CoreWeave股价周二收创历史新高150.48美元，涨超25%，公司股价自上市以来已暴涨248%。自5月15日公布强劲财报 以来，公司成为交易平台Public.com上交易量排名前二的股票，散户对其看涨期权的押注是看跌期权的四倍。 硬·AI 作者 | 李笑寅 编辑 | 硬 AI 从不起眼IPO到市值翻三倍，CoreWeave股价飙升。 01 AI算力业务疯狂扩张 资本支出料将暴增 CoreWeave的核心业务是基于英伟达芯片处理人工智能工作负载，其数据中心从零开始设计，专为AI计算 优化。 最新与Applied Digital的租赁协议，进一步巩固了其在美国和欧洲33个AI数据中心的布局。 作为英伟达的战略合作伙伴，CoreWeave不仅受益于其技术支持，还吸引了微软这一最大客户，以及 OpenAI的五年119亿美元长期合作。 2025年，公司预计资本支出将激增53%至215亿美元，彰显其对AI基础设施扩张的雄心。 02 散户狂欢，华尔街冷眼相待 短短两个月内，CoreWeave的市值从IPO时的230亿美元飙升至惊人的720亿美元，发行价从40美元飙至 目前的150.48美元。 尤其自5月15日公布强劲 ...

Core Viewpoint - Elon Musk's companies are rapidly initiating a series of financing activities, including a $300 million equity sale by xAI and a $5 billion debt offering, alongside Neuralink raising $650 million [2][4][8]. Group 1: xAI Financing Activities - xAI is launching a $300 million equity sale that allows employees to sell shares to new investors, with the company valued at $113 billion, consistent with its valuation during the acquisition of the social media platform X [4][6]. - The merger of xAI and X into a new entity called XAI Holdings has resulted in xAI being valued at $80 billion and X at $33 billion, following Musk's acquisition of X for $44 billion in October 2022 [4][6]. - A larger financing round is anticipated after this equity sale, with plans to raise approximately $20 billion for the newly merged AI startup and social media business, highlighting market enthusiasm for AI and Musk's influence [4][6]. - Musk has initiated a $5 billion debt sale for xAI, with Morgan Stanley seeking debt financing to support investments in AI infrastructure, particularly a large data center named Colossus in Memphis [4][5][6]. Group 2: Neuralink Financing - Neuralink has raised $650 million in funding, bringing its valuation to $9 billion, excluding this new investment [8][9]. - The previous valuation of Neuralink was $3.5 billion during a financing round in 2023 [10]. Group 3: Musk's Business Focus - Following Musk's departure from a role in the Trump administration, he appears eager to refocus on his business ventures, particularly xAI and Tesla, as he aims to concentrate on critical technology releases [12][13].

图 点击 上方 硬AI 关注我们 魏哲家表示，关税政策存在不确定性和风险等潜在影响，但截至目前，"尚未看到客户行为有任何转变"，AI需求仍强劲； 企业AI是另一个支持未来数年结构性趋势的AI需求来源；预计台积电今年先进制程销售占比将继续上升，可达七至八成。 硬·AI 作者 | 李笑寅 编辑 | 硬 AI 魏哲家上任周年首度主持股东大会，称2025年业绩料将"稳步增长"。 6月3日周二，台积电董事兼总裁魏哲家在股东大会上表示， 2025年按美元计的销售额增速预计在20%区间 中段，利润将创下记录高位。 01 关税尚未影响到客户行为 AI需求仍强劲 台积电计划在熊本县建设第二家工厂，外界普遍预计该工厂将于今年第一季度开工。 针对"台积电将赴中东 设厂"的传言，魏哲家明确表示"不会"。 他表示，之前类似与英特尔合资的传闻已经有出现过，这次是中东，下次会不会是非洲，所有的问题答案 还是一样"不会"。 魏哲家表示， 关税政策存在不确定性和风险等潜在影响，但截至目前，"尚未看到客户行为有任何转变"， 因此维持对今年全年营收的展望。 魏哲家强调，市场需求仍然强劲，预计晶圆制造将继续受益于AI的强劲需求以及其它终端市场的温和 ...

Core Viewpoint - The upcoming Apple Worldwide Developers Conference (WWDC) may not showcase Apple's commitment to AI, potentially highlighting its shortcomings in the field [2][6]. Group 1: Limited AI Openings - The most significant AI announcement at WWDC will be the opening of Apple's foundational models to third-party developers, allowing them to utilize the company's device-side technology for lightweight tasks [2][6]. - Apple's large language model has approximately 3 billion parameters, which is significantly lower than competitors like OpenAI and even Apple's own cloud-based models [2][6]. - Internally, Apple has various models with parameters ranging from 3 billion to 150 billion, with the 150 billion parameter model being cloud-dependent and more powerful for nuanced reasoning [2][4]. Group 2: Apple's AI Dilemma - Since the initial exposure to Apple Intelligence last August, the product has been perceived as more of a branding effort rather than a breakthrough innovation [6][7]. - Features like Writing Tools and Genmoji have been useful but do not match the innovation level of competitors [7]. - There are concerns that the upcoming announcements may further expose Apple's deficiencies in AI compared to rivals [7]. Group 3: Brand Strategy Over Technological Breakthroughs - This year's WWDC appears to focus more on brand rebranding rather than technological advancements, with plans to rename operating systems based on years [8][9]. - Existing features in applications like Safari and Photos are being rebranded as "AI-driven" without substantial new technology [10]. Group 4: Future Plans Require Patience - Apple aims to have a more compelling narrative in the coming years, with projects in progress including a revamped LLM Siri, redesigned Shortcuts app, an AI doctor service called "Mulberry," and a ChatGPT competitor referred to as "Knowledge" [12][13].

Core Viewpoint - NVIDIA's CEO Jensen Huang expressed concern over the H20 export restrictions impacting the company's access to the Chinese AI market, which is valued at $50 billion, while highlighting the robust demand for AI processing capabilities driven by the Blackwell chip production [1][8][45]. Group 1: Financial Performance and Market Impact - NVIDIA's Q1 revenue reached $44 billion, a 69% year-over-year increase, despite the challenges posed by export restrictions [25]. - The company anticipates a loss of $8 billion in H20 revenue due to new export limitations, significantly affecting future business prospects in the Chinese market [8][43]. - The data center revenue grew by 73% year-over-year, driven by the rapid ramp-up of the Blackwell product line [5][27]. Group 2: AI Demand and Technological Advancements - There is an explosive growth in AI inference demand, with token generation increasing by 500% year-over-year, particularly in complex AI workloads [12][29]. - The Blackwell architecture is designed to support this demand, offering a throughput that is 40 times higher than the previous Hopper architecture [12][10]. - The average deployment rate for major hyperscale customers is nearly 1,000 NVL72 racks per week, indicating strong market adoption [10][28]. Group 3: Strategic Insights on AI Market - Huang emphasized that winning the Chinese AI market is crucial for global leadership, as it houses half of the world's AI researchers [3][45]. - The company is exploring options to create attractive solutions for the Chinese market in light of the export restrictions [8][46]. - The rise of open-source AI models like DeepSeek and Qwen is seen as a strategic advantage for the U.S. in maintaining its leadership in AI technology [13][46]. Group 4: Future Outlook and Growth Engines - NVIDIA is optimistic about future growth, citing multiple key growth engines including surging inference demand, sovereign AI initiatives, and enterprise AI [19][49]. - The company plans to achieve $45 billion in revenue for Q2, with expected gross margins of 71.8% [20][43]. - The establishment of AI factories globally is seen as a foundational step in building the necessary infrastructure for AI deployment across industries [15][62].

Core Viewpoint - The collaboration between Alibaba and SAP exemplifies the power of AI in driving market capitalization growth despite macroeconomic challenges, highlighting AI as a key narrative in capital markets [2][21]. Group 1: SAP's Performance and Market Impact - SAP has become the highest-valued company in Europe, with its stock price increasing by 25% year-to-date and 60% over the past 12 months, surpassing €300 billion in market capitalization [7][10]. - The strong performance of SAP has significantly contributed to the DAX index, which has seen an 18.85% increase in 2024 and a 15.96% rise in early 2025 [10]. - Despite SAP's success, the German economy is facing challenges, with a 0.2% contraction in 2024 and stagnation expected in 2025, raising questions about the disconnect between SAP's performance and the broader economic environment [10][18]. Group 2: SAP's AI Strategy - SAP's AI strategy includes the launch of Joule, a generative AI tool aimed at enhancing user productivity by 30%, and Business Data Cloud (BDC), which integrates and manages vast amounts of data [11][12]. - The market for BDC is projected to reach $300 billion by 2028, with a compound annual growth rate of 24%, indicating strong demand for SAP's AI solutions [12]. - SAP's AI initiatives are designed to transform its business model, with expectations of converting €11 billion in support service revenue into over five times that in cloud service revenue [14]. Group 3: Market Reception and Analyst Consensus - Positive market feedback was evident at the Sapphire annual conference, with analysts expressing increased confidence in SAP's growth potential and revenue acceleration [16]. - Analysts from Morgan Stanley and Bank of America have highlighted SAP's strong product pipeline and the attractiveness of its AI narrative, supporting a favorable valuation outlook [16][21]. - SAP's global operations and AI-enhanced solutions address common challenges faced by enterprises worldwide, indicating that its growth is not solely tied to the German economy [17]. Group 4: Future Collaboration between Alibaba and SAP - The partnership between Alibaba and SAP will initially focus on technology integration and market expansion, particularly in China and Southeast Asia [20]. - This collaboration aims to create tailored, scalable, and secure AI solutions for regional markets, potentially accelerating the AI adoption in traditional industries [20]. - The alliance reflects a broader trend of cooperation in AI development, which is becoming essential for sharing innovation risks in a complex technological landscape [21].

Core Viewpoint - Anthropic's release of the Claude 4 series models marks a new era in AI capabilities, particularly in programming, potentially reshaping the software development industry landscape [4][17]. Group 1: Model Capabilities - Claude Opus 4 is touted as the "best programming model globally," capable of maintaining stable performance over long tasks requiring focus and effort, verified by Rakuten's 7-hour continuous operation [3][8]. - Claude Sonnet 4 shows a significant accuracy improvement, achieving 72.7% in the SWE-bench test compared to Sonnet 3.7's 62.3% [5][6]. - Both models utilize a hybrid design, allowing for immediate responses and deeper reasoning, enhancing their utility in complex coding and problem-solving scenarios [5][9]. Group 2: Extended Functionality - The new models introduce "extended thinking and tool usage," enabling Claude to utilize web searches and other tools during reasoning, improving response accuracy [11]. - Opus 4 significantly enhances memory capabilities, allowing it to create and maintain "memory files" when granted local file access, improving long-term task awareness and coherence [11][12]. Group 3: Product Launch and Integration - Claude Code has officially launched, receiving positive feedback during testing, and integrates seamlessly with platforms like GitHub Actions, VS Code, and JetBrains [12][13]. - The pricing structure remains consistent with previous models, with Opus 4 charging $15 and $75 per million tokens for input and output, respectively, and Sonnet 4 charging $3 and $15 [6]. Group 4: Competitive Landscape - The release of Claude 4 series intensifies competition among AI giants, with recent announcements from Microsoft, Google, and OpenAI highlighting the race for leading AI models [15]. - Investors are encouraged to reassess the competitive landscape, particularly Anthropic's position relative to OpenAI and Google, as the capabilities of the Claude 4 series may provide opportunities for increased market share [17].

Core Viewpoint - OpenAI is expanding its global AI infrastructure by constructing the largest AI data center in Abu Dhabi, UAE, as part of its "Stargate" initiative, in collaboration with local AI company G42 [2][5]. Group 1: Project Scale and Details - The Abu Dhabi data center will have a total capacity of 5GW and cover approximately 10 square miles, with an energy demand equivalent to that of five nuclear power plants [5]. - OpenAI will be the primary tenant of this data center, planning to utilize 1GW of computing resources, with the first phase involving a smaller 1GW cluster expected to begin operations in 2026 [5][7]. - This project significantly surpasses OpenAI's initial "Stargate" project in Abilene, Texas, which has an estimated total power usage of only 1.2GW [5]. Group 2: Mutual Investment and Collaboration - The partnership with G42 includes a mutual investment clause, where G42 commits to equal investment in AI infrastructure in the United States, fostering technology exchange between the Middle East and the U.S. [10]. - The UAE government plans to provide free ChatGPT Plus subscriptions to its citizens and integrate ChatGPT into various government services, including energy, healthcare, and administrative functions [10]. Group 3: Expansion into Asia-Pacific - OpenAI is also looking to expand into the Asia-Pacific region, with plans to visit countries such as Japan, South Korea, Australia, India, and Singapore to discuss potential collaborations on AI infrastructure and software applications [12].

Core Viewpoint - Nvidia is collaborating with Navitas Semiconductor to develop a next-generation 800V high-voltage direct current (HVDC) architecture, which is expected to significantly enhance the power supply systems for AI data centers, particularly for supporting GPU workloads like Rubin Ultra [3][4]. Group 1: Collaboration and Technology - The partnership aims to leverage Navitas's gallium nitride (GaN) and silicon carbide (SiC) technologies to improve energy efficiency and reduce copper usage in data centers [4]. - The new 800V HVDC architecture represents a major technological leap in data center infrastructure, addressing the increasing power demands of AI computing [4][6]. Group 2: Current Limitations and Innovations - Current data center architectures utilize traditional 54V rack power distribution systems, which are limited to several hundred kilowatts and face physical constraints when power exceeds 200kW [6]. - Nvidia's solution involves converting 13.8kV AC grid power directly to 800V DC, eliminating multiple conversion steps to maximize efficiency and reliability [6]. Group 3: Benefits of 800V HVDC - The higher voltage level of 800V HVDC allows for a reduction in copper wire thickness by up to 45%, addressing sustainability concerns for next-generation AI data centers [7]. - Using traditional 54V systems, supplying power to a 1MW rack requires over 200kg of copper, which is unsustainable for the gigawatt-level power needs of future AI data centers [7].

Core Viewpoint - OpenAI's acquisition of the AI hardware startup io, co-founded by Jony Ive, marks its largest acquisition to date, valued at approximately $6.5 billion, with a focus on developing consumer-centric AI-driven devices [2][6]. Group 1: Acquisition Details - OpenAI announced the acquisition of io for a total valuation of nearly $6.5 billion, which includes $5 billion for equity and the remainder from a previous partnership that granted OpenAI a 23% stake in io [2]. - The acquisition is expected to be completed in the summer of 2023, pending regulatory approval [2]. - The deal will bring around 55 hardware engineers, software developers, and manufacturing experts to OpenAI, forming a dedicated team for AI-driven device development [2][3]. Group 2: Leadership and Team Structure - Jony Ive and his design firm LoveFrom will not join OpenAI but will oversee the creative and design aspects, including software development [3]. - The io team will merge with OpenAI's research, engineering, and product teams in San Francisco to enhance collaboration [3]. Group 3: Future Product Expectations - The first device resulting from the acquisition is anticipated to debut in 2026, with Ive and Altman having explored initial concepts over the past two years [8]. - Both Ive and Altman aim to create a novel product that transcends traditional screen experiences, addressing consumer desires for innovative technology [8]. - While the new device is not expected to replace smartphones, it will represent a new type of interaction with AI, moving beyond current device limitations [9].