（原标题：玻璃，革命芯片？） 过去半个世纪以来，人类一直痴迷于让事物"更小"。这是在单个芯片上集成更多晶体管的唯一途径。将 晶体管缩小到10纳米、5纳米和3纳米，正是半导体技术的定义。但最终，物理定律给出了冷酷的判 决："你不可能再缩小了。" 对此，人们改变了想法。 "如果我们不能缩小单个单元的尺寸，为什么不把几个单元组合起来建造更大的东西呢?" 这一个问题改变了游戏规则。如今，核心关注点不再是芯片内部的微观电路，而是连接芯片的"桥梁"以 及支撑它们的"接地"。纳米级战争已经结束，微米级战争已经开始。 而在这片战场的正中央，矗立着一块透明的玻璃。 为什么芯片不能再大了 人工智能模型越来越大，芯片上需要集成的晶体管数量也随之增加。为了容纳更多晶体管，芯片尺寸必 须增大——但芯片尺寸存在一个无法突破的极限。 这就是良率问题。不能再小了，也不能再大了。单片芯片是条死路。 所以，业界的应对之策是反其道而行之。 分开，然后重新连接 想象一下，用3D打印机一次性打印出一座霍格沃茨城堡。如果打印过程中出现一个地方出错，整个作 品就得扔掉。但如果你用乐高积木搭建呢?一块坏的积木——换一块就行了。 Chiplet就像乐高积木。 ...

Core Insights - Generative AI is viewed as a transformative technology by Amazon's CEO Andy Jassy, indicating its potential to significantly enhance customer experiences across the company [1] - Elon Musk predicts that by 2040, humanoid robots could create a market worth $250 trillion, representing a major shift in the global economy driven by AI innovation [2][3] - Major firms like PwC and McKinsey acknowledge the potential of AI to unlock multi-trillion-dollar opportunities, reinforcing the optimistic outlook on AI's economic impact [3] Company and Industry Analysis - A breakthrough in AI technology is believed to be redefining work, learning, and creativity, leading to increased interest from hedge funds and top investors [4] - There is speculation about an under-owned company that may play a crucial role in the AI revolution, suggesting that it could be a significant investment opportunity [4][6] - Prominent figures in technology and finance, including Bill Gates and Warren Buffett, recognize AI as a major technological advancement with the potential for substantial social benefits [8]

2月21日消息，据英国《金融时报》（Financial Times）引述知情人士报导称，人工智能（AI）芯片龙头大厂英伟达（Nvidia）即将完成对AI技术大厂 OpenAI的投资，将以300亿美元的投资取代双方于去年达成的1,000亿美元长期合作意向。 去年9月22日，OpenAI 和 英伟达共同宣布了一份长期的战略合作伙伴关系意向书， OpenAI 的下一代 AI 基础设施将部署至少 10 吉瓦的英伟达AI系统， 以训练和运行其下一代模型。英伟达则打算在OpenAI部署新的英伟达系统时向其投资高达 1000 亿美元（即这笔资金将用于资助OpenAI采购英伟达的AI系 统）。第一阶段计划于 2026 年下半年使用英伟达Vera Rubin 平台上线。 目前OpenAI已经宣布的投资与合作项目金额累计已经超过了10000亿美元，这也使得外界对于其是否能够获得足够的资金来完成这些计划充满了质疑。而 为了推进这些计划，OpenAI目前也正在积极地进行新一轮融资。 据路透社今年1月报道，OpenAI在最新一轮募资中计划筹集多达1000亿美元资金，使得该公司估值高达约8300亿美元。 《金融时报》最新的报道显示，Op ...

硅谷创始人与公司之间的"生死同盟"正在瓦解 自2025年中期以来，Meta、谷歌、英伟达三大巨头已经在所谓"收购性招聘"的游戏中豪掷超过360亿美元，这不是传统收购，而是对顶尖人才的直接掠 夺。与此同时，OpenAI与Anthropic之间也上演着"你挖我墙角、我偷你核心"的滑稽循环，顶级研究员的年薪被推至千万乃至数亿美元，创始人不再相 信"使命"，员工也不再相信"股权"。 几个赤裸的现实，勾勒出这场战争的残酷轮廓。 硅谷，曾以"改变世界"为信仰的科技乌托邦，如今已沦为赤裸裸的人才角斗场。创始人与公司之间的"生死同盟"正在瓦解，取而代之的是一份份价值数亿 美元的报价单和永不停歇的"转会"闹剧。 薪酬已无天花板：Meta被曝向前苹果AI基础模型团队负责人庞若鸣开出超过2亿美元总包，九位数录用通知在顶级圈内已成常态。 "合法挖角"成主流：英伟达与Groq的200亿美元"技术许可"协议实为规避反垄断的"后门收购"，谷歌24亿美元"买走"Windsurf核心技术团队却抛弃公司，留 下满地狼藉。 忠诚彻底破产：创始人带头"跳船"，早期员工沦为交易筹码，"使命驱动"沦为笑谈，"自由球员"心态席卷硅谷。 资本紧急自救：风投 ...

另外，从产品本身而言，随着头部科技公司的资本开支加大，其与中腰部公司的差距并非不断被缩小，而是越来越大。且随着头部公司的产品迭代速度加 快，市场则会更加青睐优质的产品。在这种背景下小公司逆袭的故事，只会越来越少。 例如当下的AI行业，市场并不相信小公司产品的逆袭，AI公司的底层的竞争力往往体现在算力、算法与模型上，这些都是创业型小公司难以与巨头匹敌的 客观事实。目前，在AI的全球语境下，谷歌、OpenAI、Meta、亚马逊等头部科技公司资本开支逐年提升，算力基本处于过剩状态，这种优势是小公司所不 具备的。 需要说明的是，当下AI的竞争需要实打实地投入，没有弯路可以取巧。因此，尽管目前AI是全球最大的风口，但并不意味着小型创业企业可以突围。 如果采用之前对互联网的估值方式对AI大模型公司进行估值，那么很有可能会错失一个时代。 但这并不意味着，所有的大模型公司都能获配超溢价估值。 过去很长一段时间，对于创业型科技企业我们习惯找一家头部公司进行对标，然后贴上相应的标签。例如在芯片领域，一些创业型芯片企业很容易被对标英 伟达，然后给予超出公司实际营收水平且梦幻般的估值。 实际上，从公司本身的质地而言，差距十分明显。而 ...

智通财经2月22日讯（编辑 赵昊）本周美股整体反弹，三大指数周线均录得涨幅，其中标普500指数和 纳斯达克综合指数双双累涨超1%。 | 日期 | 时段 | 标的 | | --- | --- | --- | | 周一 | 益前 | 达美乐比萨、道明尼资源 | | (2月23日) | 盘后 | Hims & Hers Health, BWX Technologies | | 周二 | 盘前 | 家得宝、NRG Energy、亚玛芬体育 | | (2月24日) | 盘后 | AMC院线、纳微半导体、惠普、第一太阳能 | | 周三 | 盘前 | Hut 8、TJX公司、Circle、劳氏 | | (2月25日) | 盘后 | 英伟达、塞富时、IonQ、新思科技、携程网 | | 周四 | 盘前 | D-Wave、Vistra Energy、百度、百济神州 | | (2月26日) | 盘后 | CoreWeave、SoundHound、MP Materials、載尔科技 | | 周五 | 盘前 | Energy Fuels、全球星 | | (2月27日) | | | 财报方面，英伟达是重中之重，公司已被广泛视为人工智能热 ...

Core Insights - Congressman Daniel Meuser disclosed sales of Nvidia stock in early 2026, marking his first transactions since February 2025, where he sold between $1,000 to $15,000 on January 14 and again on January 30 [2][3] Group 1: Trading Activity - In 2025, Meuser sold a total of $580,000 in stock without making any new purchases, contrasting with his 2024 activity where he bought $1.13 million and sold $1.94 million [3] - The recent Nvidia stock sales are attributed to a spouse, suggesting Meuser may not have been aware of these transactions [4] Group 2: Historical Performance - Meuser has consistently sold Nvidia stock since 2022, potentially realizing profits while missing out on significant gains, as Nvidia's stock has increased over 1,150% in the last five years [5] - A detailed history of Meuser's Nvidia transactions shows multiple sales, with significant amounts sold at various price points, including $250,000 to $500,000 on February 11, 2025, at a daily high of $134.48 [5]

Core Viewpoint - Microsoft and Nvidia are currently trading at historically low valuations, presenting a significant buying opportunity for investors [1][2]. Group 1: Stock Valuation - The price-to-forward earnings ratio is highlighted as the best metric for valuing companies in the rapidly evolving AI landscape, as it reflects future potential rather than past performance [4][5]. - Both Nvidia and Microsoft are currently valued at low levels compared to their historical performance, making them attractive investments [5][7]. Group 2: Company Performance and Growth - Nvidia is positioned to benefit significantly from AI spending, with its GPUs being essential for AI model training and inference, and analysts project a 65% growth in fiscal year 2027 [8]. - Microsoft is experiencing substantial growth in its cloud computing platform, Azure, which has risen 39% year over year in Q2 FY 2025, driven by demand from AI companies [10]. - The integration of AI features into Microsoft's core software suite is expected to contribute to its growth alongside its cloud services [10]. Group 3: Market Context - The recent tech sell-off is viewed as an opportunity for investors to acquire shares of Nvidia and Microsoft, which are seen as resilient and high-growth stocks [11].

Core Viewpoint - Nvidia and Alphabet are major players in the artificial intelligence sector, with Nvidia primarily providing GPUs and Alphabet developing its own TPU, which serves a specific niche in AI computing [1][3][4]. Group 1: Company Overview - Nvidia specializes in graphics processing units (GPUs) that are widely used for AI, cryptocurrency mining, drug discovery, engineering simulations, and gaming graphics [3]. - Alphabet has developed the Tensor Processing Unit (TPU) over the past decade in collaboration with Broadcom, positioning itself as a competitor in the AI computing space [4]. Group 2: Market Dynamics - While Nvidia's GPUs are versatile, they may not be fully utilized for specific workloads, leading to inefficiencies for clients [3]. - Alphabet's TPUs are available exclusively through Google Cloud, which may limit clients' flexibility in switching providers, thus encouraging a balanced use of both Nvidia's GPUs and Alphabet's TPUs [5]. Group 3: Revenue and Future Outlook - A significant portion of Nvidia's revenue is derived from AI, raising concerns about potential revenue impacts if AI spending declines in the future [6].

Market Overview - The stock market rebounded last week, with the Nasdaq ending a five-week losing streak, rising 1.9% due to strong performances from major tech companies like Meta Platforms, Nvidia, and Amazon [1] - The S&P 500 increased by 1.1%, breaking a two-week decline, aided by a Supreme Court ruling against President Trump's emergency tariffs [1] Supreme Court Ruling - The Supreme Court ruled 6-3 against Trump's tariffs, stating that no president had previously used the statute to impose tariffs of such magnitude, requiring clear congressional authorization for such actions [1] - Following the ruling, the S&P 500 rose 0.7%, although some companies like Nike experienced a decline due to ongoing tariff concerns [1] Big Tech Performance - Major tech stocks saw significant gains, with Meta up 2.5% and Nvidia up 3.8% after Meta announced plans to use Nvidia's chips in its data centers, highlighting strong AI demand [1] - Amazon's shares surged 5.6% after a regulatory filing revealed that Bill Ackman's Pershing Square increased its position in the company [1] - Alphabet's stock initially lagged but later rallied to end the week up 3% [1] Private Credit Concerns - Concerns in the private credit market arose from Blue Owl Capital's decision to restrict withdrawals from its private debt fund, causing a nearly 6% drop in its shares [1] - Major private asset managers like Ares Management and Blackstone faced significant declines, with Ares down 8% and Blackstone down 6.6% [1] - Despite these concerns, BlackRock's exposure to private credit did not raise alarms, as its shares only dropped 1% before recovering [1] Portfolio Adjustments - Capital One was the only financial stock traded last week, with additional shares purchased [1] - The company exited its position in Texas Roadhouse due to concerns over ongoing beef inflation issues [1]