晚安咖啡GoodNightCoffee

晚安咖啡GoodNightCoffee 的全部节目

共 1380 集。

《晚安咖啡》科技夜话(英文版):英伟达财报超预期,Anthropic IPO火爆

《晚安咖啡》科技夜话(英文版):英伟达财报超预期,Anthropic IPO火爆

晚安咖啡GoodNightCoffee

Hi everyone~~ Good evening! It’s Wednesday, November 18, 2026. Welcome to Goodnight Coffee—I’m your host, An’an! Tonight I’ll walk you through some interesting business and tech stories. We’ll keep it light and fun, and wrap up today on an upbeat note. Ready? Let’s get started! @@@ According to CNBC, NVIDIA shares rose 3.72% in Wednesday premarket trading after the company once again reported better-than-expected results. Its third-quarter revenue was $115.84 billion, more than $6 billion above analysts’ $109.10 billion estimate; fourth-quarter revenue guidance was $130 billion, above LSEG’s $125.1 billion consensus. Boosted by the results, new cloud providers CoreWeave, Nebius, and Iren rose 6.1%, 8.2%, and 7.2%, respectively. Shortly before the earnings release, CEO Jensen Huang, who resigned last week, made his first public statement, saying he was willing to return to the company after several prominent tech figures voiced support. But new board chair Harvey Jones—NVIDIA’s first-ever board chair—said on the earnings call that the board does not plan to have Huang serve as CEO again. Interim CEO Colette Kress told CNBC that hyperscalers are urgently seeking more available compute, that claims of a shift in customer types are not true, and that all types of customers are still working with NVIDIA, whose customer base is very broad. The US Commerce Department announced regulatory adjustments on Monday to restrict Chinese companies’ remote access to compute in overseas data centers, aiming to stop China from obtaining NVIDIA’s most advanced chips. NVIDIA said in its earnings commentary that its outlook assumes no data center compute revenue from China. @@@ According to a report from Jiemian News published with authorization from 36Kr, Gerstner, CEO of Altimeter Capital, said the firm applied to subscribe for $5 billion worth of shares in Anthropic’s IPO—its largest allocation request in 20 years. The IPO is 10 to 20 times oversubscribed, and the order book is more than five times the size of the previous largest IPO order. However, the request has not yet been filled, and the oversubscription multiple is based on its own market judgment. Anthropic plans to price the deal local time on November 18 and list on Nasdaq the next day under the ticker ANT, with an offering range of $70 to $78. The IPO could raise up to $64.4 billion, plus a private placement of about $26 billion, for a total of up to $90.4 billion. At the top end, the fully diluted valuation would be about $1.4 trillion. Investors including Sequoia, Iconiq, Lightspeed, Accel, and Altimeter plan to subscribe billions of dollars. Private placement participants include Google and NVIDIA at $10 billion each, AMD at $2.5 billion, Broadcom, Micron, and Microsoft at $1 billion each, and a TSMC affiliate at $500 million. After completion, Amazon and Google will hold 14.7% and 13.7% of outstanding shares, respectively. @@@ According to Readhub, Microsoft has launched six AI models spanning coding, cybersecurity, image generation, and voice. Its coding model MAI Code 2 delivers 98% of GPT-6.1 Sol’s performance at only 40% of the price; against Claude Opus 5.5, it reaches 96% performance at just 5% of the cost. Cybersecurity model MAI Cyber 2 Flash outperformed GPT-6 Astra in relevant security tests, with per-task cost less than 2% of the latter’s. Image model MAI Image 2.7 ranked first in image editing. Mustafa Suleyman, Microsoft’s AI chief, said the company is accelerating work on its own general AI models to reduce dependence on external models, and plans to open model weights to enterprises so customers can customize models with their own data. @@@ According to CNBC, JPMorgan is testing with Hyundai Card how AI agents can handle financial transactions without direct human intervention and assessing whether they could eventually negotiate and make payments with each other. The two companies have completed an initial proof of concept simulating an FX transaction. In the test, Hyundai Card’s AI agent requested an FX transaction to maintain euro liquidity in its Kinexys blockchain deposit account. JPMorgan said this was the first test of an autonomous bank agent versus a client agent on a private blockchain for corporate treasury management. Wee Kee Toh, global head of Kinexys Insights at JPMorgan, said the next step will be to test autonomous agents going beyond executing transactions and interacting with one another in the ecosystem—the biggest part of expanding the first phase. The two companies plan to launch a more complex phase involving multiple AI agents in the first half of 2027. After the initial trial, JPMorgan has not set a commercialization timetable. Toh said the biggest obstacle is not the technology itself, but governance, internal controls, and accountability and risk without direct human oversight. AvePoint’s 2026 State of AI report shows that unauthorized autonomous agent behavior is the third most common AI agent-related security vulnerability, affecting 34.1% of companies that reported such incidents. CEO Tianyi Jiang warned that agent interactions can amplify errors; in the worst case, a compromised or hallucinating agent gives bad data to another, potentially triggering a high-speed cascade of erroneous fund transfers before humans notice. @@@ According to TechWeb, JD.com announced unaudited results for the three months ended September 30, 2026: revenue was RMB309.6 billion, up 3.5% year over year; operating profit was RMB5.2 billion, compared with a loss of RMB1.1 billion a year earlier, swinging to profit; net income attributable to shareholders was RMB6.5 billion, compared with RMB5.3 billion a year earlier; non-GAAP net income attributable to shareholders was RMB8.3 billion, up 44%. By segment, JD Retail revenue was RMB257.966 billion, up 2.9%, with operating profit of RMB15.1 billion and a 5.8% margin; JD Logistics revenue was RMB63.953 billion, up 16.1%; new business revenue was RMB7.831 billion, down 49.8%, with an operating loss of RMB9.642 billion, narrowing from RMB15.736 billion a year earlier, as JD Food Delivery and Jingxi drove a clear reduction in losses. Marketing expenses were RMB15.5 billion, down 26.2%; operating margin was 1.7%, compared with negative 0.4% a year earlier. JD completed its voluntary public takeover offer for CECONOMY at €1.335 billion, acquiring 59.8% of share capital and voting rights; CECONOMY has been consolidated since November 12. As of September 30, cash and short-term investments totaled RMB214.7 billion; third-quarter free cash flow was negative RMB23.861 billion. @@@ According to TechWeb, after new US export control rules took effect, a $480 million loan for ByteDance to buy NVIDIA GPUs in Malaysia was halted. At least five banks evaluating the deal have internally discussed withdrawing or not participating, and one banker said the loan is “effectively dead.” The financing was borrowed by Singapore cloud provider Aolani Cloud and led by DBS Group; the funds were to buy NVIDIA GPUs in Malaysia and then lease the compute to ByteDance for remote use. On November 16, the US Bureau of Industry and Security designated Malaysia, Singapore, Thailand, Indonesia, and Brazil as high-risk transshipment countries; exports, reexports, or in-country transfers of advanced AI chips to these countries require US licenses. The new rules also regulate remote access: if an operating entity knows or should know that the end user is in China, and the aggregate processing performance exceeds 12 million—roughly equivalent to 700 to 800 H100s—the relevant chip supply must undergo license review. Banks are not worried about ByteDance’s ability to repay, but that once they lend, they could become liable for violating US export controls. Aolani said it is coordinating with regulators and financing partners to ensure compliance. As of publication, ByteDance had not publicly commented. @@@ According to Huxiu 24 Hours, Zong Fuli has made another personnel adjustment, with 173 managers removed from their posts. @@@ Okay, that’s it for today’s show! Thank you for spending time with me and chatting about so many light and fun little stories. I hope you’re in a good mood today too. If you enjoyed the show, you’re very welcome to subscribe and share it with your friends—it helps us a lot. See you next time!

9 分钟 23 次播放 16小时前

《晚安咖啡》科技夜话:英伟达财报超预期,Anthropic IPO火爆

《晚安咖啡》科技夜话:英伟达财报超预期,Anthropic IPO火爆

晚安咖啡GoodNightCoffee

嗨~~大家晚上好呀!今天是2026年11月18日,星期三,欢迎来到晚安咖啡,我是主播安安!今晚我会陪你一起看看有趣的商业科技动态,轻松聊、开心听,给今天画个元气满满的句号。准备好啦,我们开始吧! @@@ 据CNBC报道,英伟达股价周三盘前交易中上涨3.72%,此前公司再次公布超预期财报。其第三季度营收为1158.4亿美元,比分析师预期的1091.0亿美元高出逾60亿美元;第四季度营收指引为1300亿美元,高于LSEG统计的1251亿美元共识。受业绩带动,新型云服务商CoreWeave、Nebius和Iren股价分别上涨6.1%、8.2%和7.2%。在财报发布前不久,上周辞职的CEO黄仁勋首次公开表态,称在多位知名科技人士声援后,他愿意回归公司;但新任、也是英伟达首位董事会主席Harvey Jones在财报电话会上表示,董事会不打算让他重新担任CEO。临时CEO Colette Kress对CNBC表示,超大规模云服务商正急切寻求更多可用算力,所谓客户类型发生变化并不属实,各类客户都仍在与英伟达合作,公司客户群非常广泛。美国商务部周一宣布调整法规,限制中国企业远程访问海外数据中心算力,以阻止中国获取英伟达最先进芯片;英伟达则在财报评论中称,其展望不假设来自中国的任何数据中心计算收入。 @@@ 据界面新闻经36氪授权发布的消息,Altimeter Capital首席执行官格斯纳称,公司申请认购Anthropic IPO价值50亿美元的股票,为其20年最大配售申请;此次IPO超额认购10至20倍,订单规模是此前最大IPO订单五倍以上,但申请尚未完成购买,超额倍数基于其市场判断。Anthropic计划当地时间11月18日定价,次日以代码ANT在纳斯达克挂牌,发行区间70至78美元,IPO最多募资644亿美元,另安排约260亿美元私募配售,合计最高904亿美元,按上限定价完全稀释估值约1.4万亿美元。红杉、Iconiq、光速、Accel和Altimeter等投资方计划认购数十亿美元。私募配售方包括谷歌、英伟达各100亿美元,AMD25亿美元,博通、美光、微软各10亿美元,台积电关联5亿美元;完成后亚马逊、谷歌分别持有流通股14.7%和13.7%。 @@@ 据Readhub报道,微软推出六款AI模型,覆盖编程、网络安全、图像生成及语音等领域。其中,编程模型MAI Code 2的性能可达GPT-6.1 Sol的98%,价格仅为其40%;相较Claude Opus 5.5,其性能达到96%,成本仅为5%。网络安全模型MAI Cyber 2 Flash在相关安全测试中的表现超过GPT-6 Astra,单项任务成本不到后者的2%。图像模型MAI Image 2.7位列图像编辑排名第一。微软AI负责人苏莱曼称,公司正加速研发自主通用AI模型,减少对外部模型的依赖,并计划向企业开放模型权重,支持客户利用自有数据定制模型。 @@@ 据CNBC报道,摩根大通正与韩国现代信用卡测试AI代理如何在无直接人工干预下处理金融交易,并评估其最终能否相互谈判和付款。两家公司已完成一项模拟外汇交易的初步概念验证。测试中,现代信用卡的AI代理请求进行外汇交易,以维持其在Kinexys区块链存款账户中的欧元流动性;摩根大通称,这是首个基于私有区块链、面向企业财资管理的自主银行代理对客户代理测试。摩根大通Kinexys Insights全球主管Wee Kee Toh表示,下一步将测试自主代理不再局限于执行交易、而是在生态系统中彼此互动,这是扩大第一阶段范围的最大部分;两家公司计划2027年上半年启动涉及多个AI代理的更复杂阶段。初步试验完成后,摩根大通尚未设定商业化时间表。Toh称,技术本身并非最大障碍,更难的是治理、内部控制以及无直接人工监督下的责任和风险。AvePoint《2026年AI现状》报告显示,未经授权的自主代理行为是第三常见的AI代理相关安全漏洞,影响34.1%报告此类事件的公司。其首席执行官Tianyi Jiang警告,代理互动可能放大错误,最坏情况下,一个被入侵或产生幻觉的代理向另一个提供错误数据,可能在人类察觉前引发高速连锁的错误资金转移。 @@@ 据TechWeb报道,京东集团发布截至2026年9月30日止三个月未经审计业绩:收入3096亿元,同比增长3.5%;经营利润52亿元,上年同期亏损11亿元,实现扭亏;归母净利润65亿元,上年同期53亿元;非GAAP归母净利润83亿元,同比增长44%。分业务看,京东零售收入2579.66亿元,同比增长2.9%,经营利润151亿元,利润率5.8%;京东物流收入639.53亿元,同比增长16.1%;新业务收入78.31亿元,同比下降49.8%,经营亏损96.42亿元,较上年同期157.36亿元收窄,京东外卖和京喜带动明显减亏。营销开支155亿元,同比下降26.2%;经营利润率1.7%,上年同期为负0.4%。京东以13.35亿欧元完成对CECONOMY的自愿公开收购要约,取得59.8%股本及投票权,后者自11月12日起并表。截至9月30日,现金及短期投资合计2147亿元;第三季度自由现金流为负238.61亿元。 @@@ 据TechWeb消息,美国出口管制新规生效后,字节跳动一笔用于在马来西亚采购英伟达GPU的4.8亿美元贷款被叫停。至少五家评估该交易的银行已内部讨论退出或不参与,有银行人士称贷款“实际上已经死亡”。融资由新加坡云服务商Aolani Cloud借款,星展集团牵头安排,资金用于在马来西亚采购英伟达GPU,再以算力租赁供字节跳动远程调用。11月16日,美国商务部工业与安全局将马来西亚、新加坡、泰国、印尼和巴西列为转运高风险国家,向这些国家出口、再出口或境内转移先进AI芯片须申请美国许可;新规还把远程访问纳入管制:若经营实体知道或应知最终用户位于中国,且算力总处理性能超过1200万、约相当于700至800块H100,相关芯片供应须接受许可审查。银行担忧的并非字节还款能力,而是一旦放贷,自身可能成为违反美国出口管制的责任主体。Aolani称正与监管机构和融资伙伴协调确保合规。截至发稿,字节跳动未公开置评。 @@@ 据虎嗅24小时报道,宗馥莉再次进行人事调整,此次共有173名管理人员被免职。 @@@ 好啦,今天的节目到这里就结束啦!谢谢你们陪我聊了这么多轻松有趣的小事,希望今天也有好心情。如果您喜欢我们的节目,也非常欢迎订阅和转发给你的小伙伴,这对我们非常有帮助。那我们下期再见吧!

9 分钟 44 次播放 16小时前

《晚安咖啡》科技夜话:Anthropic拟万亿美元IPO,苹果2027推智能眼镜,阿里发Qwen4-Max

《晚安咖啡》科技夜话:Anthropic拟万亿美元IPO,苹果2027推智能眼镜,阿里发Qwen4-Max

晚安咖啡GoodNightCoffee

嗨~~大家晚上好呀!今天是2026年11月16日,星期一。欢迎来到晚安咖啡,我是主播安安~今天咱们一起聊聊最新的商业科技动态,干货满满,别走开哦! @@@ 据The Verge报道,围绕英伟达联合创始人、前首席执行官黄仁勋的事态正快速演变。报道提到,有传闻称黄仁勋可能正在寻找重返CEO职位的路径。与此同时,Megaspeed发表声明称,Alice Huang自2025年8月以来未在该公司工作。报道称,这份声明引发的疑问多于它能回答的问题。The Verge还表示,其Stepback新闻通讯最新一期已梳理相关情况,而自当天上午以来,事件又出现多项新进展。 @@@ cnBeta援引彭博社报道,Anthropic预计11月19日或20日登陆纳斯达克,寻求1.4万亿美元估值。公司计划通过IPO和一项独立私募配售合计融资最多904亿美元,规模将超过SpaceX IPO的750亿美元基础融资,成为有史以来最大上市交易;知情人士称,此次IPO已获近七倍超额认购。Anthropic拟发行8.25亿股,发行价每股70至78美元。截至10月31日,公司年化营收810亿美元,较9月底增加100亿美元;现金及等价物超过1300亿美元。高盛预计其2030年才能实现年度自由现金流为正;到2045年支出承诺超9000亿美元,其中约710亿美元在2027年9月底前到期。上周五,CEO阿莫代伊与摩根大通CEO戴蒙在纽约IPO活动上同台推介;知情人士称,Anthropic不准中国大陆和香港机构投资者参与IPO认购。此前黄仁勋声誉危机致英伟达股价重挫,Anthropic被寄望接替英伟达成为AI新标杆。 @@@ 据The Verge报道,彭博社记者马克·古尔曼称,苹果计划在2027年6月的全球开发者大会(WWDC)上发布首款智能眼镜,并在当年年底前开始出货。该产品不配备增强现实功能,核心卖点是新版Siri人工智能和iPhone应用生态。原型代号N50,目前运行精简版visionOS,采用基于旧款智能手表芯片的新处理器。眼镜配有类似MagSafe的扁平矩形磁吸充电器,大小约相当于一粒Tic Tac,用户戴着眼镜也能充电。苹果还强调外观、隐私和音质,员工称其音质达到行业领先水平,支持通话、音乐播放和FaceTime音频。眼镜具备Wi-Fi,离开iPhone后,部分非第三方应用功能仍可继续使用,例如通过Wi-Fi继续听音频,但它不会成为完全独立的设备。新版Siri将增加对应用的细粒度控制,苹果先让开发者在iPhone上适配,以期眼镜推出后能直接使用相关功能。 @@@ 据The Verge报道,美国联合发射联盟(ULA)的“火神”(Vulcan)火箭在一次发射中发生异常,未能将40颗亚马逊“Leo”卫星送入轨道。ULA在社交平台X发表声明称,根据发射主管收到的数据,这枚从41号航天发射复合体升空的火箭在飞行10分钟后出现异常,导致亚马逊任务未能入轨;目前尚不清楚故障原因,公司承诺在调查过程中保持透明。这据报是ULA成立20年来首次任务失败。此次发射原计划是亚马逊向ULA购买的38次发射中的第一次。“Leo”是亚马逊对标SpaceX“星链”的卫星项目,但此前进展艰难,迄今仅有少量所需卫星入轨。亚马逊表示,任何发射都有风险,首次任务尤其如此,ULA一直是其可信赖的合作伙伴,相信ULA会从此次任务中吸取教训,并落实“火神”的长期计划。 @@@ 据TechWeb援引快科技消息,美国加州北区联邦地区法院法官诺埃尔·怀斯11月14日全盘驳回三星、美光、SK海力士的撤诉动议,认定原告已合理指控三家违反《谢尔曼法》第一条、达成非法协议,DRAM集体反垄断诉讼进入实质审理。诉讼由美国五州17名消费者及小企业于2026年6月提起,指控三家协同限制常规DRAM供应、推高价格。诉状称,截至2025年第三季度,三家合计占全球DRAM营收逾91%;自2022年底同步削减传统消费级DRAM产能,2023年将大量晶圆产能转向HBM,18个月内传统DRAM价格累涨约697%。法院列举的同步行为包括三星与SK海力士拒绝长期供货协议、三家推迟或叫停新晶圆厂、同季度减产DDR4、美光2025年12月关闭30年历史的Crucial品牌、2026年1月上线高度相似客户审核机制,并提及其1998至2002年曾涉价格操纵案。三家企业称决策为适配AI市场变化,且类似争议2018年已被驳回;法官认为本案事实和证据有本质区别。三家须在2026年12月7日前提交完整答辩文件,首次管理会议定于2027年1月12日。 @@@ 据新浪科技报道,11月16日,据千问消息,阿里巴巴发布全新一代基座大模型Qwen4-Max,总参数2.4T,具备原生视觉能力,支持百万上下文。该模型基于新的架构和训练优化,在智能体、编程等关键能力上持续进化,并在金融、法律等专业垂直领域实现从“信息检索”到“专业交付”。上线期间,千问APP会员可限免体验,使用Qwen4-Max不消耗会员额度。相关报道显示,其预览版已于11月11日首发千问AI平台Token Plan,同步接入千问办公、Qoder、千问APP和QwenStudio,预览窗口持续至11月25日24时,主要用于能力验证和收集反馈。另据澎湃新闻报道,Qwen4-Max已上线千问AI平台提供API服务,缓存命中每百万Tokens低至0.5元,完成单任务平均费用较上代模型下降六成;用户也可通过阿里Token Plan调用,Qwen4系列中小尺寸模型将陆续开源。 @@@ 据新浪科技报道,腾讯文化发文披露,腾讯在28岁生日当天将一年一度总办LinkTime做成特别播客,总办就AI时代机会、挑战与选择发声。讲话提出,所有核心业务团队都要面向AI时代再创业:AI native业务、模型和产品要逆流而上,核心业务一半人守住、一半人打新战役,组织扁平化、以小团队作战,打破山头和中间传话阻碍。腾讯称要做以人为本的AI,坚持造福人类和社会、由人控制、赋能创造力;AI仍处早期,算力基建需自建且价格飞涨,投入会兼顾产出比。其AI金字塔以模型为核心,已进入国内第一梯队;小微定位私人AI助理,目标是“token自由”;元宝未来定位决策和资讯参谋;WorkBuddy在8个月内成为中国提供最多token的单一AI产品。总办还要求心态归零、加快反馈决策和行动、提升人才密度,并提到游戏、金融、安全等团队已与模型Co-Design带来Hy能力提升。 @@@ 据新浪科技报道,华为终端11月16日上午宣布,鸿蒙终端设备数已超过1亿台。此前,华为常务董事、终端BG董事长余承东通过社交媒体宣布,全新子品牌OSTA星芒将于11月20日在“鸿蒙智行享界V8及星芒品牌发布会”上正式亮相。消息称,星芒将共享鸿蒙操作系统、麒麟芯片技术,以及华为覆盖线上线下的渠道与售后体系。 @@@ 新浪科技消息,11月16日,因算力资源供给紧张,DeepSeek面向普通用户的App及网页版对话服务推出限额措施:存续超过7天且无活跃交互的长对话将无法接续;单次输入文字上限约4万字;每日对话额度用尽会提示“因服务器繁忙,使用额度已达上限”。分析认为,此举意在回收闲置会话占用的KV缓存,优先保障活跃用户推理。10月以来其服务多次承压,官方状态页记录8起故障,发生在10月2日、8日、12日(两次)、15日、20日、27日和11月3日,多为响应变慢等性能降级;10月15日故障持续超4小时,C端服务及V4 Pro、V4.1 Flash API同步受限,事后对API用户赠金补偿。8月至10月,V4.1 Flash API可用性99.46%,V4 Pro为99.74%。DeepSeek近期开放全家桶应用API,API侧实行峰谷分时定价和独立并发限额。OpenRouter周榜显示,10月19日至25日,V4.1-Flash-1017以33.6万亿Token周调用量居首,环比增801%。 @@@ 据新浪科技报道,11月12日超聚变创业板IPO过会,从5月22日受理到过会仅用174天,80亿元募资额使其成为创业板年内最大IPO项目。公司营收从2023年250.92亿元增至2025年582.46亿元,2026年上半年为372.86亿元,同比增20.78%;AI服务器收入占比从25.03%升至53.29%。但主营业务毛利率从14.27%降至2025年8.58%,净利率1.77%;2026年上半年毛利率回升至15.49%,净利润21.08亿元,已超2025年全年两倍。2025年研发费用率2.02%,低于中科曙光、紫光股份和浪潮信息。2026年上半年经营现金流净额-156.85亿元,应收账款153.03亿元,较2025年末增超170%,存货356.62亿元,增68%。截至2026年7月末在手订单301.13亿元,其中AI服务器占73.70%。 @@@ 好啦,今天的节目到这里就结束啦!谢谢你的陪伴,愿你今晚有个好心情~如果您喜欢我们的节目,也非常欢迎订阅和转发给你的小伙伴,这对我们非常有帮助。那我们下期再见吧!

12 分钟 99+ 次播放 2天前

《晚安咖啡》科技夜话(英文版):Anthropic拟万亿美元IPO,苹果2027推智能眼镜,阿里发Qwen4-Max

《晚安咖啡》科技夜话(英文版):Anthropic拟万亿美元IPO,苹果2027推智能眼镜,阿里发Qwen4-Max

晚安咖啡GoodNightCoffee

Hi there~~ Good evening, everyone! Today is Monday, November 16, 2026. Welcome to Goodnight Coffee. I'm your host, An'an. Today we're going to talk about the latest business and tech developments—packed with useful insights, so stay tuned! @@@ The Verge reports that events surrounding Nvidia co-founder and former CEO Jensen Huang are evolving rapidly. The report notes rumors that Jensen Huang may be looking for a path back to the CEO role. Meanwhile, Megaspeed issued a statement saying that Alice Huang has not worked at the company since August 2025. The report says the statement raises more questions than it answers. The Verge also says the latest issue of its Stepback newsletter has reviewed the situation, and since that morning, several new developments have emerged. @@@ cnBeta, citing Bloomberg, reports that Anthropic is expected to list on Nasdaq on November 19 or 20, seeking a $1.4 trillion valuation. The company plans to raise up to $90.4 billion through an IPO and a separate private placement, which would exceed SpaceX's $75 billion base IPO raise and become the largest listing ever; people familiar with the matter say the IPO is nearly seven times oversubscribed. Anthropic plans to issue 825 million shares at $70 to $78 each. As of October 31, the company's annualized revenue was $81 billion, up $10 billion from the end of September; cash and equivalents exceed $130 billion. Goldman Sachs expects it to achieve positive annual free cash flow only in 2030; spending commitments by 2045 exceed $900 billion, about $71 billion of which comes due before the end of September 2027. Last Friday, CEO Amodei and JPMorgan CEO Dimon appeared together at an IPO event in New York to pitch the deal; people familiar with the matter say Anthropic will not allow institutional investors from mainland China and Hong Kong to participate in IPO subscriptions. Earlier, Jensen Huang's reputation crisis caused Nvidia shares to slump, and Anthropic is expected to succeed Nvidia as the new AI benchmark. @@@ The Verge reports that Bloomberg's Mark Gurman says Apple plans to unveil its first smart glasses at its Worldwide Developers Conference in June 2027 and begin shipping them by the end of that year. The product will not have augmented reality features; its core selling points are the new Siri AI and the iPhone app ecosystem. The prototype is codenamed N50, currently runs a stripped-down version of visionOS, and uses a new processor based on an older smartwatch chip. The glasses come with a flat rectangular MagSafe-like magnetic charger about the size of a Tic Tac, and users can charge them while wearing them. Apple also emphasizes design, privacy and audio quality. Employees say its audio quality is industry-leading, and it supports calls, music playback and FaceTime audio. The glasses have Wi-Fi; after leaving the iPhone, some non-third-party app functions can still be used, such as continuing to listen to audio via Wi-Fi, but it will not be a fully independent device. The new Siri will add granular app controls; Apple is first having developers adapt on the iPhone so that related functions can be used directly once the glasses launch. @@@ The Verge reports that United Launch Alliance's Vulcan rocket experienced an anomaly during a launch and failed to send 40 Amazon Leo satellites into orbit. ULA said in a statement on X that, according to data received by the launch director, the rocket, which lifted off from Space Launch Complex 41, experienced an anomaly 10 minutes after liftoff, causing the Amazon mission to fail to reach orbit; the cause is not yet known, and the company promised transparency during the investigation. This is reportedly ULA's first mission failure in its 20-year history. The launch was originally planned as the first of 38 launches Amazon purchased from ULA. Leo is Amazon's satellite project aimed at competing with SpaceX's Starlink, but it has struggled, with only a small number of the needed satellites in orbit so far. Amazon says any launch carries risk, especially a first mission; ULA has always been a trusted partner, and it believes ULA will learn from this mission and carry out Vulcan's long-term plan. @@@ TechWeb, citing Kuaikeji, reports that on November 14, Judge Noelle Wise of the U.S. District Court for the Northern District of California denied in full motions to dismiss by Samsung, Micron and SK Hynix, finding that the plaintiffs have plausibly alleged that the three violated Section 1 of the Sherman Act and reached an illegal agreement, moving the DRAM class-action antitrust lawsuit into substantive proceedings. The lawsuit was filed in June 2026 by 17 consumers and small businesses in five U.S. states, alleging that the three coordinated to restrict conventional DRAM supply and push up prices. The complaint says that as of the third quarter of 2025, the three together accounted for more than 91% of global DRAM revenue; since late 2022, they synchronized cuts to conventional consumer DRAM capacity, and in 2023 shifted substantial wafer capacity to HBM, with conventional DRAM prices cumulatively rising about 697% in 18 months. The court listed synchronized conduct including Samsung and SK Hynix refusing long-term supply agreements, the three delaying or halting new fabs, cutting DDR4 production in the same quarter, Micron closing its 30-year-old Crucial brand in December 2025, launching highly similar customer audit mechanisms in January 2026, and mentioned that they were involved in a price-fixing case from 1998 to 2002. The three companies say their decisions were adapting to changes in the AI market and that similar disputes were dismissed in 2018; the judge found the facts and evidence in this case fundamentally different. The three must submit complete answer documents by December 7, 2026, and the first management conference is set for January 12, 2027. @@@ Sina Tech reports that on November 16, according to Qwen, Alibaba released its new-generation foundation model Qwen4-Max, with 2.4 trillion total parameters, native vision capabilities and support for a million-token context. Based on a new architecture and training optimizations, the model continues to evolve in key capabilities such as agents and coding, and in professional verticals such as finance and law it moves from "information retrieval" to "professional delivery." During the launch period, Qwen app members can try it free for a limited time, and using Qwen4-Max does not consume membership quota. Related reports show that its preview version first launched on November 11 on the Qwen AI platform Token Plan and was simultaneously integrated into Qwen Office, Qoder, the Qwen app and QwenStudio; the preview window runs until 24:00 on November 25, mainly for capability validation and collecting feedback. The Paper also reports that Qwen4-Max is now available via API on the Qwen AI platform, with cache hits as low as RMB0.5 per million tokens, and the average cost per completed task down 60% from the previous generation; users can also call it through Alibaba Token Plan, and the small- and medium-sized models in the Qwen4 series will be open-sourced successively. @@@ Sina Tech reports that an article from Tencent Culture disclosed that on its 28th birthday, Tencent turned its annual Executive Office LinkTime into a special podcast, with the Executive Office speaking out on the opportunities, challenges and choices in the AI era. The remarks proposed that all core business teams must start a second venture for the AI era: AI-native businesses, models and products must go upstream against the current; half of the people in core businesses should defend the core, while the other half should fight new battles; the organization should be flattened and operate in small teams, breaking down fiefdoms and barriers of middle-layer relayed messaging. Tencent said it wants to build human-centered AI, adhering to benefiting humanity and society, being controlled by humans and empowering creativity; AI is still in its early stages, compute infrastructure must be self-built and prices are soaring, and investment will balance input and output efficiency. Its AI pyramid is centered on models and has entered China's first tier; Xiaowei is positioned as a personal AI assistant with the goal of "token freedom"; Yuanbao will be positioned as a decision-making and information adviser; WorkBuddy became the single AI product in China providing the most tokens within eight months. The Executive Office also required a zero-based mindset, faster feedback, decision-making and action, and higher talent density, and mentioned that teams in gaming, finance, security and other areas have already co-designed with models, bringing improvements in Hy capabilities. @@@ Sina Tech reports that Huawei Terminal announced on the morning of November 16 that HarmonyOS terminal devices have exceeded 100 million units. Previously, Huawei's Executive Director and Chairman of the Terminal BG Yu Chengdong announced via social media that the new sub-brand OSTA Xingmang will officially debut on November 20 at the "HarmonyOS Intelligent Mobility Stelato V8 and Xingmang Brand Launch Event." Reports say Xingmang will share HarmonyOS, Kirin chip technology, and Huawei's online and offline channel and after-sales systems. @@@ Sina Tech reports that on November 16, due to tight supply of compute resources, DeepSeek introduced usage limits for its consumer app and web chat services: long conversations that have lasted more than seven days without active interaction can no longer be continued; single input text is capped at about 40,000 characters; when the daily chat quota is exhausted, users will see the prompt "Server busy, usage limit reached." Analysts believe the move is intended to reclaim KV cache occupied by idle sessions and prioritize inference for active users. Since October, its services have come under pressure multiple times; the official status page has recorded eight incidents, occurring on October 2, October 8, October 12 (twice), October 15, October 20, October 27 and November 3, mostly performance degradation such as slower responses; the October 15 incident lasted more than four hours, with consumer services and the V4 Pro and V4.1 Flash APIs simultaneously restricted, and API users were later compensated with credits. From August to October, V4.1 Flash API availability was 99.46%, and V4 Pro was 99.74%. DeepSeek recently opened APIs for its full suite of applications, with time-of-use peak/off-peak pricing and separate concurrency limits on the API side. OpenRouter's weekly ranking shows that from October 19 to 25, V4.1-Flash-1017 ranked first with 33.6 trillion tokens in weekly calls, up 801% week-on-week. @@@ Sina Tech reports that on November 12, xFusion's ChiNext IPO was approved at the listing review meeting. It took only 174 days from acceptance on May 22 to approval, and its RMB8 billion fundraising makes it the largest ChiNext IPO project this year. The company's revenue rose from RMB25.092 billion in 2023 to RMB58.246 billion in 2025, and was RMB37.286 billion in the first half of 2026, up 20.78% year-on-year; AI server revenue share rose from 25.03% to 53.29%. However, its main business gross margin fell from 14.27% to 8.58% in 2025, with a net margin of 1.77%; in the first half of 2026, gross margin rebounded to 15.49%, and net profit was RMB2.108 billion, more than twice the full-year 2025 figure. Its 2025 R&D expense ratio was 2.02%, lower than Sugon, Unisplendour and Inspur. In the first half of 2026, net operating cash flow was -RMB15.685 billion; accounts receivable were RMB15.303 billion, up more than 170% from the end of 2025; inventory was RMB35.662 billion, up 68%. As of the end of July 2026, its backlog was RMB30.113 billion, of which AI servers accounted for 73.70%. @@@ All right, that's all for today's show! Thank you for spending time with us. I hope you're in a great mood tonight. If you like our show, you're very welcome to subscribe and share it with your friends—it really helps us. See you next time!

13 分钟 34 次播放 2天前

《晚安咖啡》科技夜话(英文版):字节4000亿估值回购,DeepSeek开放API,SpaceX捕获星舰

《晚安咖啡》科技夜话(英文版):字节4000亿估值回购,DeepSeek开放API,SpaceX捕获星舰

晚安咖啡GoodNightCoffee

Hi there~~ Good evening, everyone! Today is Sunday, November 15, 2026. I’m An’an, your host on Goodnight Coffee~ Tonight, let’s catch up on the latest business and tech headlines together—plenty of useful takeaways, and a relaxed, easy listen. So find a comfortable spot, and let’s get tonight’s show started! @@@ According to TechWeb, citing foreign media, ByteDance plans to buy back up to $6 billion worth of shares from existing investors at a valuation of about $400 billion. The buyback price is $241.35 per share, up from $200.41 in the previous buyback last November. External shareholders have the right to require ByteDance to repurchase all of their holdings by a specified date. In its latest shareholder letter, ByteDance proposed extending that deadline again, from December 11 this year to December 11, 2030. ByteDance also said it plans to offer employees working on AI a total of 300 million shares tied to the value of the company’s AI business, as a way to reward and retain AI talent. Earlier reports showed that in the first half of 2026, ByteDance’s revenue was about $120 billion, up 30% year on year, with net profit of about $20 billion. The company also recently secured about $30 billion in loans from the banking system—its largest single loan to date. @@@ According to QbitAI, DeepSeek has announced that it has officially opened up the API for its “full suite” of applications. Web search, deep-thinking search, file upload, link reading, speech recognition, and speech synthesis all now have standalone interfaces and can be called and billed separately. Voice capabilities are being opened up for the first time. Developers can combine them freely, or even recreate the complete DeepSeek App one-to-one. DeepSeek also open-sourced a lightweight chat front-end demo that supports importing chat history. Previously, Search and Fetch, opened on October 17, required developers to chain tool calls themselves. Now, once web search is enabled, DeepSeek can automatically decide when to search, what to search for, which web pages to read, and how to answer follow-up questions. V4.1-Flash already supports these conversation extensions through the Messages, Responses, and Chat Completions APIs. On pricing, speech recognition costs 0.78 yuan per hour at the original price, or 0.39 yuan at 50% off—about 1/4.6 of OpenAI GPT-Transcribe, 1/3.1 of GPT-4o mini Transcribe, and 1/16.5 of Google Cloud’s standard service. Speech synthesis drops from 3 yuan to 1.5 yuan per 10,000 characters; search and web scraping cost 5 yuan per 1,000 successful calls; and OCR costs 1.5 yuan per 1,000 pages. The announcement said the existing conversational product is only a relatively basic form. @@@ According to Readhub, Huawei Executive Director Richard Yu released a video reviewing the development of Huawei’s Mate X series of foldable phones and announced that the Mate X8 is coming soon. At the end of the video, the new phone’s appearance was revealed. Yu said the new device has breakthroughs in the display, camera, and performance. Footage circulating online shows Yu holding a purple new phone believed to be the actual Mate X8, featuring a circular camera module. Related leaks say the Mate X8 is a large foldable with a conventional aspect ratio, equipped with a new 8-inch wide-color-gamut BOE display and ultra-flat crease technology, a rear 50MP 18EV ultra-large-sensor flagship camera system, an “N-generation Kirin chip,” and a large circular deco with a jewel-ring design. @@@ Readhub reports that the large unmanned cargo aircraft W5000 has successfully completed its test flight. The aircraft was developed and manufactured in Changzhou, Jiangsu. All 220,000 of its components are domestically produced. It has a maximum payload of 5 tons and a maximum range of 2,600 kilometers. So far, W5000 has received 835 intended orders, with customers including several domestic logistics companies. According to plan, the aircraft is expected to enter commercial operation by the end of 2027. @@@ According to Huxiu 24 Hours, 12 drugs from Hengrui Pharmaceuticals have been included in China’s 2026 national medical insurance drug list, three of them for the first time. @@@ According to The Verge, SpaceX has for the first time successfully caught the Starship spacecraft itself using the robotic arms on its launch and catch tower, “Mechazilla.” The company had previously caught boosters multiple times using the “chopsticks” method, but this was the first time it completed the maneuver with the Starship vehicle. During this mission, SpaceX originally planned to catch the booster on Starship’s 15th flight, but the booster aborted its landing and then exploded over the Gulf of Mexico; the company is still investigating the cause. The report says that because the entire launch vehicle was not recovered, this was not a complete success, but it is still considered a milestone for the Starship program. Video shows Starship descending through the clouds, flipping vertically, and entering smoothly between Mechazilla’s waiting robotic arms at about 4:20 p.m. ET. SpaceX CEO Elon Musk said on the third-quarter earnings call that these first flights are not guaranteed to succeed, but that the next flight is more likely than not to succeed. @@@ According to Readhub, Nvidia’s board has issued a formal response to outside attention on the CEO candidate issue. The board said it is currently in the process of selecting an executive search firm and will launch recruitment for the company’s CEO position. The board explicitly stressed that, as of now, it has not had formal contact with any potential candidate, and speculation about any individual internal or external candidate is premature. @@@ According to Readhub, Italy has issued an €805 million fine against Meta. @@@ cnBeta reports that Google has launched a personal financial analysis feature for Gemini. The feature can connect to bank accounts and credit cards to help users analyze their personal income and expenses. @@@ According to The Verge, Tesla has invited “select customers” to buy the two-seat, steering-wheel-free Cybercab for personal use at $29,990, but they must also subscribe to a $299-per-month “Autonomy Subscription.” The vehicle is only being sold in “supported service areas” covered by Tesla’s Robotaxi service, with limited initial supply. Buyers must also pay a $1,000 non-refundable deposit to secure a December delivery window, though delivery is not guaranteed. If the subscription is canceled, because the vehicle has no manual-driving fallback, the Cybercab will be completely unusable on public roads. The order terms also prohibit commercial use: it cannot be used for Uber or DoorDash, nor can it be connected to Tesla’s Robotaxi network to offset the subscription fee. Over a five-year ownership period, the mandatory subscription adds $17,940, bringing the base cost to about $47,600 before taxes, charging, and insurance—higher than a new Model 3 or Model Y. Tesla began offering Cybercab rides in its Austin Robotaxi service in September; according to Robotaxi Tracker, 278 Cybercabs are currently in operation, accounting for about 40% of the fleet. Tesla AI head Ashok Elluswamy has suggested that customer vehicles may be allowed to join the network in the second quarter of 2027. @@@ All right, that wraps up today’s show! Thank you for spending this time with us. May you carry a good mood with you as you keep moving forward. If you enjoy our program, we’d really appreciate it if you subscribe and share it with your friends—it helps us a lot. See you next time!

8 分钟 37 次播放 3天前

《晚安咖啡》科技夜话:字节4000亿估值回购,DeepSeek开放API,SpaceX捕获星舰

《晚安咖啡》科技夜话:字节4000亿估值回购,DeepSeek开放API,SpaceX捕获星舰

晚安咖啡GoodNightCoffee

嗨~~大家晚上好呀!今天是2026年11月15日,星期日,我是晚安咖啡的主播安安~今天咱们一起聊聊新鲜的商业科技动态,干货满满、轻松开听,快找个舒服的姿势,和我一起开启今晚的分享吧! @@@ 据TechWeb援引外媒消息,字节跳动计划以约4000亿美元估值,从现有投资者手中回购不超过60亿美元股份。此次回购价格为每股241.35美元,高于去年11月上一轮回购的每股200.41美元。外部股东有权要求字节跳动在特定日期前回购其全部持股,字节跳动在最新股东信中提议,将这一期限从今年12月11日再次延长至2030年12月11日。字节跳动还表示,计划向从事AI业务的员工提供合计3亿股与公司AI业务价值挂钩的股份,以奖励并留住AI人才。此前消息显示,2026年上半年,字节跳动营收约1200亿美元,同比增长30%,净利润约200亿美元;近期公司还从银行体系获得约300亿美元贷款,为其历史上规模最大的单笔贷款。 @@@ 据量子位报道,DeepSeek宣布正式开放“全家桶”应用API,联网搜索、深度思考搜索、文件上传、链接阅读及语音识别、语音合成等均提供独立接口,支持单独调用、分别计费。语音能力系首次开放;开发者可自由组合,甚至1:1复刻完整DeepSeek App。DeepSeek还开源轻量级聊天前端Demo,支持导入历史对话。此前10月17日开放的Search和Fetch需开发者自行串联工具调用,如今打开联网搜索后,何时搜索、搜什么、读取哪些网页及后续回答可由DeepSeek自动完成;V4.1-Flash已在Messages、Responses和Chat Completions接口下支持相关对话扩展。价格上,语音识别原价0.78元/小时,五折后0.39元,约为OpenAI GPT-Transcribe、GPT-4o mini Transcribe和谷歌云标准服务价格的1/4.6、1/3.1和1/16.5;语音合成每万字符由3元降至1.5元;搜索和网页抓取每1000次成功调用5元,OCR每1000页1.5元。公告称现有对话产品只是较基础的一种形态。 @@@ 据Readhub报道,华为常务董事余承东发布视频回顾华为Mate X系列折叠屏手机发展历程,并宣布Mate X8即将推出。视频结尾公开了新机外观,余承东称新机在屏幕、拍照、性能等方面有新突破。流传画面显示,余承东手持一台紫色新机,被认为是Mate X8真机,其采用圆形镜头模组。相关爆料称,Mate X8为常规比例大折叠产品,搭载8英寸广色域京东方新屏与超平整折痕技术,后置50MP、18EV超大底旗舰影像,配备“N代麒麟芯”,并采用大圆Deco搭配宝石环设计。 @@@ Readhub消息,大型无人货运飞机W5000已顺利完成试飞。该飞机由江苏常州研发制造,整机22万个零部件全部实现国产化,最大载重5吨,最远可飞2600公里。截至目前,W5000意向订单已达835架,客户覆盖国内多家物流企业。按计划,该飞机预计于2027年底正式投入商业运营。 @@@ 据虎嗅24小时报道,恒瑞医药有12款药品纳入2026年国家医保目录,其中3款为首次纳入。 @@@ 据The Verge报道,SpaceX首次使用发射与捕获塔“Mechazilla”的机械臂成功捕获星舰(Starship)飞行器本身。此前该公司已多次以“筷子”方式捕获助推器,但这是首次对星舰飞行器完成这一操作。此次任务中,SpaceX原计划捕获星舰第15次飞行任务的助推器,但助推器中止着陆,随后在墨西哥湾上空爆炸,公司仍在调查原因。报道称,由于未能回收整个运载器,这并非完全成功,但仍被视为星舰项目的里程碑。视频显示,星舰穿云而下、垂直翻转,于美东时间下午约4点20分平稳进入Mechazilla等待的机械臂之间。SpaceX CEO马斯克曾在第三季度财报电话会上表示,这些首飞不保证成功,但下一次飞行成功的可能性大于不成功。 @@@ 据Readhub报道,英伟达董事会就外界关注的CEO人选问题作出正式回应。董事会称,目前正处于遴选高管猎头公司的阶段,将启动公司首席执行官职位的招聘工作。董事会明确强调,截至目前尚未与任何潜在候选人进行正式接触,现阶段外界关于任何内部或外部个别候选人的猜测都为时尚早。 @@@ 据Readhub消息,意大利对Meta开出8.05亿欧元罚单。 @@@ cnBeta消息,Google为Gemini推出个人财务分析功能,该功能可接入银行账户和信用卡,帮助用户分析个人收支情况。 @@@ 据The Verge报道,特斯拉已邀请“精选客户”以29,990美元购买两座、无方向盘的Cybercab用于个人使用,但须同时订阅每月299美元的“Autonomy Subscription”。该车仅在特斯拉Robotaxi服务覆盖的“支持服务区域”销售,初期供应有限;买家还需支付1,000美元不可退还定金,争取12月交付窗口,但不保证交付。取消订阅后,因车辆没有手动驾驶后备方案,Cybercab在公共道路上将完全无法使用。订单条款还禁止商业用途,不能用于Uber、DoorDash,也不能接入特斯拉Robotaxi网络抵消订阅费。按五年持有期计算,强制订阅费增加17,940美元,使税前、充电和保险前的基础成本约达47,600美元,高于全新Model 3或Model Y。特斯拉9月在奥斯汀Robotaxi服务中开始提供Cybercab乘坐;据Robotaxi Tracker,目前278辆Cybercab运营,约占车队40%。特斯拉AI负责人埃卢斯瓦米曾提出,可能在2027年第二季度允许客户车辆加入该网络。 @@@ 好啦,今天的节目到这里就结束啦!谢谢你的陪伴,愿你带着好心情继续向前。如果您喜欢我们的节目,也非常欢迎订阅和转发给你的小伙伴,这对我们非常有帮助。那我们下期再见吧!

7 分钟 73 次播放 3天前

《晚安咖啡》科技夜话(英文版):黄仁勋辞任英伟达CEO,微软云服务全球宕机,谷歌发布Gemini 4 Hydrogen

《晚安咖啡》科技夜话(英文版):黄仁勋辞任英伟达CEO,微软云服务全球宕机,谷歌发布Gemini 4 Hydrogen

晚安咖啡GoodNightCoffee

Hi there, and good evening, everyone! Today is Saturday, November 14, 2026. Welcome to Goodnight Coffee. I’m your host, An’an! Tonight, let’s talk about the freshest, piping-hot business and tech updates—relaxed, engaging, and packed with insight. So put on your headphones and let’s get tonight’s show started! @@@ According to QbitAI, Nvidia has announced that Jensen Huang has resigned as the company’s president, CEO, and director, effective immediately. Colette Kress, currently executive vice president and CFO, will serve as interim CEO while retaining her current role until a permanent successor is appointed. After the news broke, Nvidia shares in premarket trading fell as low as $150.13, down about 13.58% from the previous close; after the open, the decline remained above 10%, before narrowing to 4.63% by press time. Jensen Huang co-founded Nvidia in 1993 and led the company for 33 years. The board has launched a formal CEO search and will consider both internal and external candidates; it has also appointed Harvey C. Jones as the first board chair in the company’s history. He previously served as CEO and executive chairman of Synopsys. Kress has been with Nvidia for more than 13 years, and the announcement said she has played a key role in the company’s growth and transformation. On January 27, 2025, when DeepSeek triggered a reassessment of AI compute spending, Nvidia fell about 17% in a single day and lost roughly $593 billion in market value—at the time the largest single-day market-value loss for a U.S. company. The board expects the company’s operations, customer commitments, and product roadmap to continue as planned, and it will report third-quarter fiscal 2027 results and hold its earnings call on November 17 as scheduled. @@@ According to Huxiu 24 Hours, OpenAI has disbanded its standalone robotics team and is instead betting on general-purpose large models. @@@ According to Sina Tech, SpaceX completed the 15th test flight of Starship on Friday but failed to recover either of the two stages. The rocket lifted off from a launch site in southern Texas at 4:47 p.m. local time, 5:47 p.m. Eastern. This was the most ambitious Starship mission to date, with the company attempting for the first time to catch both the booster and the upper stage. After the booster catch was called off, it made a controlled splashdown in the Gulf of Mexico and caught fire on the water. Earlier that day, SpaceX received a launch license from the Federal Aviation Administration. Hours later, the agency issued a nationwide ground stop due to communications failures at multiple air traffic control centers, but said the license was unaffected. After the booster catch failed, the upper stage continued along its planned trajectory, with plans to orbit Earth 14 times and, if all went well, land at Starbase on Saturday. SpaceX said the mission deployed 26 Starlink V3 satellites; its constellation already has more than 11,000 active low-Earth-orbit satellites. A single Starship can carry up to 60 V3 satellites, with roughly 20 times the capacity of Falcon 9, aiming for Starlink to carry 51% of global internet traffic. William Blair analysts said that even with the failed catch this time, a delay of a few months would have limited impact. In addition, Zelensky said Ukraine’s military Starshield service was again restricted without warning, and several U.S. senators condemned the move and called for a congressional investigation. @@@ According to Sina Tech, Microsoft is gradually recovering from an outage that lasted all day. The disruption began on Thursday and initially affected only the U.S. and the Americas, but as engineers struggled to restore downed servers, the problem quickly spread worldwide. It grew from outages in Outlook email and Teams video meetings to personal and enterprise cloud services such as email, file storage, and document storage. Copilot AI tools, Defender security software, and the Intune management console were also down for a time, and some users reported being unable to access cloud Windows virtual machines. Microsoft has not announced a root cause, saying only that the incident involved “a large number of servers unexpectedly unavailable,” and that the investigation is ongoing. By Friday morning on the U.S. East Coast, Outlook began to recover and some emails started arriving. Doug Madory, director of internet analysis at Kentik, said that after about 18 hours of downtime, traffic to Microsoft’s services began to recover. In a Friday status update, Microsoft said there was no evidence that the incident was related to a malicious attack. When media asked whether Microsoft had logs or other technical means to detect malicious activity and whether it could reconstruct the incident, its spokesperson had not yet responded. The last major Microsoft cloud outage occurred in September, which Microsoft attributed to a “core identity authentication configuration issue” that prevented users from signing in to services. @@@ According to 36Kr, on November 12, Google released Gemini 4 Hydrogen and rolled it out globally, replacing Gemini 3.8 Flash as the series’ “efficiency execution layer.” It launched fully the same day across the Gemini app, API, enterprise edition, and Antigravity; Search AI Mode made it the default model. Artificial Analysis shows that its default high setting scores 51, surpassing the flagship Argon’s 50; compared with 3.8 Flash, it is 2.5 times faster, Terminal-Bench 4.1 rises from 18.2% to 57%, and hallucinations are reduced by more than fivefold. API pricing is $1.5 per million input tokens and $7.5 per million output tokens. The Orchestrator can switch between Carbon and Hydrogen within the same session; Google says this saves 24% of quota and is 19% faster. This is the third Gemini 4 model in six weeks. Its price is double that of 3.8 Flash, and the new pricing will apply uniformly after the old promotional pricing ends on December 31. @@@ According to TechWeb, on November 14, DeepSeek released its “full-suite” application API, packaging the complete set of mature interaction capabilities from its own app into standardized modules for developers. Unlike previous offerings that only provided model inference interfaces, this release covers application-layer capabilities such as search, web crawling and parsing, file parsing, OCR, speech recognition, and speech synthesis, which can be called individually or embedded in combination into workflows. By enabling the corresponding parameters in the existing API, developers can get conversational features such as deep-thinking search and file input, with tool calls and context assembly handled automatically on the server side. Currently, the V4.1-Flash model, deepseek-flash, supports web search, deep-thinking search, file input, link reading, and model safety in the Messages, Responses, and Chat Completions interfaces. Web search can automatically initiate webpage or image searches and answer with citations; deep-thinking search can independently go online to read links during multi-step reasoning; file input supports parsing formats such as PDF, Office, and images; and safety checks can review model inputs and outputs and return refusal markers for risks. Capabilities such as Search, Fetch, OCR, ASR, and TTS can also be called independently. DeepSeek said that in addition to token fees, it will charge tool-call fees for the extended capabilities; those providing application services externally must enable safety parameters and assume obligations such as content safety review, large-model registration, and AI labeling. @@@ According to Readhub, Moonshot AI, the developer of the Kimi large model, is considering an IPO in Hong Kong as early as early 2027 and has begun dismantling its overseas holding structure to pave the way. The restructuring path is as follows: first, the company will repurchase overseas shares from U.S. dollar-denominated shareholders, then arrange for the relevant funds to be invested in the domestic entity. Previously, Moonshot AI completed its final private funding round at a valuation of about $50 billion; the company is considering raising up to $5 billion in the IPO and has confidentially submitted a listing application to the Hong Kong Stock Exchange. Moonshot AI said it does not comment on market rumors or speculation. The company was founded by Yang Zhilin and focuses on the Kimi series of large models. If it completes a listing at a $50 billion valuation, it would be one of the largest IPOs among Chinese large-model companies. @@@ According to Readhub, China’s 2026 national medical insurance drug catalog and commercial insurance innovative drug catalog were announced. A total of two innovative drugs were transferred from the first commercial insurance innovative drug catalog into basic medical insurance, further deepening the coordination and linkage between the two catalogs. Among them, JW Therapeutics’ relma-cel injection was included in basic medical insurance for the first time. It is a million-yuan-class CAR-T cancer therapy, and after medical insurance reimbursement, patients’ out-of-pocket costs can fall to under 100,000 yuan. The core reason the drug entered basic medical insurance with a significant price cut is that key raw materials have been localized, greatly reducing production costs. @@@ According to cnBeta, China’s 10-ton-class electric commercial aircraft CE-25A has successfully completed its first piloted flight. @@@ According to Huxiu 24 Hours, the first meeting of the Trade Council under the China-U.S. economic and trade consultation mechanism has been held. @@@ Alright, that’s all for today’s show! Thank you for spending time with me today. May you rest easy and sleep well. If you enjoy our program, you’re very welcome to subscribe and share it with your friends—it really helps us a lot. See you next time!

10 分钟 33 次播放 4天前

《晚安咖啡》科技夜话:黄仁勋辞任英伟达CEO,微软云服务全球宕机,谷歌发布Gemini 4 Hydrogen

《晚安咖啡》科技夜话:黄仁勋辞任英伟达CEO,微软云服务全球宕机,谷歌发布Gemini 4 Hydrogen

晚安咖啡GoodNightCoffee

嗨~~大家晚上好呀!今天是2026年11月14日,星期六,欢迎来到《晚安咖啡》,我是主播安安!今天咱们一起聊聊新鲜热乎的商业科技动态,轻松又有料,快戴上耳机,和我一起开启今晚的分享吧! @@@ 据量子位消息,英伟达宣布黄仁勋已辞去公司总裁、首席执行官及董事职务,即刻生效。现任执行副总裁兼CFO Colette Kress出任临时CEO,并继续兼任原职,直至正式继任者获任命。消息公布后,英伟达盘前股价一度跌至150.13美元,较前一交易日收盘下跌约13.58%;开盘后跌幅仍超10%,发稿前收窄至4.63%。黄仁勋1993年联合创办英伟达,掌舵33年。董事会已启动正式CEO遴选,将同时考虑内外候选人;并任命Harvey C. Jones为公司历史上首位董事会主席,他曾任新思科技CEO和执行董事长。Kress在英伟达工作超过13年,公告称她在公司增长与转型中发挥关键作用。2025年1月27日DeepSeek引发AI算力投入重估时,英伟达曾单日收跌约17%,市值蒸发约5930亿美元,创当时美股单家公司单日市值损失纪录。董事会预计公司运营、客户承诺及产品路线图将继续推进,并按原定安排于11月17日公布2027财年第三季度业绩、召开财报电话会议。 @@@ 据虎嗅24小时消息,OpenAI已解散独立机器人团队,转而押注通用大模型。 @@@ 新浪科技消息,SpaceX周五完成星舰第15次试飞,但未能回收两级箭体。火箭于当地时间下午4时47分(美国东部时间下午5时47分)从得州南部发射场升空。此次是迄今目标最高的星舰任务,公司首次尝试同时捕获助推器与上面级;助推器捕获取消后受控溅落墨西哥湾,并在海面起火。当天早些时候SpaceX获美国联邦航空管理局发射许可,数小时后该局因多个空管中心通信故障发布全国停飞指令,但称该许可未受影响。助推器捕获失败后,上面级继续沿预定轨道飞行,计划环绕地球14圈,若顺利周六降落在星基地发射场。SpaceX称本次释放26颗星链V3卫星;其星座已有超过11000颗在轨活跃低轨卫星,一枚星舰最多可搭载60颗V3卫星,运力约为猎鹰9号20倍,目标是让星链承载全球51%互联网流量。威廉·布莱尔分析师称,即便本次捕获失败,推迟数月影响有限。此外,泽连斯基称乌军星盾服务再遭无预警限制,多名美国参议员谴责并呼吁国会调查。 @@@ 据新浪科技报道,微软正从持续一整天的服务中断中逐步恢复。故障始于周四,最初仅影响美国及美洲区域,但因工程师难以恢复宕机服务器,问题很快蔓延至全球;影响从Outlook邮件、Teams视频会议无法使用,扩大到邮件、文件、文档存储等个人与企业云服务,Copilot AI工具、Defender安全检测软件及Intune管理控制台也一度瘫痪,另有用户反馈云Windows虚拟机无法访问。微软尚未公布根因,仅称事故是“大量服务器意外不可用”,调查仍在推进。美国东海岸周五上午,Outlook开始恢复,部分邮件陆续送达;Kentik互联网分析总监道格·马多里称,经过约18小时停机后,微软各项服务流量开始回升。微软在周五状态公告中表示,暂无证据表明事件与恶意攻击有关。媒体询问微软是否具备日志等技术手段侦测恶意行为、能否复盘事故经过,其发言人暂未回应。上一次微软大规模云服务事故发生在9月,微软归因于“核心身份验证配置问题”,导致用户无法登录服务。 @@@ 据36氪报道,11月12日谷歌发布Gemini 4 Hydrogen并全球推送,接替Gemini 3.8 Flash成为系列“效率执行层”。它当天在Gemini应用、API、企业版和Antigravity等全量上线;搜索AI Mode将其设为默认模型。Artificial Analysis显示,其默认high档51分,超过旗舰Argon的50分;相较3.8 Flash,速度快2.5倍,Terminal-Bench 4.1从18.2%升至57%,幻觉降低5倍以上。API定价每百万输入/输出1.5/7.5美元。Orchestrator可在同一会话切换Carbon与Hydrogen,谷歌称省24%配额、快19%。这是六周内第三款Gemini 4,其价格较3.8 Flash翻倍,旧优惠价12月31日结束后统一新价。 @@@ TechWeb消息,11月14日,DeepSeek发布“全家桶”应用API,将自有App中已成熟的整套交互能力打包为标准化模块向开发者开放。与以往仅提供模型推理接口不同,此次覆盖搜索、网页抓取与解析、文件解析、OCR、语音识别、语音合成等应用层能力,支持单独调用或组合嵌入工作流。开发者在原有API中开启对应参数,即可获得深度思考搜索、文件输入等对话功能,工具调用与上下文拼接由服务端自动完成。目前,V4.1-Flash模型(deepseek-flash)已在Messages、Responses和Chat Completions接口支持联网搜索、深度思考搜索、文件输入、阅读链接及模型安全;联网搜索可自动发起网页或图片搜索并附引用作答,深度思考搜索可在多步推理中自主联网阅读链接,文件输入支持PDF、Office、图片等格式解析,安全校验可审核模型输入输出并对风险返回拒答标识。Search、Fetch、OCR、ASR、TTS等能力也可独立调用。DeepSeek表示,扩展能力除Token费用外还将收取工具调用费用;对外提供应用服务须开启安全参数,并承担内容安全审核、大模型登记、AI标识等义务。 @@@ 据Readhub报道,Kimi大模型开发商月之暗面正考虑最早于2027年初在香港上市,并已着手拆除海外控股架构,为IPO铺路。相关重组路径为:先由公司从美元股东手中回购海外股份,再安排相关资金投资境内主体。此前,月之暗面已完成最后一轮私募融资,估值约500亿美元;公司正考虑在IPO中募资最多50亿美元,并已以保密形式向港交所递交上市申请。月之暗面方面表示,不对市场传言或猜测置评。该公司由杨植麟创立,主推Kimi系列大模型。若按500亿美元估值完成上市,这将成为中国大模型公司中规模最大的IPO之一。 @@@ 据Readhub消息,2026年国家医保药品目录及商保创新药目录公布,共有2款创新药从首版商保创新药目录转入基本医保,医保与商保双目录协同衔接进一步深化。其中,药明巨诺的瑞基奥仑赛注射液首次纳入基本医保,这是一款百万级CAR-T抗癌疗法,患者经医保报销后自付费用可降至10万元以内。该药此次大幅降价进入基本医保,核心原因是关键原材料实现国产化,生产成本因此大幅压缩。 @@@ 据cnBeta消息,中国10吨级电动商用飞机CE-25A已完成有人驾驶首飞,首飞取得成功。 @@@ 据虎嗅24小时消息,中美经贸磋商机制项下贸易理事会首次会议已举行。 @@@ 好啦,今天的节目到这里就结束啦!谢谢你今天的陪伴,愿你安心入睡。如果您喜欢我们的节目,也非常欢迎订阅和转发给你的小伙伴,这对我们非常有帮助。那我们下期再见吧!

9 分钟 61 次播放 4天前

《晚安咖啡》科技夜话(英文版):黄仁勋辞职,Outlook故障

《晚安咖啡》科技夜话(英文版):黄仁勋辞职,Outlook故障

晚安咖啡GoodNightCoffee

Hi~~ good evening, everyone! It's Friday, November 13, 2026. I'm An'an, your host at Goodnight Coffee, and I'm super happy to be with you again! Today I've got a fresh batch of business and tech updates—plenty of substance, easy to follow. Let's jump right into today's show! @@@ According to CNBC, Nvidia is nearing its longest weekly losing streak since 2012. As of Thursday's close, the stock had fallen 11% this week and 4.3% on Thursday, on track for a sixth consecutive weekly decline. In this round, its market value has shrunk by 29%, or about $1.6 trillion, and it has fallen below Apple. On the same day, the board announced that co-founder and CEO Jensen Huang had resigned, appointed CFO Colette Kress as interim CEO, and is searching for a successor. Earlier, two extramarital affairs involving Huang had been made public: one involved a person alleged to be a Chinese spy, and the other involved the company's vice president of its healthcare business. He wrote to employees over the weekend apologizing for being "unfaithful to his wife." Before Huang's departure last week, Nvidia shares had already fallen about 9%; after the midterm elections, a decisive Democratic victory and the passage of several ballot measures that could slow AI construction deepened the selloff. Nvidia plans to report quarterly results next Tuesday, with analysts expecting revenue to rise 91% year on year to $108.9 billion, 18 times the same period in 2022. Oppenheimer said Huang's departure could create a persistent share overhang, while D.A. Davidson cut its price target from $300 to $200. @@@ According to The Verge, Microsoft's Outlook email service has still not recovered. Since early Thursday afternoon, users in many parts of the world have reported being unable to access Outlook mailboxes via the web or the app. Microsoft later acknowledged the issue on its Microsoft 365 Status account, saying some infrastructure was experiencing elevated load in multiple regions, which could cause connection errors; Teams also had reports of problems, though less severe. Microsoft said the impact involves the Microsoft 365 commercial service suite, mainly Exchange Online, and that some users may be unable to access mailboxes, with email delivery possibly delayed. Microsoft Communications Director Ben Hope said the company was aware of the issue and investigating, thanked customers for their patience, and would restore service as soon as possible. Several hours later, Microsoft updated its progress multiple times, saying it was encountering difficulties restarting servers and monitoring server health. By about 6 p.m. Eastern Time, Microsoft said it had restored service to some affected servers, which were currently stable, and had shifted traffic and connections to alternate paths, with limited recovery in some regions. The company also simplified the recovery process and implemented protective traffic management at specific locations to reduce load; if the results are positive, it will expand the measure to more locations. However, there is still no timeline for full recovery. @@@ According to VentureBeat, China-linked hackers implanted backdoors into executives' laptops via USB, exploiting a vulnerability targeted by a fix that companies already had but had not deployed. The report says the fix already existed, but the company involved had not enabled it, allowing attackers to breach executives' devices. @@@ According to cnBeta, Mercedes-Benz is pausing all investment in its German plants and considering closing some factories to cope with high vehicle manufacturing costs. The investment pause follows a failure to reach consensus with the works council and union representatives on cutting labor costs. Production chief Michael Schiebe said management had jointly decided to halt follow-up investment at German sites. In September, he warned that two German plants could close if there was no progress; Thursday's information indicates that management is preparing for possible closures. Management board member Britta Seeger said in an employee memo that she wants to avoid closures at all costs, but the option must be discussed further. Negotiations between the company and the IG Metall union are still underway. The crisis in Germany's auto industry is escalating: Volkswagen Group previously carried out its largest restructuring ever, threatening tens of thousands of jobs; Schaeffler announced last week that, hit by weak demand for combustion-engine vehicles, it plans to cut 3,460 jobs in Europe and close five facilities, four of them in Germany. Analyst Dudenhöffer said this is not surprising; unless the cost structure is thoroughly reformed, the industry's future is at risk. @@@ According to TechWeb, JD.com announced on the evening of November 12 that it had completed its takeover offer for German consumer electronics retail group CECONOMY. Through a wholly owned indirect subsidiary, it acquired 59.8% of the share capital and voting rights for €4.60 per share in cash, with total consideration of €1.335 billion. CECONOMY has become a JD subsidiary, and its results will be consolidated into JD's financial statements from November 12, 2026; together with the 25.3% stake retained by partner Convergenta, the two sides hold a combined 85.2%. Merger control, foreign investment, and EU foreign subsidy approvals have all been completed. CECONOMY operates more than 1,000 stores in 11 European markets under the MediaMarkt, Saturn, and MediaWorld brands, employs more than 50,000 people, and had more than 53 million loyalty members in 2025. The announcement said CECONOMY will maintain independent operations and its local technology architecture, with no plans to adjust staffing, employee agreements, or office locations. JD CEO Xu Ran said the two sides will build a leading next-generation consumer electronics platform in Europe; CECONOMY CEO Remko Rijnders said the partnership is now in place and implementation is moving forward. @@@ According to TechWeb, after news of organizational changes and personnel adjustments emerged from Tencent's internal all-hands meeting marking its 28th anniversary, rumors that "Tencent will lay off 50% of staff" appeared on social platforms. On November 12, Tencent PR Director Zhang Jun denied the rumors on Weibo, calling them "nothing but a misunderstanding and rumor." A Tencent insider said it was a summary of an internal speech, not an internal document, and that management hopes businesses that have not yet been mobilized will get moving. On November 11, Tencent's core executives held their regular online dialogue with all employees, with Ma Huateng and Liu Chiping speaking. The leaked speech shows Ma Huateng used the analogy of "shifting from land warfare to naval warfare," called the current situation "a life-or-death transformation," and urged employees to shed complacency and respond with agility. He said the AI shock has arrived and that the company must prepare new playbooks and get on board. Tencent is taking on heavy-asset businesses for the first time. Around 2021, cost reduction and efficiency improvement had already escalated into a deep organizational restructuring; the goal is not to cut costs but to adapt to a new battlefield, and the hope is to let small teams do big things with existing headcount. He proposed that for traditional core businesses, half the people should defend and half should fight new campaigns; on staffing, the company must decisively choose people who can adapt to the new playbook, and stop coaxing people or negotiating conditions before they get on board. Ma Huateng concluded that this LinkTime was intentionally stricter than in the past. @@@ According to 36Kr, Unitree Robotics' share price, after halving two months ago, rebounded; as of November 11, it had doubled in less than a month, and its market cap returned to RMB 300 billion. JPMorgan upgraded its rating two notches from "underweight" to "overweight." On October 22, Wang Xingxing judged that physical AI had begun to show signs of a breakthrough, with a window that could arrive in the next one to two years, or two to three years; JPMorgan subsequently raised its 2028 humanoid robot shipment forecast from about 41,600 units to 111,600 units. On November 4, an Apple research team published HALO results: Unitree G1 was tested 360 times across six unfamiliar environments for six categories of individual skills, achieving an average success rate of 91% on individual tasks. None of these environments were used in training or pre-scanned for mapping; training used data generated from about 1,000 virtual environments. The base version of Unitree G1 now starts at RMB 85,000, down from RMB 99,000; after the 2025 price cut, the overall gross margin remained 63.18%. In 2025, Unitree shipped 5,511 humanoid robots, more than ten times the previous year; in October this year, Wang Xingxing disclosed that cumulative production of the G1 had reached about 10,000 to 20,000 units. In the first three quarters of this year, R&D investment was RMB 263 million, up 191.48% year on year, higher than the 55.71% revenue growth in the same period. @@@ According to Sina Tech, DJI's official website published a notice on November 13 regarding suggested retail price adjustments for some products, saying that due to the sustained sharp rise in costs of multiple key components, including global semiconductors and memory, and after careful evaluation, DJI will adjust the suggested retail prices of some products from midnight on November 18, 2026. Previously, social media rumors said that DJI Pocket 4, Pocket 4P, and the entire Nano series would see across-the-board price increases starting November 18, with increases of up to 13% and up to RMB 500 per unit. @@@ According to Sina Tech, on the evening of November 12, a full-featured integrated innovative terminal jointly developed and promoted by China UnionPay and Ant Group made its national debut at Intime Commercial's Xi'an Bell Tower Kaiyuan Mall. It is authorized for distribution and operation by UnionPay Merchant Services. The terminal can accept major payment methods on one device, including Alipay's "Tap!", domestic and international bank cards, aggregated QR code payments, mobile Pay, and digital RMB. It is currently the most comprehensive financial-grade smart POS terminal on the market in terms of payment method compatibility, meeting requirements such as UnionPay enterprise technical specifications and Alipay IoT device security specifications. The new device also offers value-added digital operation capabilities such as AI business analysis, intelligent customer-flow analysis, precision member operations, and government consumption-promotion programs. Intime Commercial's Xi'an Bell Tower Kaiyuan Mall is the first national chain high-end mall to deploy it, and will be among the first to offer services such as member recognition, payment settlement, and voice-reminder shopping guidance. China UnionPay Chairman Dong Junfeng said this is the industry's first innovative device to achieve integrated payment acceptance; Ant Group CEO Han Xinyi said the new device is fully compatible with various payment methods, and the two sides will continue to deepen cooperation. Since Alipay launched "Tap!" in July 2024, the service has covered 2,200 segmented scenarios such as food ordering and transportation, connected 30 million offline locations, and served 400 million users. @@@ All right, that's all for today's show! Thank you for spending time with us—and remember to take your good mood with you! If you enjoy our program, you're very welcome to subscribe and share it with your friends; it really helps us. See you next time!

12 分钟 55 次播放 5天前

《晚安咖啡》科技夜话:黄仁勋辞职,Outlook故障

《晚安咖啡》科技夜话:黄仁勋辞职,Outlook故障

晚安咖啡GoodNightCoffee

嗨~~大家晚上好呀!今天是2026年11月13日,星期五。我是晚安咖啡的主播安安,超开心又和大家见面啦!今天我会带来一波新鲜热乎的商业科技动态,干货满满、轻松好懂,快跟我一起开启今天的精彩内容吧~ @@@ 据CNBC报道,英伟达正接近2012年以来最长的周线连跌纪录。截至周四收盘,该股本周累跌11%,周四跌4.3%,势将连续第六周走低;本轮市值缩水29%、约1.6万亿美元,并跌破苹果。同日,董事会宣布联合创始人兼CEO黄仁勋辞职,任命首席财务官Colette Kress为临时CEO并寻找继任者。此前,黄仁勋两起婚外情被公开:一起涉及一名被指为中国间谍的人,另一起涉及公司医疗健康业务副总裁;他周末曾致信员工,为“对妻子不忠”道歉。上周黄仁勋离职前,英伟达股价已跌约9%,中期选举后民主党大胜及多项可能放缓AI建设的公投通过,加剧跌势。英伟达计划下周二公布季度业绩,分析师预计营收同比增长91%至1089亿美元,较2022年同期增长18倍。Oppenheimer称黄仁勋离职或带来持续股票过剩,D.A. Davidson将目标价从300美元下调至200美元。 @@@ 据The Verge报道,微软Outlook邮件服务仍未恢复。自周四下午早些时候起,全球多地用户反映无法通过网页或应用访问Outlook邮箱。微软随后在Microsoft 365状态账号承认问题,称部分基础设施在多个地区出现负载升高,可能导致连接错误;Teams也有故障报告,但程度较轻。微软称,受影响范围涉及Microsoft 365商业服务套件,主要是Exchange Online,部分用户可能无法访问邮箱,邮件投递也可能延迟。微软传播总监本·霍普表示,公司已知悉并正在调查,感谢客户耐心,将尽快恢复服务。数小时后,微软多次更新进展,称在重启服务器和监控服务器健康状态方面遇到困难。到美东时间约下午6点,微软表示已将部分受影响服务器恢复服务,目前保持稳定,并把流量和连接转移到替代路径,部分地区出现有限恢复。公司还简化恢复流程,并在特定地点实施保护性流量管理以降低负载,若效果积极将扩大到更多地点。不过,全面恢复仍无时间表。 @@@ 据VentureBeat报道,与中国有关联的黑客通过USB向高管笔记本电脑植入后门,利用的是企业已经拥有但未部署的修复方案所针对的漏洞。报道称,相关修复本已存在,但涉事企业没有启用,攻击者据此侵入高管设备。 @@@ cnBeta消息,梅赛德斯-奔驰正暂停德国工厂所有投资,并考虑关闭部分工厂,以应对高昂的汽车制造成本。暂停投资因未能就削减劳动力成本与劳资委员会及工会代表达成共识。生产负责人迈克尔·希贝表示,管理层已共同决定停止德国各基地的后续投资。9月他曾警告,若无进展,德国两家工厂可能关闭;周四信息表明,管理层正为可能关闭做准备。管理委员会成员布丽塔·泽格在员工备忘录中称,希望不惜代价避免关闭,但必须更深入探讨这一选项。公司与德国金属工业工会(IG Metall)的谈判仍在进行。德国汽车业危机正升级:大众汽车集团此前实施有史以来最大规模重组,威胁数万岗位;舍弗勒上周宣布,受燃油车需求疲软影响,计划在欧洲削减3460个岗位、关闭5处设施,其中4处在德国。分析师杜登霍费尔称这不令人意外,除非彻底改革成本结构,行业未来岌岌可危。 @@@ 据TechWeb报道,京东集团11月12日晚间公告,已完成对德国消费电子零售集团CECONOMY的收购要约。京东通过全资间接子公司,以每股4.60欧元现金、总对价13.35亿欧元收购其59.8%股本及投票权。CECONOMY成为京东子公司,业绩自2026年11月12日起并入京东合并报表;连同合作方Convergenta保留的25.3%股权,双方合计持股85.2%。经营者集中、外商投资及欧盟外国补贴审批均已完成。CECONOMY以MediaMarkt、Saturn及MediaWorld品牌在欧洲11个市场经营逾1000家门店,雇用逾5万员工,2025年拥有逾5300万忠诚会员。公告称,CECONOMY将保持独立运营和当地技术架构,无计划调整人员、员工协议及办公地点。京东CEO许冉称,双方将打造欧洲领先新一代消费电子平台;CECONOMY CEO Remko Rijnders称合作已落地、正推进落实。 @@@ 据TechWeb报道,腾讯28周年内部全员会传出组织变革和人员调整消息后,社交平台出现“腾讯将裁员50%”传闻。11月12日,腾讯公关总监张军微博辟谣,称这是“纯纯的误会和谣言”。腾讯内部人士称,这是内部讲话概述,并非内部文件,管理层希望尚未动员的业务动起来。 11月11日,腾讯核心高管按例与全体员工线上对话,由马化腾和刘炽平讲解。网传讲话显示,马化腾以“陆战转海战”作比,称当前是“生死攸关的变革”,呼吁丢掉安逸、敏捷迎战。他称AI冲击已至,要准备新打法、上船。腾讯首次做重资产业务,2021年前后降本增效已升级为组织架构深度调整,目的不是降本而是适应新战场,希望原有人力下让小团队做大事。他提出传统核心业务一半人守、一半人打新战役;用人上要干脆选择适应新打法者,不再哄着、谈条件才上船。马化腾最后称,这次LinkTime有意比以往严厉。 @@@ 36氪消息,宇树科技股价两个月前腰斩后反弹,截至11月11日不到一个月翻倍,市值重回3000亿元;摩根大通将评级从“低配”连升两级至“增持”。10月22日,王兴兴判断物理AI已出现突破苗头,窗口可能在未来一两年或两三年;随后摩根大通把2028年人形机器人出货量预测从约4.16万台上调至11.16万台。11月4日,苹果研究团队公布HALO成果:宇树G1在6个陌生环境针对6类单项技能测试360次,单项任务平均成功率91%,这些环境均未参与训练或提前扫描建图;训练使用约1000个虚拟环境生成数据。宇树G1基础版起售价已从9.9万元降至8.5万元;2025年降价后,整体毛利率仍为63.18%。2025年宇树人形机器人出货5511台,是上年十倍多;今年10月王兴兴披露,G1累计生产下线约1万至2万台。今年前三季度研发投入2.63亿元,同比增长191.48%,高于同期55.71%的收入增速。 @@@ 据新浪科技报道,大疆官网11月13日发布关于部分产品建议零售价调整的说明,称受全球半导体及存储等多项关键零部件成本持续大幅上涨影响,经审慎评估,大疆将于2026年11月18日零时起调整部分产品的建议零售价。此前,社交媒体传出消息称,大疆Pocket 4、Pocket 4P及Nano全系列将于11月18日起全面上调价格,最高涨幅达13%,单台最高上涨500元。 @@@ 新浪科技消息,11月12日晚,中国银联与蚂蚁集团联合研发推广的全功能融合创新终端在银泰商业西安钟楼开元商城全国首发,由银联商务授权经销和运营。该终端可一机通收支付宝“碰一下”、境内外银行卡、二维码聚合支付、手机Pay和数字人民币等主流支付方式,是当前市场上支付方式兼容维度最全面的金融级智能POS终端,满足银联企业技术规范、支付宝IoT设备安全规范等要求。新设备还具备AI经营分析、智能客流研判、精准会员运营、政府促消费等数字化经营增值能力。银泰商业西安钟楼开元商城是首个落地的全国连锁高端商场,将率先开放会员识别、支付结算、语音提醒导购等服务。中国银联董事长董俊峰称,这是业内第一款实现支付一体化受理的创新设备;蚂蚁集团首席执行官韩歆毅表示,新设备全面兼容各类支付方式,双方将继续深化合作。自2024年7月支付宝推出“碰一下”以来,该服务已覆盖点餐、出行等2200个细分场景,接入3000万个线下点位,服务4亿用户。 @@@ 好啦,今天的节目到这里就结束啦!谢谢大家的陪伴,记得把好心情带走哦!如果您喜欢我们的节目,也非常欢迎订阅和转发给你的小伙伴,这对我们非常有帮助。那我们下期再见吧!

10 分钟 99+ 次播放 5天前

《晚安咖啡》科技夜话:黄仁勋去留,Anthropic IPO,星舰双捕获

《晚安咖啡》科技夜话:黄仁勋去留,Anthropic IPO,星舰双捕获

晚安咖啡GoodNightCoffee

嗨~~大家晚上好呀!今天是2026年11月12日,星期四,欢迎来到晚安咖啡,我是主播安安!今晚咱们一起聊聊新鲜热乎的商业科技动态,干货满满、轻松有趣,快把耳朵借给我,咱们马上出发吧! @@@ 据The Verge援引《华尔街日报》报道,Nvidia董事会正准备解雇CEO黄仁勋,最早12日开会决定其去留。此事源于《纽约时报》报道黄仁勋存在第二段被指婚外情,对象是向他汇报的Nvidia医疗保健副总裁金伯利·鲍威尔。匿名消息人士称,两人均已婚,3月在圣何塞一家酒店被拍到在一起。CNBC称,相关传闻长期在公司内部流传,部分因鲍威尔没有健康或科学背景。几天前,黄仁勋承认与Megaspeed高管、被指为中国间谍的Alice Huang存在婚外情,公司随后表示已聘请前美国联邦检察官Joon H. Kim处理调查;The Verge称,董事会似乎不会等调查结果再行动。Nvidia股价周三盘后下跌超过7%。据Wired报道,黄仁勋去年曾聘请以处理公开性丑闻闻名的危机传播公司Heller。Nvidia及Heller均未立即回应The Verge置评请求。 @@@ 据Readhub消息,SpaceX计划11月14日进行星舰第15次试飞,首次尝试用发射塔“筷子”机械臂同时捕获超重型助推器和星舰飞船。发射拟从得州星际基地进行,窗口为美国得州当地时间11月13日16时45分起的75分钟,目前仍待最终监管许可。若两级同时捕获成功,将成为轨道级运载系统首次实现两级完整回收,推动星舰快速重复使用。此前SpaceX已三次成功捕获超重型助推器,但从未尝试捕获星舰飞船。按计划,助推器将在发射约6分30秒后返场被捕获;星舰将入轨运行超过22小时、绕地球约14圈,部署26颗星链V3卫星后再入大气层返回发射场被捕获,难度更高。任务设置多重安全方案,状态不满足要求时将改为海上溅落;本次还将使用改进后的新一代星舰硬件与发射塔机械臂。 @@@ 据新浪科技报道,知情人士称,人工智能公司Anthropic的IPO已获数倍超额认购,订单达到可供发行股份的数倍,需求主要来自大型基石投资者的意向认购。公司及联合创始人计划通过IPO和260亿美元私募配售合计募资最高904亿美元,有望超过SpaceX的IPO,成为史上最大规模上市融资;IPO拟以每股70至78美元发行8.25亿股,目标完全稀释估值约1.4万亿美元,低于部分投资者此前预估区间。知情人士称,发行条款仍可能变动,Anthropic代表暂未回应置评请求。公司正加大机构推介力度,本周五将举行线上直播,CEO达里奥·阿莫迪将与摩根大通CEO戴蒙同台。部分获配投资者正转售部分认购额度,以赚取费用并降低上市后交易风险;一些上市前股东通过专项载体转让股权。申报文件显示,高盛和摩根士丹利联席主承销,另有20家投行参与,股票将在纳斯达克全球精选市场挂牌,代码ANT。 @@@ 据The Verge报道,欧盟委员会周三公布拟议《数字公平法案》,拟对游戏、社交媒体和电商平台施加新规则。法案针对误导性价格或可用性、游戏开箱等成瘾性设计、未披露的网红营销等利用消费者投入更多时间和金钱的做法。欧盟委员会称,法案将保护各年龄段在线消费者,促进企业公平竞争,预计每年为欧盟消费者节省至多137亿欧元(151亿美元)。提案还可能影响YouTube、TikTok、Meta的“兔子洞”算法,以及AI聊天机器人和DuoLingo等促使用户持续互动的应用。一些游戏组织批评,对虚拟货币、开箱和向未成年人提供微交易的限制“不可行、不合理且不成比例”。该法案仍需欧洲议会和欧盟理事会批准,当前措施可能被修改或删除;若生效,违规企业最高可被罚全球年营业额的6%。欧盟委员麦格拉思称,规则旨在让企业凭产品和服务竞争,而非利用消费者或规避规则。 @@@ 据The Verge报道,谷歌DeepMind研究人员发布新AI模型AlphaGenome 2,用于解析DNA变化如何影响不同细胞。该模型是去年发布的AlphaGenome的后续版本。人体几乎所有细胞携带相同DNA,但不同细胞类型和状态会开启或关闭不同基因,同一遗传变异的影响也会不同。AlphaGenome 2可预测变异对单个细胞类型和状态的影响。谷歌在博客中称,它比AlphaGenome提供更精细视角,可能揭示特定细胞中此前被忽略的疾病相关效应。模型基于覆盖68种组织、逾1亿个单细胞档案训练;公司称,其识别出的可改变细胞基因表达的变异数量接近此前最先进模型的两倍,还能无需重新训练预测新细胞类型和生物学状态,并识别实验数据空白,帮助科学家决定下一步实验室研究。DeepMind研究科学家Adam Gayoso称,重大突破在于“将人类遗传学与细胞生物学联系起来”;基因组学负责人Žiga Avsec表示,药物针对特定蛋白和特定细胞类型,这种颗粒度至关重要。 @@@ 36氪消息,字节跳动联合创始人张一鸣在近期AI团队全员大会上表示,字节有能力与Anthropic和OpenAI竞争,但真正追平前沿最早要到2029年或2030年。他坦言,Seed语言模型预训练数据集比领先对手小几个数量级,并重申不应依赖“蒸馏”更大模型输出训练。字节2025年从Google DeepMind招来吴永辉,领导约1000人的Seed部门;Seed语言主模型最近一次大版本仍是今年6月的2.1。据彭博社援引知情人士消息,字节计划未来几周发布类似Meta Muse的AI助手。字节还讨论2026年在AI基础设施上支出最高700亿美元,并以Seedance视频生成模型填补OpenAI收缩Sora服务后的空白;2026年10月单独设立安全与对齐团队。豆包月活、日活国内领先,但更多依赖运营、推荐和买量,而非Seed模型的代差优势。 @@@ 新浪科技消息,在百度世界2026大会开幕式上,百度集团执行副总裁沈抖发布百度天池1024卡超节点,这套算力集群计划于明年推向市场。该超节点搭载光引擎技术,卡间互联带宽提升3倍;对比上一代产品,训练性能提升2.5倍、推理性能提升3倍。其配套G3.5 KV Cache显存加速柜,强化缓存复用能力,长上下文推理性能提升2倍;在多轮对话缓存命中场景,首token等待时间(TTFT)降低90%,整体吞吐提升7倍。单套天池1024卡超节点可支撑5万亿参数大模型训练,并具备横向扩展能力,可扩展至万卡规模集群。 @@@ 据36氪报道,元宝搜索已经上线,腾讯再次加入默认入口的争夺。 @@@ 新浪科技消息,由优必选、沐曦股份、锋龙股份共同出资成立的具身智能端侧芯片公司曦选创智宣布完成5亿元天使轮融资,投后估值40亿元。本轮由纪源资本、经纬创投联合领投,和利资本、鼎晖龙芯创投、上汽恒旭资本等机构跟投。曦选创智今年5月成立,聚焦自主可控的具身智能端侧芯片与计算平台,首颗自研芯片定位面向机器人的“大小脑一体”SoC,以全面超越英伟达Thor芯片为研发目标,并针对多模态感知、认知决策与运动协同开展芯片定义和系统设计。本轮资金将投入核心芯片与计算平台研发、软件工具链搭建、研发团队扩充,并推进与机器人厂商、行业客户的联合验证及量产交付准备。公司正推进首款机器人主控SoC及配套软硬件系统研发,计划2027年向产业伙伴提供开发验证平台,2028年推出首款SoC及配套具身智能整体解决方案。 @@@ 据虎嗅24小时消息,汽车供应商开始盯上小电驴的智能驾驶,相关报道以“你会买智驾小电驴吗”向消费者发问,关注汽车供应链向两轮电动车智驾领域延伸后,用户是否愿意购买智驾小电驴。 @@@ 好啦,今天的节目到这里就结束啦!谢谢大家的陪伴,希望今天的内容让你轻松又开心~如果您喜欢我们的节目,也非常欢迎订阅和转发给你的小伙伴,这对我们非常有帮助。那我们下期再见吧!

10 分钟 99+ 次播放 6天前

《晚安咖啡》科技夜话(英文版):黄仁勋去留,Anthropic IPO,星舰双捕获

《晚安咖啡》科技夜话(英文版):黄仁勋去留,Anthropic IPO,星舰双捕获

晚安咖啡GoodNightCoffee

Hi~~ Good evening, everyone! Today is Thursday, November 12, 2026. Welcome to Goodnight Coffee; I’m your host, An’an! Tonight, let’s talk about the latest and hottest business and tech developments—full of useful insights, relaxed and fun. Lend me your ears, and let’s get started! @@@ According to The Verge, citing The Wall Street Journal, Nvidia’s board is preparing to dismiss CEO Jensen Huang and could meet as early as the 12th to decide whether he stays or goes. The move comes after The New York Times reported that Huang had a second alleged extramarital affair—this time with Kimberly Powell, Nvidia’s vice president of healthcare, who reports to him. Anonymous sources said both are married and were photographed together at a hotel in San Jose in March. CNBC said rumors had long circulated inside the company, partly because Powell lacks a health or science background. A few days earlier, Huang acknowledged an extramarital affair with Megaspeed executive Alice Huang, who has been accused of being a Chinese spy. The company then said it had hired former U.S. federal prosecutor Joon H. Kim to handle the investigation. The Verge said the board appears unlikely to wait for the investigation’s findings before acting. Nvidia shares fell more than 7% after hours on Wednesday. Wired reported that last year Huang hired Heller, a crisis communications firm known for handling public sex scandals. Nvidia and Heller did not immediately respond to The Verge’s request for comment. @@@ According to Readhub, SpaceX plans to conduct the 15th Starship test flight on November 14, making its first attempt to use the launch tower’s “chopstick” robotic arms to catch both the Super Heavy booster and the Starship spacecraft at the same time. The launch is planned from Starbase, Texas, with a 75-minute window starting at 4:45 p.m. local Texas time on November 13, and it is still awaiting final regulatory approval. If both stages are caught simultaneously, it would mark the first time an orbital-class launch system has achieved full recovery of both stages, advancing rapid Starship reuse. SpaceX has previously caught the Super Heavy booster three times but has never attempted to catch the Starship spacecraft. Under the plan, the booster will return and be caught about 6 minutes 30 seconds after launch; the Starship will orbit for more than 22 hours, circle Earth about 14 times, deploy 26 Starlink V3 satellites, then re-enter the atmosphere and return to the launch site to be caught—a more difficult feat. The mission has multiple safety scenarios; if conditions do not meet requirements, it will switch to an ocean splashdown. This flight will also use improved next-generation Starship hardware and launch tower robotic arms. @@@ According to Sina Tech, people familiar with the matter said the IPO of AI company Anthropic has been several times oversubscribed, with orders reaching several times the shares available for issuance. Demand mainly comes from indications of interest by large cornerstone investors. The company and its co-founders plan to raise up to $90.4 billion in total through the IPO and a $26 billion private placement, potentially surpassing SpaceX’s IPO to become the largest listing financing in history. The IPO plans to offer 825 million shares at $70 to $78 per share, targeting a fully diluted valuation of about $1.4 trillion, below some investors’ previous estimates. The sources said the terms may still change, and Anthropic representatives did not immediately respond to a request for comment. The company is stepping up institutional marketing and will hold an online livestream this Friday, where CEO Dario Amodei will appear alongside JPMorgan CEO Jamie Dimon. Some allocated investors are reselling part of their allocations to earn fees and reduce post-listing trading risk; some pre-IPO shareholders are transferring equity through special-purpose vehicles. The filing shows Goldman Sachs and Morgan Stanley as joint lead underwriters, with 20 other banks participating. The shares will list on the Nasdaq Global Select Market under the ticker ANT. @@@ According to The Verge, the European Commission on Wednesday unveiled the proposed Digital Fairness Act, which would impose new rules on games, social media and e-commerce platforms. The bill targets misleading pricing or availability, addictive designs such as loot boxes in games, undisclosed influencer marketing, and other practices that exploit consumers’ time and money. The European Commission said the bill would protect online consumers of all ages, promote fair competition among businesses, and is expected to save EU consumers up to €13.7 billion ($15.1 billion) a year. The proposal could also affect YouTube, TikTok and Meta’s “rabbit hole” algorithms, as well as AI chatbots and apps like Duolingo that keep users engaged. Some gaming groups criticized restrictions on virtual currencies, loot boxes and offering microtransactions to minors as “unworkable, unreasonable and disproportionate.” The bill still needs approval from the European Parliament and the EU Council, and the current measures may be amended or removed; if it takes effect, violating companies could be fined up to 6% of global annual turnover. EU Commissioner McGrath said the rules are meant to make companies compete on their products and services rather than exploit consumers or circumvent the rules. @@@ According to The Verge, Google DeepMind researchers have released a new AI model, AlphaGenome 2, to parse how DNA changes affect different cells. The model is a follow-up to AlphaGenome, released last year. Nearly all cells in the human body carry the same DNA, but different cell types and states turn different genes on or off, so the same genetic variant can have different effects. AlphaGenome 2 can predict the effects of variants on individual cell types and states. Google said in a blog post that it offers a finer-grained view than AlphaGenome and may reveal disease-related effects in specific cells that were previously overlooked. The model was trained on more than 100 million single-cell profiles covering 68 tissues. The company said it identifies nearly twice as many variants that can alter cellular gene expression as the previous state-of-the-art model, can predict new cell types and biological states without retraining, and can identify gaps in experimental data to help scientists decide on next steps in the lab. DeepMind research scientist Adam Gayoso said the major breakthrough lies in “connecting human genetics with cell biology”; head of genomics Žiga Avsec said drugs target specific proteins and specific cell types, and this granularity is crucial. @@@ According to 36Kr, ByteDance co-founder Zhang Yiming said at a recent AI team all-hands meeting that ByteDance is capable of competing with Anthropic and OpenAI, but truly catching up to the frontier will not happen until 2029 or 2030 at the earliest. He acknowledged that Seed language model pretraining datasets are orders of magnitude smaller than those of leading rivals, and reiterated that the company should not rely on training by “distilling” outputs from larger models. ByteDance hired Wu Yonghui from Google DeepMind in 2025 to lead its approximately 1,000-person Seed department; the latest major version of Seed’s main language model is still 2.1 from June this year. Bloomberg, citing people familiar with the matter, reported that ByteDance plans to release an AI assistant similar to Meta Muse in the coming weeks. ByteDance has also discussed spending up to $70 billion on AI infrastructure in 2026 and using its Seedance video-generation model to fill the gap after OpenAI scaled back its Sora service; it set up a separate safety and alignment team in October 2026. Doubao leads China in monthly and daily active users, but relies more on operations, recommendations and paid user acquisition than on a generational edge from Seed models. @@@ According to Sina Tech, at the opening ceremony of Baidu World 2026, Baidu executive vice president Shen Dou unveiled Baidu’s Tianchi 1024-card Supernode, a computing cluster planned to hit the market next year. The supernode features optical engine technology, with card-to-card interconnect bandwidth increased 3x; compared with the previous generation, training performance is up 2.5x and inference performance up 3x. Its accompanying G3.5 KV Cache memory acceleration cabinet strengthens cache reuse, improving long-context inference performance 2x; in multi-turn dialogue cache-hit scenarios, first-token wait time, or TTFT, is reduced by 90%, while overall throughput increases 7x. A single Tianchi 1024-card Supernode can support training of large models with 5 trillion parameters and offers scale-out capability to a 10,000-card cluster. @@@ According to 36Kr, Yuanbao Search has gone live, with Tencent once again joining the battle for default entry points. @@@ According to Sina Tech, Xixuan Chuangzhi, an embodied AI edge chip company jointly funded by UBTech, MetaX and Fenglong, announced that it has completed a RMB 500 million angel round at a post-money valuation of RMB 4 billion. The round was co-led by GGV Capital and Matrix Partners China, with Heli Capital, CDH Longxin Venture Capital, SAIC Hengxu Capital and other institutions participating. Founded in May this year, Xixuan Chuangzhi focuses on independently controllable embodied AI edge chips and computing platforms. Its first in-house chip is positioned as a robot “brain-and-cerebellum-integrated” SoC, with the R&D goal of comprehensively surpassing Nvidia’s Thor chip, and it is conducting chip definition and system design around multimodal perception, cognitive decision-making and motion coordination. The funds will go toward R&D of core chips and computing platforms, building a software toolchain, expanding the R&D team, and advancing joint validation with robot makers and industry customers as well as preparations for mass production and delivery. The company is advancing R&D of its first robot main control SoC and supporting hardware and software systems, and plans to provide a development and validation platform to industry partners in 2027 and launch its first SoC and an accompanying end-to-end embodied intelligence solution in 2028. @@@ According to Huxiu 24 Hours, auto suppliers have begun to set their sights on smart driving for e-bikes. The report asks consumers, “Would you buy a smart-driving e-bike?” focusing on whether users are willing to buy smart-driving e-bikes as the automotive supply chain extends into smart driving for two-wheeled electric vehicles. @@@ Alright, that’s all for today’s show! Thank you for spending time with us. I hope today’s content made you feel relaxed and happy~ If you like our show, you’re very welcome to subscribe and share it with your friends—it helps us a lot. See you next time!

11 分钟 42 次播放 6天前

播放列表

自动播放下一个

播放列表还是空的

去找些喜欢的单集添加进来吧