and postponed data center leases totaling over 2GW in the US and Europe, indicating a surplus in data center supply [1][7][18] - The cancellation of leases has led to a decline in semiconductor stocks, with the Philadelphia Semiconductor Index dropping approximately 3.3% and Nvidia falling over 5.7% [2] - Despite Microsoft's actions, overall demand for data centers remains strong, with Google and Meta filling the void left by Microsoft [4][19] Group 1: Microsoft’s Data Center Strategy - Microsoft has decided to halt new capacity leases primarily due to the decision not to support additional OpenAI training workloads, reflecting a strategic shift in its data center operations [7][18] - The company aims to retain core market capacity for its cloud and inference workloads while canceling leases that exceed its updated mid-term capacity needs [7][18] - Microsoft plans to invest approximately $80 billion in AI data centers in the fiscal year ending in June, although growth is expected to slow thereafter [15] Group 2: Market Impact and Opportunities - The cancellation of Microsoft’s leases has created opportunities for other major data center operators, with Google and Meta experiencing significant growth in demand [8][19] - OpenAI is increasingly seeking to build its own data center capacity, with plans for multiple "Stargate" projects that could total over 6GW in long-term capacity needs [9][19] - The overall demand for data centers is still growing year-on-year, despite the supply adjustments made by Microsoft [4][18] Group 3: Equipment Orders and Supply Chain Dynamics - Data center equipment orders are experiencing a slowdown due to large-scale operators redesigning their facilities for higher rack density, which began in January [10][20] - The postponement of equipment orders is expected to negatively impact Vertiv's order volume in the first half of 2025 [10][20] - The redesign efforts are causing delays in procurement decisions, as operators await final designs before placing orders [10][20] Group 4: Future of Data Center Infrastructure - Microsoft is planning to retrofit existing data centers to support liquid cooling, which is seen as a cost-effective way to enhance cooling efficiency without overhauling entire facilities [12][21] - The push for liquid cooling is driven by the need to quickly unlock power capacity for inference workloads, indicating a shift in operational priorities among major data center operators [21]