Core Viewpoint - Nvidia and Microsoft are both approaching a $4 trillion market valuation, but they represent fundamentally different AI investment logics: Nvidia focuses on direct bets on core infrastructure, while Microsoft emphasizes long-term belief in the proliferation of application ecosystems [2][4]. Group 1: Nvidia's Position - Nvidia's value surge is attributed to its unique position in the AI value chain, acting as a "supplier" where any company looking to enter the AI field must first procure its chips, leading to explosive growth with annual sales increasing over tenfold in the past three years [4][9]. - Analysts predict Nvidia's average annual growth rate will remain at 32% over the next three years, although its growth is contingent on the demand from its largest customers and the potential emergence of disruptive technologies [9][10]. Group 2: Microsoft's Strategy - Microsoft plays the role of a "service provider," betting on deeply integrating AI technology into its extensive product matrix, such as Azure cloud services and Office software, and convincing users to pay a premium for these AI-enhanced services [4][5]. - Microsoft's market capitalization increased by $1 trillion in less than three months, but achieving a $4 trillion valuation would result in the highest expected earnings multiple in over 20 years, indicating that the market has high expectations for its future performance [2][6]. Group 3: Challenges Facing Microsoft - Microsoft faces significant challenges, including a potential rift with early partner OpenAI, which may restructure their relationship as OpenAI seeks to become a standard profit-driven company [7]. - The company is also struggling to reduce its dependency on Nvidia, encountering difficulties in developing its own AI chips, while undergoing large-scale internal restructuring, including layoffs aimed at increasing efficiency and investment in AI [7][8]. - Financially, AI's contribution to Microsoft remains limited, with AI services in the Azure cloud generating approximately $11.5 billion, which is only about 4% of total sales, providing some downside protection [9].

Group 1 - Meta is shifting its strategy towards pursuing "superintelligence" after acquiring Scale AI, focusing on attracting top talent like Alexandr Wang and his team [2][22][23] - Apple is struggling to attract top AI researchers due to its conservative culture, lack of competitive salaries, and insufficient computational resources compared to companies like Meta and OpenAI [2][63][72] - OpenAI faces significant risks regarding Microsoft's control over intellectual property (IP), which could create uncertainty for its researchers [3][37][44] Group 2 - Nvidia has built a strong moat with its hardware and software ecosystem but has made strategic missteps, such as competing directly with cloud service providers after acquiring Lepton [3][41] - Device-side AI is viewed pessimistically, as consumers are more price-sensitive and prefer free cloud-based AI services over paid device-side solutions [3][73] - The future of AI is expected to see more complex and valuable functionalities relying on cloud infrastructure rather than device-side capabilities [3][66] Group 3 - The competition for "superintelligence" is led by OpenAI, with Meta closely following, as both companies are expected to attract sufficient top talent to remain competitive [4][29] - The AI landscape is characterized by a lack of differentiation among companies, as many are utilizing similar underlying technologies and methods [4][60] - The ultimate goal of AI is to reduce human working hours, with the initial impact expected to be on white-collar creative jobs rather than manual labor [4][73]

Core View - CoreWeave has received the first market deployment of AI server systems based on NVIDIA's latest high-end chips, highlighting its significant position in providing cloud computing capabilities for AI software training and operation [2][4]. Group 1: Company Developments - CoreWeave's GB300 NVL72 server rack represents NVIDIA's latest generation of AI chip technology, manufactured by Dell Technologies [2]. - The new system is expected to enable clients, including OpenAI, to develop and deploy larger and more complex AI models at unprecedented speeds [2][4]. - Following the announcement, CoreWeave's stock price increased by nearly 9% [1][2]. Group 2: Industry Trends - The demand for AI infrastructure remains strong, with major cloud computing companies competing to establish AI capabilities [3][4]. - Dell is revitalizing its business by focusing on selling AI servers to companies, including Musk's xAI, reflecting the ongoing growth in the tech industry's demand for AI infrastructure [4]. - Despite the strong demand for AI infrastructure, the launch of new chips occurs amid uncertainties in chip trade policies and the broader global economic environment [4][5].

Core Insights - DeepSeek is actively seeking to attract overseas talent by posting 10 job openings on LinkedIn, including three positions focused on Artificial General Intelligence (AGI) [1][2] - The job postings are significant as they are the first from DeepSeek on LinkedIn in several months, indicating a strategic move to tap into the global talent pool [2] - The positions are located in Beijing and Hangzhou, with all descriptions provided in Chinese, suggesting a focus on local language despite targeting international candidates [2] Industry Context - The global competition for AI talent is intensifying, with companies like OpenAI and Meta also vying for top talent [2][3] - U.S. AI firms are employing various strategies to attract talent, including workforce reductions to reallocate funds for investments in data centers and offering high salaries to top engineers [3] - This talent war highlights the fierce competition within the AI industry to control transformative technologies and maintain a competitive edge in a rapidly evolving field [3]

Core Viewpoint - Microsoft is adjusting its internal AI chip development strategy to focus on less aggressive designs by 2028, aiming to overcome delays in development while maintaining competitiveness against Nvidia [2][4]. Group 1: Development Delays and Strategic Adjustments - Microsoft has faced challenges in developing its second and third-generation AI chips, leading to a strategic shift towards more pragmatic and iterative designs [2][4]. - The Maia 200 chip's release has been postponed from 2025 to 2026, while the new Maia 280 chip is expected to provide a 20% to 30% performance advantage per watt over Nvidia's 2027 chip [2][4][5]. - The company acknowledges that designing a new high-performance chip from scratch each year is not feasible, prompting a reduction in design complexity and an extension of development timelines [2][5]. Group 2: Chip Development Timeline - The Braga chip's design was completed six months late, raising concerns about the competitiveness of future chips against Nvidia [5]. - A new intermediate chip, Maia 280, is being considered for release in 2027, which will be based on the Braga design and consist of multiple Braga chips working together [5][6]. - The Maia 400 chip, initially known as Braga-R, is now expected to enter mass production in 2028, featuring advanced integration technologies for improved performance [6][7]. Group 3: Impact on Partners - The revised roadmap has negatively impacted Marvell, a chip design company involved in the Braga-R project, leading to a decline in its stock price due to project delays and economic factors [9]. - Not all of Microsoft's chip projects are facing issues; CPU projects, which are less complex than AI chips, are progressing well [9][10]. - Microsoft's Cobalt CPU chip, released in 2024, is already generating revenue and is being used internally and by Azure cloud customers [10].

作者 | 龙 玥 编辑 | 硬 AI Meta正以前所未有的薪酬水平在AI领域展开人才争夺战。 扎克伯格为AI顶尖人才开出高达4年3亿美元、首年超1亿美元的天价薪酬包，并向OpenAI员工发出至少十份类似邀约， 引发业内震动。OpenAI高层震怒，感觉"就像有人砸了我们的家偷走了东西"，并紧急调整薪酬应对。不过，Meta发言人 否认相关报道称"这些说法不实"。 硬·AI 据《连线》（WIRED）杂志最新获悉，马克·扎克伯格在为其新成立的Meta超级智能实验室招兵买马之际， 向顶尖研究人才开出了高达3亿美元（约合人民币21.8亿元）的四年期薪酬方案，其中首年总薪酬超过1亿 美元（约合人民币7.27亿元）。 知情人士透露，Meta已向OpenAI员工发出至少十份令人咋舌的天价邀约。Meta曾试图挖角一名高级研究 员担任首席科学家一职，但遭到拒绝。消息人士表示，虽然薪酬方案包含股权，但首年的股票可立即归属 （即立即兑现）。 "这个报价才可能打动我去Meta工作，"一位OpenAI员工私下说道。其他员工则表示，他们正在权衡金钱与 在Meta可能产生的影响力，与留在OpenAI相比孰轻孰重。一些人认为在OpenAI能发 ...

Core Viewpoint - Figma has officially filed for an IPO in the United States after a failed acquisition by Adobe, aiming for aggressive expansion through acquisitions and investments post-listing [1][2]. Group 1: IPO Details - Figma plans to list on the New York Stock Exchange under the ticker "FIG" and is considered one of the most anticipated IPOs in recent years due to its rapid growth and high valuation in the private market [2]. - In Q1 of this year, Figma's revenue grew by 46% year-over-year, increasing from $156.2 million to $228.2 million, while net profit rose from $13.5 million to $44.9 million [2]. Group 2: Customer Base and Growth - Figma's rapid growth is attributed to its large and high-quality customer base, with over 13 million monthly active users, of which only one-third are designers [4]. - Approximately 85% of monthly active users are from outside the United States, contributing to 53% of the company's revenue [4]. - Figma has around 450,000 enterprise customers, with 1,031 clients generating over $100,000 in annual revenue, marking a 47% increase [4]. Group 3: Investor Returns - The IPO is expected to provide much-needed returns for Silicon Valley venture capital firms, with Index Ventures being the largest external shareholder at 17%, followed by Greylock at 16%, Kleiner Perkins at 14%, and Sequoia at 8.7% [5]. - CEO Dylan Field, who co-founded Figma in 2012, holds 51.1% of the voting rights prior to the IPO [5]. Group 4: Future Growth Strategy - Figma plans to adopt an aggressive expansion strategy, with CEO Field indicating that investors should "expect bold moves," including acquisitions and investments [7]. - As of the end of March, Figma had $1.54 billion in cash and cash equivalents, providing ample resources for future investments and acquisitions [7]. - Recent acquisitions include a $14 million purchase of a tech company's assets and team, a $35.5 million acquisition of a content management system software company, and investments in cryptocurrency [9].

Core Viewpoint - Apple is secretly developing at least seven head-mounted devices, including three Vision series products and four smart glasses, with a focus on the smart glasses market as a significant growth opportunity [2][4][7] Group 1: Vision Product Line - The Vision Pro M5 version is expected to start mass production in Q3 2025, with an estimated shipment of 150,000 to 200,000 units, featuring an upgraded M5 processor while maintaining other specifications [4] - The Vision Air is planned for mass production in Q3 2027, designed to be over 40% lighter than the Vision Pro, utilizing new materials and a lower price point [4][5] - The second generation of Vision Pro is anticipated to begin production in H2 2028, featuring a new design and a Mac-level processor [5] Group 2: Smart Glasses Roadmap - Smart glasses similar to Ray-Ban are projected to start mass production in Q2 2027, with expected shipments of 3 million to 5 million units or more, offering various frame and lens material options [7] - These smart glasses will not have display capabilities but will include features like audio playback, camera, video recording, and AI environmental awareness, potentially replacing some TWS earphones and smartphone camera functions [7] - Advanced XR glasses are planned for mass production in H2 2028, supporting voice control and gesture recognition, with AI functionality being crucial for success [7] Group 3: Challenges and Market Dynamics - A previously planned display accessory project has been paused since Q4 2024 due to competitive disadvantages, with a potential for re-evaluation and relaunch [9] - Apple’s competitive advantage in the Vision series and smart glasses lies in its strong hardware development capabilities and ecosystem integration, while challenges remain in AI-driven operating systems and software development [9][10] - Before Apple launches its display-capable smart glasses, multiple brands are expected to release similar products to capture early market share, although these early products may remain niche for the next two years [10]

Core Viewpoint - The global humanoid robot market is on the brink of explosion, with major tech companies competing for ecosystem dominance, and China potentially leading due to its manufacturing base and policy support [2][8]. Group 1: Market Potential - By 2050, the global humanoid robot market is expected to reach an annual revenue of $5 trillion, with cumulative adoption reaching 1 billion units, which is approximately double the total revenue of the top 20 global automotive manufacturers in 2024 [2][12]. - The humanoid robot index has risen by 14.4% this year, outperforming the S&P 500 by about 11 percentage points, indicating strong market interest and performance [2]. Group 2: Competitive Landscape - Major tech companies like Apple, Meta, Google, and Huawei have recently accelerated their open-source strategies in the humanoid robot field to attract developers to their ecosystems [2][4]. - Meta has launched the V-JEPA 2 world model with 1.2 billion parameters, while Google DeepMind introduced Gemini Robotics On-Device, allowing local operation of its robot models [6]. Group 3: Government Support - The Chinese government has significantly increased its support for the humanoid robot industry, with investment funds totaling approximately 187 billion RMB [9]. - Various local governments have established substantial funds to support the development of the humanoid robot value chain, including a 10 billion RMB fund in Wuhan [9]. Group 4: Commercial Deployment - Companies like Foxconn and NVIDIA are discussing deploying humanoid robots in new AI server manufacturing facilities, while Amazon is developing systems for humanoid robots to operate as delivery workers [11]. - The transition from concept validation to actual commercial application of humanoid robots is evident, particularly in manufacturing and logistics, where efficiency and cost reduction are being realized [11].

Core Viewpoint - OpenAI has begun renting Google's TPU chips for the first time on a large scale, marking a significant shift away from its reliance on NVIDIA chips, which have dominated the AI chip market [2][4]. Group 1: OpenAI's Shift to Google TPU - OpenAI's decision to use Google TPU chips is driven by a surge in demand for computational power, with paid ChatGPT subscriptions increasing from 15 million to over 25 million since the beginning of the year [2][3]. - The collaboration with Google allows OpenAI to reduce its dependency on Microsoft's data centers and presents an opportunity for Google to challenge NVIDIA's GPU market dominance [2][4]. - OpenAI's spending on NVIDIA server chips exceeded $4 billion last year, with projections indicating that spending on AI chip services could approach $14 billion by 2025 [3]. Group 2: Industry Trends and Competitors - The increasing demand for AI capabilities has prompted several companies, including Amazon, Microsoft, OpenAI, and Meta, to develop their own inference chips to lessen reliance on NVIDIA [2][4]. - Meta is also considering the use of TPU chips, indicating a broader trend among major AI players to explore alternatives to NVIDIA [5]. - Google has been developing TPU chips for about a decade and has been offering this service to cloud customers since 2017, with other companies like Apple and Cohere also utilizing Google Cloud's TPU [4][5].