阿克雷说,市场普遍忽视 AI 交易以外的高质量公司,这给了他们把新公司纳入组合的机会:三季度新建了 TransDigm 和 Rollins 的小仓位,以后可能加大;如果估值允许,还想买入其他长期欣赏的公司。
We initiated small positions in TransDigm and Rollins that may grow over time, and there are other long-admired businesses we hope to add as well if valuations permit.
达摩达兰分析 AI 资本开支如何流向芯片、网络设备、电力和数据中心公司,英伟达是最大受益者。他说自己大部分组合选择跟随市场共识,持有七巨头中除特斯拉和英伟达以外的五家,其中亚马逊、Alphabet、Meta、微软都是 AI 资本开支大户;但过去一两年新增资金大多放在收益 4% 的短期国债里,组合现金比平时多,他承认少赚了,但睡得更好。
I am personally going with the "market consensus' choice for the bulk of my portfolio, since I do hold five of the Mag Seven (all except Tesla and Nvidia) and four of my holdings in this group (Amazon, Alphabet, Meta and Microsoft) are heavy investors in AI cap ex
面对 AI 估值过高和互联网泡沫的对比,Cramer 仍偏爱有多种方式受益于 AI 的科技巨头,以 Meta、微软为例(他的慈善信托都持有)。Meta 年初表现不佳,但他相信扎克伯格和团队会想出办法让 AI 投入变现,现在靠的是个人智能体 Muse,股价已回到约 738 美元。在利率上升、其他地方难找增长时,AI 相关公司是少数还能保持强劲增长的。
I thought that Mark Zuckerberg and his dream team are brilliant and they'll figure something out.
Friedberg 认为开源和开放权重模型已经无处不在,想集中管控这类软件是徒劳的,风险会随之上升,因此必然带来对防御系统的大量投入。他点名 Palo Alto Networks 这类公司,说本周在华盛顿听到一些财富 50 强 CEO 非常担心网络防御,会把它列为董事会层面的头号优先事项,网络防御预算会“直线飙升”;而 AI 时代的网络防御必须用 AI 来做。
companies like PaloAlto Networks I think the investment in this space um what I heard in DC this week with some Fortune50 CEOs are very worried about cyber defense.
AMD 一周涨近 13%,9 月涨近 30%,一度突破 1 万亿美元市值,部分因为 Meta 推出个人 AI 智能体平台 Muse。Cramer 认为 AI 智能体会让 CPU 需求爆发,而 AMD 仍是 CPU 领导者;GPU 方面它也拿到 Meta、OpenAI 的大单,英伟达产能都满足不了需求,市场足够容纳多个赢家。苏姿丰 2014 年接手时公司只值 25 亿美元,此后股价涨了 18000% 以上。
Lisa Su has orchestrated one of the greatest turnarounds in history at AMD, arguably the greatest of all time, and I see no sign that it's going to stop any time soon,
Jason 说他一直在同时用 Grokbot 和 Meta 的 Muse,这是普通人第一次能从 AI 得到实际价值:用 Muse 可以立刻连接所有服务、真正把事办成,相当于有了免费的执行助理。他提到 Muse 上周五登上 App Store 第一,发布后 Meta 股价涨了 10%,约十天下载 300 万次;他称这个产品非凡,虽不如 Grokbot 放得开,但很会把事情做完。
This product is extraordinary. I've been playing with it, comparing it to Grockbot.
It's really quite excellent, I have to be honest with you, because it simplifies a lot of these more complicated technical capabilities into a very simple, usable interface
Sacks 认为耐克是“go woke, go broke”的又一个例子:品牌本是靠乔丹这类伟大运动员、靠运动表现和胜利建立起来的,后来的营销却推出与运动无关的形象和政治化内容。另一半问题是新 CEO John Donahoe 带来的咨询式“为变而变”:把直销当成新风口,砸掉了零售业务和零售合作关系,原本属于耐克的货架让给了竞争对手,对手因此拿走大量份额。
direct to consumer is the new hotness. So they blow up all their retail operations all their retail relationships.
will go down as one of the most important transactions in AI because you are creating now in the largest competitor and the largest most well capitalized company a bull work against all of this closed source oligopoly insanity
I’ve been going through the lululemon (LULU) 10-Q, conference call transcript and the 8-K press release, and clearly things have changed for the worse. My IV15 fell rather dramatically under new assumptions and probabilities.
We have a lot of visibility from within our companies as to how we're using A.I., what type of benefits it's delivering. So that brought incremental interest.
Friedberg 讲了自己公司的经历:一个周末用 Claude Code 和 Cursor 搭了内部 CRM,后来发现权限、安全、数据存储等功能要一直补,最终换成了 Salesforce,发现好用得多。他的结论是公司的时间和钱应该花在自己独有的流程上(比如他们的植物育种软件),而不是重做 Slack、Gmail、Salesforce 这类别人已经做好的软件。他认为 CRM 这类横向平台跨行业通用,真正会被 AI 冲击的是只服务单一行业的垂直 SaaS。
I think the horizontal platform companies these tools that are horizontal CRM works across many different verticals industry verticals and Gmail works across many industry verticals and Slack and so on.
And so the people that control that today, the large systems of record, what I said before and what I'll double down on now is that those guys hold an incredibly special place in the ecosystem if they do it right.
I do not believe that core systems of record like CRM are going to get ripped and replaced with something vibe coded. I've been saying that for a long time.
Moderna 市值因所谓「癌症疫苗」数据从 200 亿美元涨到 600 亿美元。Friedberg 解释说这是一种个性化新抗原免疫疗法:测序肿瘤 DNA、找出特有蛋白、再让免疫系统攻击它,这套方法从 90 年代末就有,主要靠 NIH 等公共资金研究出来,测序和制备成本都很低,现在已有诊所收 5 万美元做类似治疗。他对 Moderna 打算收 50 万美元感到失望,认为这是借监管流程形成的监管俘获;他承认 Moderna 在自家 mRNA 工具上专利很强,但说这种疗法不一定需要 mRNA,预计未来几年海外会出现便宜得多的同类治疗。
And so what I don't like about this story is that a private pharmaceutical company has seen their market cap go from 20 billion to 60 billion cuz they went through this regulatory process to create what in my mind is a form of regulatory capture
Abercrombie & Fitch might need some time to digest yesterday's gains, but I think [CEO] Fran Horowitz is doing an incredible job here, and if the stock gives you a pullback, may I suggest you do some buying,
If you don't own Dick's, you dodged a bullet today, but based on the last time the stock fell apart, you might want to be a buyer over the next couple of months, because this company has a history of coming back from the dead
Its business should benefit from the shift toward hybrid and battery-electric vehicles, which require significantly more of the company’s products than internal combustion engine (ICE) vehicles.
Let me tell you why I think what Jensen did was so brilliant. It's because the numbers are getting so big that the TAM is getting constrained by the ability to finance this buildout.
What I love about it is that he's creating these huge compute clusters, right? And that gives him the ability to train a big model, a big potentially frontier model.
Under new management, the company refreshed the brand, returned the company to “sexy” (in its marketing and in its new ticker), generated renewed customer interest and improved financial performance.
阿克曼认为 AI 正在提升 Meta 的内容推荐、广告定向和产品创新,核心广告业务增速达到 2021 年以来最快;即便算力自用不完,也能在紧缺市场中出租获利。他预计 Meta 未来几年盈利增速 20% 以上,18 倍远期市盈率是很大的折价。
We view Meta as one of the clearest beneficiaries of AI, which is driving higher engagement through better content recommendation, greater ad monetization through improved targeting, and meaningful product innovation in consumer and business agents.
微软是阿克曼 2026 年新建的仓位。二季度信中他说微软估值处于十年来最低区间之一,反映了市场对 AI 颠覆 M365、Azure 增长和巨额算力投入的担忧。他认为 M365 深度嵌入日常工作流、性价比高,加上自有的 Copilot,比单点软件更能抵御 AI 冲击;Azure 在算力紧缺的市场中位置有利,资本开支是加速增长的审慎投资。
We believe M365's deep embeddedness in daily workflows, superior price-to-value proposition and proprietary Copilot AI agent make it far more resilient to AI disruption than point solution software.
In our view, time reallocated toward short-form video is far more likely to come from share donors like linear TV or lower-quality streaming services than from a utility-like service such as Netflix.
2026 年 2 月 Anthropic 发布 Claude Cowork 后,标普全球股价从高点下跌超过 25%,市盈率降到 19 倍。阿克曼认为评级、指数和普氏三大基准业务贡献八成以上利润,都是寡头格局下的高利润 IP 授权生意,几乎不受 AI 冲击;市场担心的 Capital IQ 占收入不到 7%。
Collectively, S&P’s benchmark businesses have formidable competitive positions and attractive growth prospects, with next to no AI disruption risk.
Given the uncertainty that is associated with my estimate, and you can see it in the simulation that I reported in my post, it is unlikely that I would ever have sufficient conviction to make SpaceX the biggest or only investment in my portfolio, but I stand ready to buy the stock as part of a portfolio, where it is one of many bets that I make on markets.
Jason 指出 Google Cloud 收入同比增长 82%,是云厂商历史上从未见过的增速。他认为谷歌仍会是第一的 AI 公司,因为它已有大量用户在产品里使用 AI:Android、搜索、Gmail、Chrome、YouTube 五款产品月活都超过 30 亿,且正在变成 AI 优先的产品;Gemini 二季度月活超过 9.5 亿、同比翻三倍。他还说自己 95% 的工作用非前沿模型就够了,看好开源和 Gemini 领先。
And if you look at Google, I still think Google will be the number one uh AI company because they have so many people using AI inside of their products already.
Yeah, look, I thought it was a very bullish earnings call. I was a little bit surprised that the stock went down after the earnings call because not only was it a beat and raise, but also I think Elon spoke to a lot of their plans.
Demis Hassabis 转任董事长、Jeff Dean 离职创业后,Friedberg 认为这是谷歌董事会和管理层在重新分配资本:投资 AI 算力基础设施有税收优惠、需求旺盛,回报率高且确定;而在开源和开放权重模型快速追赶的情况下,花数百亿美元做前沿模型很难赚回同样的回报。他说谷歌拥有全球最大的企业和消费者用户基础,不需要拥有最好的模型也能做成很好的生意,可以保持模型中立,同时托管 Anthropic、OpenAI 和各种开放权重模型。他的概括是:数据中心资本开支「高 alpha、低 beta」,模型研发风险高得多。
the way I would frame it is CapEx is high alpha, low beta in data center infrastructure, that capital. And model development theoretically could be high alpha, but it's very high beta.
Once viewed as a cyclical European cement producer, CRH has transformed itself into what we consider one of North America’s leading providers of essential construction materials and infrastructure solutions.
Jason 说苹果过去十年把约 9000 亿美元用于回购和分红,如果拿出一半投给 Anthropic、SpaceX 之类的公司,或用来做新产品、发布汽车、收购有意思的公司,会更有野心。他认为 Tim Cook 不是做这件事的人,但工程师出身的新 CEO 可能会改变这种做法。
I just think Apple could have been more ambitious if they just spent half that money instead of on buybacks and dividends on creating new products and actually releasing their car, maybe buying some interesting companies.
谷歌财报出现上市以来首次负自由现金流、股价下跌,Chamath 反而更看多。他说谷歌上市以来 20 多年平均投入资本回报率约 32%,这样的团队应当被给予信任,他们是在有条不紊地投资自己的优势。他认为谷歌搜索的 AI 转型做得很好,云和芯片业务也很强,模型越分散对谷歌越有利,因为它能在芯片层和云层赚钱。
This is when you are a machine and a group of people and a business model that compounds money at 32% over 20 year average. You give these guys the benefit of the doubt.