Nvidia 是一个看涨逻辑：容器化的 DSX/DGX AI 工厂模块可将建设时间从 6 个月缩短至 14 天，降低客户和贷款方的风险，并消除竞争对手面临的实施障碍；这应有助于销售扩张，说话者预计这些模块将彻底改变市场。
逐段跳播
Nvidia 也在最近的 GTC 柏林大会上宣布了一些进展。我们看到 Nvidia 的股票在市场上上涨了约2.5%。
Nvidia also announced some developments at its recent GTC Berlin conference. we see that Nvidia's stock has risen by approximately 2.5% in the market.
遗憾的是，我们看到盘后出现小幅下跌，我认为这是由于当前一些全球紧张局势和事件，Nvidia 的股票下跌了约1.2%。
Unfortunately, we are seeing a slight decline after official trading hours, which I believe is due to some current global tensions and events , with Nvidia's stock falling by approximately 1.2%.
现在，我想先谈谈 Nvidia。Nvidia 今天举行了年度 GTC 柏林大会。坦白说，展示的很多内容都很熟悉。如果你关注这个频道，你以前就听说过这些，对吧？他们谈到了 AI 智能体的安全性、机器人、Vera 平台提供的不同类型解决方案，以及使用 Vera 或 Blackwell 带来的性能提升。
所以，如果你一直关注这个频道，其中很多内容都相当相似。我知道很多人不关注这个频道，所以这些信息对你来说可能看起来很新。
Now, I would like to start by talking about Nvidia. Nvidia held its annual GTC Berlin conference today. Frankly, much of what was shown was familiar. If you follow this channel, you will have heard of this before, right?
They talked about the safety of AI agents , about robots, about the different types of solutions offered by the Vera platform, and about the performance improvements you get by using Vera or Blackwell.
So, much of it was pretty similar if you were following this channel. I know that many people do not follow this channel, so this information may seem new to you.
但真正让我震惊和意外的是，Nvidia 正在打造一个 DSX 模块。这是一张示意图；我稍后会播放视频。这些是集装箱，这些集装箱里装有图形处理器（GPU）。这也极大地方便了供电和散热。
背后的原因是，他们能够将人工智能工厂的建设时间从6个月缩短到仅14天。我认为这是一项巨大的成就，你不觉得吗？因为数据中心最大的问题之一，就是中心开始产生利润之前所需的等待时间，因为需要搭建系统。
这是 Jensen 试图解决的问题，也就是消除许多额外成本。你甚至不需要担心如何搭建计算单元；我们会把它建好并完整配置好，几乎准备就绪后运给你。这样做的好处是大大降低了财务风险。
在财务方面，当你借钱给别人去建大型数据中心时，等待期可能长达六个月，对吧？我们已经看到了。可能已经过去六个月，数据中心公司已经为算力付了钱，但还没有开始产生收入。现在，情况不再是这样了，对吧？
你不再需要等六个月才开始产生收入，而只需等几周。现在，作为贷款方，你会看到，如果我们使用 Nvidia 产品，风险就降低了。这是关键点；使用 Nvidia 产品是目前唯一可用的选择。
通过使用 Nvidia 产品，风险降低了，产生收入的机会变得容易得多，你收回这笔钱的可能性也提高了。
But what really shocked and surprised me was that Nvidia is creating a DSX module . This is an illustrative image; I will show the video later. These are shipping containers, and these containers contain graphics processing units (GPUs).
This also greatly facilitates power delivery and cooling . The reason behind this is their ability to reduce the production time of an AI factory from 6 months to just 14 days.
I think this is a tremendous achievement, don't you ? Because one of the biggest problems with data centers is the time it takes before a center starts generating profits, due to the need to set up the system.
This is a problem that Jensen is trying to solve, which is to eliminate many of the additional costs. You won't even need to worry about how to build the computing unit; we will build and fully set it up, and ship it to you almost ready.
The great thing about this is that it greatly reduces financial risk. On the financial side, when you lend money to someone to build huge data centers , the waiting period can be up to six months , right?
We have seen that. Six months may have passed in which the data center company has paid for the computing, but has not yet started to generate revenue. Now, that won't be the case anymore , will it?
Instead of waiting six months before you start generating revenue, you will only have to wait a few weeks. Now, as a lender, you see that the risks are reduced if we use Nvidia products.
This is the important point; using Nvidia products is currently the only available option. By using Nvidia products, the risks are reduced, the opportunity to generate revenue becomes much easier, and the likelihood of you paying for this money increases.
好，这里我们会看到一些……（我想图片在最后）。我们开始吧。这是一种集装箱。它从 Vera Rubin 和 VO72 开始。你把适当数量的它们连接在一起，然后放进一个集装箱里，集装箱内包含电力等组件，一些电力基础设施是预先集成的，冷却结构也集成好了，然后你把它运出去。
在我看来，这是我们经常听到的标准概念。当它达到约3万公斤的重量时，令人震惊。你可以加一个干冷器、一个800V 直流电源方案，现在你就有了庞大的系统。你可以一遍又一遍地重复这个过程。
如果你去 Ax，你已经能找到一些使用这个系统的公司。我认为有一家有趣的公司。我现在找不到了，但有一家有趣的公司展示了他们如何做到这一点，以及这对他们来说是一个巨大的胜利。天哪，我真希望我在这里找到了他。
是 Crush Cru-Cru-This。Cruso？Cruso？我甚至不知道怎么发音，但他们有一段大约三分钟的相关片段，这是他们的段落，或者说是他们的版本。我认为这对人工智能数据中心领域的半导体板块来说非常令人兴奋。
再说一次，它消除了很多风险。所以，这对 Nvidia 来说是一个巨大的胜利。在这里，你消除了竞争对手面临的实施障碍，并确保，作为 Nvidia，你所有的客户最终都是赢家。现在你能够扩张并增加销售。
Okay, here we'll see some of... (I think the picture is at the end). Let's begin. This is a type of shipping container. It starts with Vera Rubin and VO72. You connect an appropriate number of them together, then place them inside a shipping container that contains components such as power, some power infrastructure is pre-integrated, and the cooling structure is also integrated , then you ship them.
This is the standard concept, in my opinion, that we hear about a lot. It was astonishing when it reached a weight of approximately 30,000 kilograms. You can add a dry cooler, an 800V DC power solution , and now you have huge systems .
You can repeat this process over and over again . If you go to Ax, you will already find some companies that use this system. I think there is an interesting company. I can't find it now, but there was an interesting company showcasing how they do it and how this is a big win for them.
Oh my God, I wish I had found him here. It was Crush Cru-Cru-This. Cruso? Cruso? I don't even know how to pronounce it, but they have a clip of about three minutes about it, and this is their paragraph, or rather their version of it.
I think this is very exciting for the semiconductor sector in the field of artificial intelligence data centers. And again, it removes a lot of risks. So, it's a huge win for Nvidia.
Here you eliminate the implementation obstacle that competitors face, and ensure, as Nvidia, that all your customers are ultimately winners. Now you are able to expand and increase sales.
这些是我今天想仔细看看的几只股票。我之前对 Nvidia 非常乐观。有一个话题我觉得极其令人兴奋，那就是 Nvidia 设计的 DGX 模块，因为我相信它们将彻底改变市场。
These are some of the stocks I wanted to take a closer look at today. I was very optimistic about Nvidia. There was one topic that I considered extremely exciting , and that was the DGX modules designed by Nvidia, because I believe they will revolutionize the market.
这个频道对 $NVDA 的历次判断
Jose Najarro Stocks 在这只股票上共 20 条判断，这里只列相邻的几条。