这份清单的标题为"伊朗的新目标"，其中包括 三处英伟达设施、 五处亚马逊设施、五处微软设施、六 处 IBM 设施、三处 Palantir 设施、四处谷歌设施和三处甲骨文设施。 这些目标被描述为"敌方技术基础设施"，以三张幻灯片的形式在 Telegram 上展示，附有供应商 名称、设施性质、所在位置以及各自所从事工作的简要描述。 塔斯尼姆通讯社报道称："随着地区冲突升级为基础设施战，伊朗的合法目标也在逐步扩大。" 伊朗誓言要袭击一系列研发设施，包括： 3月12日消息，据伊朗官方媒体公布了一份清单，列出了近 30 个可能成为攻击目标的地点。其 中包括 英伟达公司的 三处设施。 据报道，伊朗已将亚马逊、谷歌、IBM、微软、英伟达、Oracle 和 Palantir 的设施列为报复性 打击的合法目标。伊朗伊斯兰革命卫队已确定了巴林、以色列、卡塔尔和阿联酋境内的 29 个地 点，这些地点设有办公室、数据中心和研究设施，伊朗计划摧毁这些目标。 5）Oracle 位于耶路撒冷的区域云服务办事处及其位于阿布扎比的总部； 6）亚马逊位于特拉维夫和海法的办事处，以及 AWS 的更多数据中心。 数小时后，伊朗中央军事指挥部 Kha ...

Core Insights - Meta is delaying the release of its new AI model, "Avocado," by approximately two months due to underperformance in internal tests compared to competitors like Google, OpenAI, and Anthropic [1][2][3] - The company is considering temporarily licensing Google's Gemini to enhance its AI products following the disappointing results [1][3] Development and Investment - Meta has invested significantly in AI, spending billions on hiring top talent and committing $600 billion to build data centers, with a projected expenditure of up to $135 billion in 2023, nearly double the $72 billion spent in 2022 [3] - The development of Avocado has been ongoing for months, with the model expected to launch around May instead of March due to its inability to meet performance benchmarks in reasoning, coding, and writing [3][10] Performance Comparison - Avocado has shown improved performance over Meta's previous AI model but still lags behind Google's Gemini 3.0, which was released in November [4] - Despite setbacks, Meta's leadership remains optimistic about the trajectory of their AI models, with expectations for significant advancements in the near future [4][7] Organizational Changes - Meta's AI division, led by Alexandr Wang, has faced some turnover, with researchers leaving before the release of Avocado [12] - The company is restructuring its AI teams, creating a new team under Andrew Bosworth to collaborate with Wang, indicating a strategic shift in its AI development approach [15][16] Competitive Landscape - Meta's AI efforts are closely monitored within the competitive landscape of AI development, where Google, OpenAI, and Anthropic are perceived as leaders [7] - The company has historically supported open-source models, which contrasts with the more restrictive approaches of its competitors [15]

v=WFxxNMxPUec&persist_gl=1&persist_gl=US&hl=en&hl=en [webPageType] => WEB_PAGE_TYPE_UNKNOWN [rootVe] => 83769 ) ) [signInEndpoint] => Array ( [nextEndpoint] => Array ( [clickTrackingParams] => CAEQptEGIhMIz-7m56mdkwMVJ7D0Bx23mybdygEEZrfNdw== [commandMetadata] => Array ( [webCommandMetadata] => Array ( [url] => /watch. v=WFxxNMxPUec&persist_gl=1&persist_gl=US&hl=en [webPageType] => WEB_PAGE_TYPE_UNKNOWN [rootVe] => 83769 ) ) [urlEndpoint] => Array ( [url] => /watch. v=WFxxNMxPUec&persist_gl=1&persist_gl=US&h ...

以下文章来源于泽平宏观商学 ，作者泽平宏观商学 泽平宏观商学 . 前沿科技企业实战研学 读万卷书行万里路 2025 年 12 月 用心打造最有品质的实战研学。读万卷书，行万里路。把教室搬到世界上最优秀的企业， 请最优秀的企业家和科学家讲课。顺势而为，把握机遇。正心正念，坚持做长期正确的 事！ 我们期待与更多的企业家朋友们携手同行，共赴 2026 实战研学之旅！ 泽 ZEP 平 INGM 宏 ACR 观 O RE 商 SEAR 学 CH 1-6月日程安排 · 2月27日 北京 国内头部智驾科技企业 人形机器人 年研学安 型成 11月4日-11日 洛杉矶、拉斯维加斯、旧金山 CES、英伟达、特斯拉 谷歌、罗宾汉、playground 斯坦福大学、伯克利大学 3月27日-28日 苏州 追觅科技、魔法原子、灵猴机器人 闭门投研会-2026 Al 的中国力量(一) 3月29日 上海 长三角校友会 4月20日-21日 香港 2026 香港 Web3 嘉年华 复星财富、HashKey、港交所 5月22日-23日 1151 11:47:50 深圳 华为、新凯来、江波龙、莫界科技 闭门投研会—2026 Al 的中国力量(二) ...

关注 「电动车公社」 和我们一起重新思考汽车 今年最变态的事情，除了金价、油价暴涨之外，莫过于内存涨价了。 就在前两天，手机圈直接放了个大招： 多款新品集体涨价，少则千百块，多则两三千，内存越大的机型涨得越多。这也是国内手机市场发展以来，首次出现全品牌同步普涨的态势。 电脑圈也不例外，内存条直接被炒成天价。 过去DIY一个台式机，组几个内存条，也就是千把块钱。但如今，最顶级的256G DDR5服务器内存，居然飙到了4.2万元的天价！ 而我们所处的汽车行业，也没能躲过这一刀。 1月初，李斌、雷军公开说过，今年最大的成本压力不是原材料，而是内存。 "（每台车）光内存涨价就要几千块，这个汽车的成本（在这里），（新SU7）加量不加价，肯定做不到。" 甚至对于车企来说，因为现在行业研发方向都在向AI大模型靠拢，内存涨价已经不只是成本压力这么简单了。 这意味着，如果不舍得花钱买内存，甚至可能会影响智能化研发的进度。 凑齐100根，就可以在北京上海这样的主城区全款买房了……这哪是内存条，简直是金条 ！ 那么问题来了：这一轮内存涨价，到底什么时候是个头？ 今天，我们就从内存涨价的原因开始聊起，看一看到底是谁在操纵大盘，又会对 ...

Core Viewpoint - Meta Platforms has delayed the release of its next artificial-intelligence model, indicating potential strategic shifts in its AI development timeline [1] Group 1 - The delay in the AI model release may impact Meta's competitive position in the rapidly evolving AI landscape [1] - This decision could reflect broader challenges the company faces in its AI initiatives and product development [1] - The postponement may also affect investor sentiment regarding Meta's innovation capabilities and future growth prospects [1]

以下文章来源于华尔街见闻Max ，作者杨宸 华尔街见闻Max . 华尔街见闻姊妹号，直击全球市场风暴，助力投资者决策。 伊朗战争的爆发打破了美国散户投资者持续逾一年的"逢跌必买"惯性。摩根大通最新零售雷达报告显示， 散户群体正出现"持续性疲软迹象" ，这是自该行追踪 以来极为罕见的信号。 与去年4月"对等关税"冲击后散户创纪录抄底的反应截然不同，本次地缘冲突爆发后， 散户周购买规模骤降约30% 。与此同时，散户ETF周度净流入减少 22%，终结了此前长达三个月的稳定买入势头。 摩根大通认为，此轮缺乏集中性抛售清仓的市场走势（不同于2025年4月标普500在数日内急跌进入调整区间的行情）或令散户普遍保持观望。 值得注意的是，尽管整体买入力度明显收缩， 散户对AI及科技方向的偏好依然显著 。本周散户继续增持英伟达、微软等科技巨头，同时大幅减持能源股，显 示出清晰的板块轮动逻辑。 买入冲动骤冷：散户整体流量跌破年均水平 根据摩根大通报告，3月5日至11日当周，散户总流入规模降至67亿美元，低于过去12个月周均71亿美元的水平。 散户持续偏好ETF（净流入63亿美元），对个股的兴趣则大幅萎缩，单股净流入仅4亿美元。 ...

Core Insights - Quarterly filed Form 13Fs are crucial for tracking stock transactions by prominent money managers, providing insights into market trends and investment strategies [1] Group 1: Stanley Druckenmiller's Investment Moves - Stanley Druckenmiller sold out of 31 stocks, reduced 16 holdings, added to 13 existing stakes, and opened 28 new positions during the fourth quarter [2] - Druckenmiller completely sold his position in Sandisk (SNDK), which he had held for a very short time, indicating a quick profit-taking strategy [4] - The selling of Sandisk may reflect skepticism towards the sustainability of the AI-driven market rally, as Druckenmiller expressed concerns about the potential overhype of AI technologies [8] Group 2: Sandisk's Market Performance - Sandisk has experienced significant price fluctuations, trading between $40 and $100 per share in Q3 and $110 to $240 in Q4, suggesting Druckenmiller likely achieved substantial returns [5] - The company benefits from high demand for NAND flash memory and solid-state drives, particularly in AI-accelerated data centers, allowing it to maintain premium pricing [7] Group 3: Alphabet's Investment Appeal - Druckenmiller increased his stake in Alphabet (GOOGL) by 277% during the fourth quarter, reflecting confidence in its market position [10] - Alphabet has achieved remarkable growth, with its shares increasing over 12,000% since its IPO, driven by its dominance in internet search and the expansion of Google Cloud services [11] - Google Cloud's sales surged 48% year-over-year in Q4, indicating strong growth potential and positioning it as a future cash flow leader for Alphabet [13]

听雨 发自 凹非寺 量子位 | 公众号 QbitAI Gemini一代更比一代强，谷歌也在持续发力，用模型能力赋能自家产品。 劈柴哥正式宣布，谷歌地图推出 "Ask Maps" 和 "沉浸式导航" 两项新功能，均由Gemini模型驱动。 不少网友欢呼，谷歌地图已经很久没有重大更新了，现在终于来了。 官方也表示，此次推出的新功能，是谷歌地图 十多年来的最大升级 。 有网友评价：单单Ask Maps功能，就足以让市面上一半的旅行规划应用黯然失色。 Ask Maps ：通过对话，直接向地图提出复杂的现实需求并获取定制化的解答，例如"最近的公共厕所在哪里"或者"附近哪个公园适合一 家人周末徒步3小时并且可以解决午餐"。 沉浸式导航 ：推出全新3D视图以更好地反映用户周围环境，并提供车道、人行横道、交通信号灯等道路细节。 大厂一动手，垂直应用全完蛋 ，这出戏码又要上演了吗？ 还有网友毫不留情地表示："以前我用苹果地图，现在我要考虑转投谷歌了。" Ask Maps：有问题不用查攻略，直接问地图 此次最令人惊喜的升级，莫过于 Ask Maps 。 我下班后有几个朋友从曼哈顿中城东区过来和我见面。你能帮我们找一家位于办公室和 ...

Core Viewpoint - The rapid expansion of data centers driven by the artificial intelligence boom is facing backlash due to rising electricity costs, with communities questioning the impact on power bills and local power grids [1][3][4]. Group 1: Electricity Price Increases - Since 2020, residential electricity prices in the U.S. have increased by over 36%, from 12.76 cents per kilowatt-hour to 17.44 cents per kilowatt-hour as of February 2026, with projections to reach 19.01 cents per kilowatt-hour by September 2027 [4]. - Retail electricity prices have risen faster than inflation since 2022, and this trend is expected to continue through 2026 [5]. - The EIA noted that regions with high residential electricity prices may experience increases above the national average [9]. Group 2: Market Mechanisms and Policy Impact - A report from SemiAnalysis indicated that market design and policy decisions, rather than just the growth of AI infrastructure, significantly contribute to rising energy prices [2]. - The Base Residual Auction mechanism in the PJM Interconnection area has been identified as a major factor in escalating energy prices, as it requires consumers to pay for expected electricity costs two years in advance [6]. - Forecasts from PJM often overestimate future demand, which can lead to inflated energy prices, especially as many planned data centers face construction delays [7]. Group 3: Corporate Responses and Commitments - Major tech companies, including Microsoft and Anthropic, have pledged to cover additional electricity costs from their data centers and invest in local communities [13]. - President Trump has engaged with AI executives to affirm the Ratepayer Protection Pledge, which aims to prevent new AI data center costs from being passed on to consumers [14]. - Despite these commitments, skepticism exists regarding the profitability of hyperscalers, which may affect their ability to fulfill these pledges [15]. Group 4: Renewable Energy and Future Opportunities - Tech companies are increasingly focusing on renewable energy sources to meet data center needs, as energy availability concerns grow [17]. - The average wait time for grid connections in primary data center markets is already between four to six years, with up to ten years in cities like Tokyo [17]. - Global energy shortfalls could create significant opportunities for energy producers, particularly in renewable energy, although skepticism about sustainability pledges remains [18].