在AWS快速增长（4960亿美元积压订单）和广告业务的推动下，AMZN具备强劲的长期潜力，但高额资本开支对自由现金流回报构成风险。
逐段跳播
今天，我将谈谈AMD、Broadcom、Amazon、Meta和Nvidia。Amazon、Microsoft、Google和Meta正在花费数十亿美元建设数据中心、购买芯片并确保电力供应。
Today , I will talk about AMD, Broadcom, Amazon, Meta, and Nvidia. Amazon, Microsoft, Google and Meta are spending billions to build data centers, buy chips and secure electricity.
接下来看下一个箭头，Amazon。Amazon很有意思，因为人们往往把它看作一家在线购物公司，但我认为你应该把Amazon看作实际上是几家不同的公司同处一个屋檐下运营。
Let's move on to the next arrow, Amazon. Amazon is interesting because people tend to think of it as an online shopping company , but I think you should look at Amazon as several different companies actually operating under one roof.
它有电子商务、第三方卖家服务、订阅和广告。我非常喜欢广告，但我发现这很难。但广告简直就是Amazon的秘密武器。Amazon Web Services也是如此。
You have e-commerce, third-party vendor services, subscriptions, and advertising. I really like advertising , but I found it difficult. But advertising is literally Amazon's secret weapon. As well as Amazon Web Services.
Amazon Web Services对于投资人工智能这一逻辑尤为关键。在2026年第二季度，Amazon Web Services实现营收约422亿美元，同比增长约37%。
Amazon Web Services is particularly important to the thesis of investing in artificial intelligence. In the second quarter of 2026, Amazon Web Services generated revenue of approximately $42.2 billion, representing a year-over-year growth of approximately 37% .
这相当于约1690亿美元的年化营收规模。这就是我们所说的云业务，其增长速度是许多规模较小的公司真正难以企及的，而Amazon却是一家巨型公司，不是吗？七大巨头之一，一家庞大的企业，却仍在非常快速地增长。
This is an annual revenue rate of approximately $169 billion. This is what we're talking about when it comes to a cloud business that's growing at a rate that many smaller companies would find truly difficult to achieve, yet Amazon is a huge company, isn't it?
One of the seven giants, a huge business, yet it is growing very rapidly .
为什么这只股票不涨？我在这里看到了一个彻底的矛盾，而作为投资者，我很喜欢Amazon。公司需要算力来训练AI模型、运行AI智能体、存储数据并部署新应用。大多数公司并不愿意花费数十亿美元自建数据中心。
因此，它们可以转而从Amazon Web Services（AWS）租用这些基础设施。它部署的AI应用越多，所需的计算资源就越多。
Why isn't the stock growing? I see a complete contradiction here, and I love Amazon as an investor. Companies need computing power to train AI models , run AI agents , store data, and deploy new applications.
Most companies do not want to spend billions to build their own data centers. Therefore, it could instead rent that infrastructure from Amazon Web Services ( AWS). The more AI applications it deploys, the more computing resources it will need.
但还有另一件极其重要的事。Amazon不仅从其他公司购买芯片，还在开发自己的人工智能芯片，包括“Trainium”，以提升性能并降低提供AI算力的成本。
But here's something else that's extremely important. Amazon is not only buying chips from other companies , but is also developing its own artificial intelligence chips, including "Trainium," to improve performance and reduce the cost of providing AI computing.
AWS可以同时受益于对计算基础设施的需求以及自身芯片设计带来的经济效益。这一经济层面非常重要，因为成本节约意味着更多的自由现金流。
AWS can benefit from the demand for computing infrastructure and the economics of its own chip designs simultaneously. This economic aspect is really important, because cost savings mean greater free cash flow.
Amazon最近宣布，其在AWS的订单积压约为4960亿美元。这是一笔可观的未来合同工作量，尽管其转化为营收的时点取决于客户的部署情况。
Amazon recently announced that its backlog of orders with AWS is approximately $496 billion. This is a substantial amount of future contractual work , although the timing of its conversion into revenue depends on customer deployments.
然后是广告，这是我最喜欢这块业务的部分。Amazon拥有巨大的优势，因为访问其网站的人往往已有购买意图。基本上，他们就是想买东西。这对广告主来说是极其有价值的信息。
Then there are the advertisements, which are my favorite part of this business. Amazon has a huge advantage because people who visit its website often have a pre-existing intention to buy.
Basically, they want to buy something. This is extremely valuable information for advertisers.
一家卖椅子的公司未必想向只是在浏览社交媒体的人投放广告。相反，你想触达的是正在积极寻找购买那把椅子、并愿意为之付钱的人。因此，Amazon可以通过推广商品列表和其他广告产品从这种购物意图中获利。
A company that sells a chair doesn't necessarily want to advertise to someone who is simply browsing social media. Rather, you want to reach someone who is actively looking to buy the chair and is willing to pay money for it.
Therefore, Amazon can profit from that shopping intent through promoted listings and other advertising products.
然而，由AI驱动的购物智能体可能会扰乱这一活动。如果顾客开始更多地依赖AI助手来寻找商品并完成购买，Amazon将需要调整其创造广告收入的方式。
However , AI-powered shopping agents may disrupt this activity. If customers begin to rely more on AI assistants to find products and complete purchases, Amazon will need to adapt its method of generating advertising revenue.
目前，我对Amazon最大的担忧是它的资本支出。建设AI基础设施需要巨额资金，我指的是那种可以用现金装满手推车和整栋楼的巨额资金。这会给，正如你猜到的，自由现金流带来很大压力。
Right now, my biggest concern about Amazon is its capital expenditures. Building an AI infrastructure requires huge sums of money, and by that I mean enormous sums that could fill wheelbarrows and entire buildings with cash.
This can put a lot of pressure on, as you guessed, free cash flow .
所以，关键问题是AWS的营收和营业利润最终是否会以快于建设该基础设施所需支出的速度增长。这正是我在密切关注的。
So, the crucial question is whether AWS's revenue and operating profits will ultimately grow at a faster pace than the spending required to build that infrastructure. This is what I am monitoring closely.
Amazon有多个潜在的增长引擎，但投资者仍需看到有吸引力的投入资本回报。
Amazon has multiple potential growth engines, but investors still need to see attractive returns on invested capital.
我持有的五只股票：AMD、Broadcom、Amazon、Meta 和 Nvidia。这五家公司都能以不同方式利用人工智能，我相信它们具有巨大的长期潜力。当出现有吸引力的机会时，我会继续买入。
My five stocks: AMD, Broadcom, Amazon, Meta, and Nvidia. All five companies are in a position to leverage artificial intelligence in different ways, and I believe they have great long-term potential.
I will continue to buy when attractive opportunities arise .
观察点
这个频道对 $AMZN 的历次判断
Invest with Henry 在这只股票上共 4 条判断，这里只列相邻的几条。