NVIDIA Corporation (NASDAQ:NVDA) GTC Financial Analyst Q&A Call March 19, 2024 11:30 AM ET

英伟达公司（纳斯达克：NVDA）GTC 财务分析师问答电话会议 2024 年 3 月 19 日上午 11:30

Company Participants 公司参与者

Jensen Huang - Founder and Chief Executive Officer

黄仁勋 - 创始人兼首席执行官

Colette Kress - Executive Vice President and Chief Financial Officer

柯蕾特·克雷斯 - 执行副总裁兼首席财务官

Conference Call Participants

电话会议参与者

Ben Reitzes - Melius Research

本·赖茨斯 - 梅利乌斯研究

Vivek Arya - Bank of America Merrill Lynch

Vivek Arya - 美国银行美林证券

Stacy Rasgon - Bernstein Research

斯泰西·拉斯贡 - 伯恩斯坦研究

Matt Ramsay - TD Cowen

马特·拉姆齐 - TD Cowen

Tim Arcuri - UBS

蒂姆·阿库里 - 瑞银

Brett Simpson - Arete Research

布雷特·辛普森 - 阿雷特研究

C.J. Muse - Cantor Fitzgerald

Joseph Moore - Morgan Stanley

约瑟夫·摩尔 - 摩根士丹利

Atif Malik - Citi

Atif Malik - 花旗

Pierre Ferragu - New Street Research

皮埃尔·费拉古 - 新街研究

Aaron Rakers - Wells Fargo

亚伦·雷克斯 - 富国银行

Will Stein - Truist Securities

威尔·斯坦 - Truist Securities

Jensen Huang 黄仁勋

Good morning. Nice to see all of you. All right. What's the game plan?

早上好。很高兴见到大家。好的。有什么计划？

Colette Kress 柯莱特·克雷斯

Okay. Well, we've got a full house and we're thanking you all for coming out for our first in-person in such a long time. Jensen and I are here to kind of really go through any questions that you have, questions from yesterday.

好的。嗯，我们座无虚席，感谢大家出席我们这么长时间以来的首次面对面活动。詹森和我在这里，真的很乐意回答你们的任何问题，包括昨天提出的问题。

And we're going to go through a series of folks that are going to be in the aisles that you can just reach-out to us, raise your hand, we'll get to you with a mic and Jensen are here to answer any questions from yesterday.

我们将逐个介绍一些人，他们将在过道中，您可以直接联系我们，举手示意，我们会给您一个麦克风，詹森在这里回答昨天的任何问题。

We thought that would be a better plan for you. I know you have already asked quite a few questions, both last night and this morning, but rather than giving you a formal presentation, we're just going to go through of good Q&A today. Sound like a good plan.

我们认为这对你来说会是一个更好的计划。我知道你昨晚和今天早上已经问了很多问题，但我们不打算给你一个正式的演示，而是今天只是进行一些好的问答。听起来是一个好计划。

I'm going to turn it to Jensen to see if he wants to add some opening remarks because we have just a quick introduction. We'll do it that way. Okay.

我将把话题转给詹森，看看他是否想要添加一些开场白，因为我们只有一个简短的介绍。我们会这样做。好的。

Jensen Huang 黄仁勋

Yeah. Thank you. First, great to see all of you. There were so many things I wanted to say yesterday and probably have said -- and wanted to say better, but I got to tell you, I've never presented at a rock concert before. I don't know about you guys, but I've never presented in a rock concert before. The -- I had simulated what it was going to be like, but when I walked on stage, it still took my breath away. And so anyways, I did the best I could.

是的。谢谢。首先，很高兴见到大家。昨天有很多事情我想说，可能已经说过了，也想说得更好，但我得告诉你们，我以前从未在摇滚音乐会上演讲过。我不知道你们怎么想，但我以前从未在摇滚音乐会上演讲过。我曾经模拟过那会是什么样子，但当我走上舞台时，还是让我屏住了呼吸。总之，我尽力而为。

Next, after the tour, I'm going to do a better job, I'm sure. I just need a lot more practice. But there were a few things I wanted to tell you. Is there a clicker -- oh, look at that. See, this is like spatial computing. It's -- by the way, if you get -- I don't know you'll get a chance, because it takes a little step up, but if you get a chance to see Omniverse in Vision Pro, it is insane. Completely incomprehensible how realistic it is.

接下来，在参观之后，我会做得更好，我相信。我只需要更多的练习。但有几件事我想告诉你。有一个点击器 - 哦，看那个。看，这就像空间计算。顺便说一句，如果你有机会看到 Vision Pro 中的 Omniverse，那简直是疯狂的。它的真实性是完全无法理解的。

All right. So we spoke about five things yesterday and I think the first one really deserves some explanation. I think the first one is, of course, this new industrial revolution. There were two -- there are two things that are happening, two transitions that are happening. The first is moving from general purpose computing to accelerated computing. If you just looked at the extraordinary trend of general-purpose computing, it has slowed down tremendously over the years.

好的。所以昨天我们谈到了五件事情，我认为第一件事情确实值得一些解释。我认为第一件事情当然是这场新的工业革命。有两个——正在发生的两个转变。第一个是从通用计算转向加速计算。如果你只看一下通用计算的非凡趋势，你会发现它在多年来已经大大减缓。

And in fact, we've known that it's been slowing down for about a decade and people just didn't want to deal with it for a decade, but you really have to deal with it now. And you can see that people are extending the depreciation cycle of their data centers as a result. You could buy a whole new set of general purpose servers and it's not going to improve your throughput of your overall data center dramatically.

事实上，我们已经知道它已经减速了大约十年，人们只是不想处理它十年，但现在你真的必须处理它。你可以看到，由此导致人们延长了数据中心的折旧周期。你可以购买一整套全新的通用服务器，但这并不会显著提高整个数据中心的吞吐量。

And so you might as well just continue to use what you have for a little longer. That trend is never going to reverse. General purpose computing has reached this end. We're going to continue to need it and there's a whole lot of software that runs on it, but it is very clear we should accelerate everything we can.

所以你最好继续使用你现有的东西更长一点。这种趋势永远不会逆转。通用计算已经达到了尽头。我们将继续需要它，有很多软件在其上运行，但很明显我们应该加快一切可能的事情。

There are many different industries that have already been accelerated, some that are very large workloads that we really would like to accelerate more. But the benefits of accelerated computing is very, very clear.

有许多不同的行业已经加速发展，有些是非常庞大的工作量，我们真的希望能够加速更多。但加速计算的好处非常非常明显。

One of the areas that I didn't spend time on yesterday that I really wanted to was data processing. NVIDIA has a suite of libraries that before you could do almost anything in a company, you have to process the data. You have to, of course, ingest the data, and the amount of data is extraordinary. Zettabytes of data being created around the world, just doubling every couple of years, even though computing is not doubling every couple of years.

昨天我没有花时间的一个领域是数据处理，这是我真的很想做的。NVIDIA 有一套库，以前在公司里几乎什么都做不了，你必须处理数据。当然，你必须摄取数据，而数据量是非常庞大的。全球每年创造的数据量达到赫兹，每隔几年就会翻一番，尽管计算能力并非每隔几年就会翻一番。

So you know that data processing, you're on the wrong side of that curve already on data processing. If you don't move to accelerated computing, your data processing bills just keep on going up and up and up and up. And so for a lot of companies that recognize this, AstraZeneca, Visa, Amex, Mastercard, so many, so many companies that we work with, they've reduced their data processing expense by 95%, basically 20 times reduction.

所以你知道数据处理，你已经在数据处理的错误一侧了。如果你不转向加速计算，你的数据处理账单就会不断上涨。因此，许多公司意识到了这一点，包括阿斯利康、Visa、美国运通、万事达卡等等，我们合作的许多公司，他们将数据处理费用减少了 95%，基本上减少了 20 倍。

To the point the acceleration is so extraordinary now with our suite of libraries called rapids, that the inventor of Spark, who started a great company called Databricks, and they are the cloud large scale data processing company, they announced that they're going to take Databricks their photon engine, which is their crown jewel and they're going to accelerate that with NVIDIA GPUs.

加速度现在非常惊人，我们的一套名为 rapids 的库使得这一点成为可能，Spark 的发明者，创立了一家名为 Databricks 的伟大公司，他们是云大规模数据处理公司，宣布他们将把 Databricks 的 photon 引擎，也就是他们的明珠，与 NVIDIA GPU 一起加速。

Okay. So the benefit of acceleration, of course, pass along savings to your customers, but very importantly, so that you can continue to sustainably compute. Otherwise, you're on the wrong side of that curve. You'll never get on the right side of the curve. You have to accelerate. The question is today or tomorrow? Okay. So accelerated computing. We accelerated algorithms so quickly that the marginal cost of computing has declined so tremendously over the last decade that it enabled this new way of doing software called generative AI.

好的。所以加速的好处，当然，是将节省传递给您的客户，但非常重要的是，这样您可以继续可持续地计算。否则，您就会站在曲线的错误一侧。您永远无法站在曲线的正确一侧。您必须加速。问题是今天还是明天？好的。所以加速计算。我们加速了算法，以至于计算的边际成本在过去十年中急剧下降，从而实现了一种称为生成式人工智能的新软件方式。

Generative AI, as you know, requires a lot of flops, a lot of flops, a lot of computation. It is not a normal amount of computation, an insane amount of computation. And yet it can now be done cost effectively that consumers can use this incredible service called ChatGPT. So, it's something to consider that accelerated computing has dropped, has driven down the marginal cost of computing so far that enabled a new way of doing something else.

生成式人工智能，正如您所知，需要大量的浮点运算，大量的浮点运算，大量的计算。这不是一般数量的计算，而是疯狂的计算量。然而，现在可以以成本效益的方式完成，消费者可以使用这种令人难以置信的服务，称为 ChatGPT。因此，值得考虑的是，加速计算已经降低了计算的边际成本，从而实现了一种新的做事方式。

And this new way is software written by computers with a raw material called data. You apply energy to it. There's an instrument called GPU supercomputers. And what comes out of it are tokens that we enjoy. When you're interacting with ChatGPT, you're getting all -- it's producing tokens.

这种新方式是由计算机编写的软件，其原材料称为数据。您向其施加能量。有一种名为 GPU 超级计算机的仪器。从中产生的是我们喜欢的令牌。当您与 ChatGPT 互动时，您得到的是所有--它正在生成令牌。

Now, that data center is not a normal data center. It's not a data center that you know of in the past. The reason for that is this. It's not shared by a whole lot of people. It's not doing a whole lot of different things. It's running one application 24/7. And its job is not just to save money, its job is to make money. It's a factory.

现在，那个数据中心不是一个普通的数据中心。这不是你过去所知道的数据中心。原因是这样的。它不是被很多人共享的。它不是在做很多不同的事情。它是全天候运行一个应用程序。它的工作不仅仅是为了节省钱，它的工作是为了赚钱。它是一个工厂。

This is no different than an AC generator of the last industrial revolution. And it's no different than the raw material coming in is, of course, water. They applied energy to it and turns into electricity. Now it's data that comes into it. It's refined using data processing, and then, of course, generative AI models.

这与上一次工业革命的交流发电机没有什么不同。当然，进来的原材料也是水。他们对其施加能量，将其转化为电能。现在进来的是数据。通过数据处理进行精炼，然后当然是生成式人工智能模型。

And what comes out of it is valuable tokens. This idea that we would apply this basic method of software, token generation, what some people call inference, but token generation. This method of producing software, producing data, interacting with you, ChatGPT is interacting with you.

从中产生的是有价值的代币。我们会应用这种基本的软件方法，代币生成，有些人称之为推理，但代币生成。这种生成软件、生成数据、与您互动的方法，ChatGPT 正在与您互动。

This method of working with you, collaborating with you, you extend this as far as you like, copilots to artificial intelligence agents, you extend the idea as long as you like, but it's basically the same idea. It's generating software, it's generating tokens and it's coming out of this thing called an AI generator that we call GPU supercomputers. Does that make sense?

这种与您合作的方法，与您合作，您可以将其扩展到您喜欢的程度，共同驾驶员到人工智能代理，您可以将这个想法延伸到您喜欢的程度，但基本上是相同的想法。它生成软件，生成令牌，并且它是从我们称之为 GPU 超级计算机的 AI 生成器中产生的。这有意义吗？

And so the two ideas. One is the traditional data centers that we use today should be accelerated and they are. They're being modernized, lots and lots of it, and more and more industries one after another. And so what is a trillion dollars of data centers in the world will surely all be accelerated someday. The question is, how many years would it take to do? But because of the second dynamic, which is its incredible benefit in artificial intelligence, it's going to further accelerate that trend. Does that make sense?

因此有两个想法。一个是我们今天使用的传统数据中心应该加速，而它们正在加速。它们正在现代化，大量的现代化，越来越多的行业一个接一个。因此，世界上的万亿美元数据中心肯定会有一天加速。问题是，需要多少年才能实现？但由于第二个动态，即人工智能的惊人好处，它将进一步加速这一趋势。这有道理吗？

However, the second data center, the second type of data center called AC generators or excuse me, AI generators or AI factories, as I've described it as, this is a brand new thing. It's a brand new type of software generating a brand new type of valuable resource and it's going to be created by companies, by industries, by countries, so on and so forth, a new industry.

然而，第二个数据中心，第二种称为交流发电机或 AI 发电机或 AI 工厂的数据中心，正如我所描述的那样，这是一种全新的东西。这是一种全新类型的软件生成一种全新类型的有价值资源，将由公司、行业、国家等创造，这是一个新的产业。

I also spoke about our new platform. People are -- there are a lot of speculations about Blackwell. Blackwell is both a chip at the heart of the system, but it's really a platform. It's basically a computer system. What NVIDIA does for a living is not build the chip. We build an entire supercomputer, from the chip to the system to the interconnects, the NVLinks, the networking, but very importantly the software.

我也谈到了我们的新平台。人们--关于Blackwell有很多猜测。Blackwell既是系统核心的芯片，但实际上它是一个平台。它基本上是一个计算机系统。英伟达的业务不是制造芯片。我们从芯片到系统再到互连、NVLinks、网络，构建了整个超级计算机，但软件也非常重要。

Could you imagine the mountain of electronics that are brought into your house, how are you going to program it? Without all of the libraries that were created over the years in order to make it effective, you've got a couple of billion dollars' worth of asset you just brought into your company.

你能想象一下进入你家的电子设备堆，你要如何对其进行编程？没有多年来创建的所有库，以使其有效，你刚刚引入公司的资产价值数十亿美元。

And anytime it's not utilized is costing you money. And the expense is too incredible. And so our ability to help companies not just buy the chips, but to bring up the systems and put it to use and then working with them all the time to make it -- put it to better and better and better use, that is really important.

任何时候如果没有被利用，都会让你损失金钱。而这个开支太惊人了。因此，我们帮助公司的能力不仅仅是购买芯片，而是提升系统并投入使用，然后与他们一起不断努力使其更好地使用，这真的很重要。

Okay. That's what NVIDIA does for a living. The platform we call Blackwell has all of these components associated with it that I showed you at the end of the presentation to give you a sense of the magnitude of what we've built. All of that, we then disassemble. This is the hard -- this is the part that's incredibly hard about what we do.

好的。这就是英伟达的生计。我们称之为Blackwell的平台拥有我在演示结束时向您展示的所有这些组件，让您感受到我们所构建的规模。然后我们将所有这些拆解。这就是我们所做的工作中极其困难的部分。

We build this vertically integrated thing, but we build it in a way that can be disassembled later and for you to buy it in parts, because maybe you want to connect it to x86. Maybe you want to connect it to a PCI-Express fabric. Maybe you want to connect it across a whole bunch of fiber, okay, optics.

我们构建了这个垂直整合的东西，但我们以一种可以后期拆卸并让您分部购买的方式构建它，因为也许您想将其连接到 x86。也许您想将其连接到 PCI-Express 结构。也许您想通过一堆光纤连接它，好的，光学。

Maybe you want to have very large NVLink domains. Maybe you want smaller NVLink domains. Maybe you can use arm, maybe so on and so forth. Does it make sense? Maybe you would like to use Ethernet. Okay, Ethernet is not great for AI. It doesn't matter what anybody says.

也许您想要拥有非常大的 NVLink 域。也许您想要更小的 NVLink 域。也许您可以使用 arm，也许等等。这有意义吗？也许您想要使用以太网。好的，以太网对于人工智能并不是很好。无论别人说什么都无所谓。

You can't change the facts. And there's a reason for that. There's a reason why Ethernet is not great for AI. But you can make Ethernet great for AI. In the case of the ethernet industry, it's called Ultra Ethernet. So in about three or four years, Ultra Ethernet is going to come, it'll be better for AI. But until then, it's not good for AI. It's a good network, but it's not good for AI. And so we've extended Ethernet, we've added something to it. We call it Spectrum-X that basically does adaptive routing. It does congestion control. It does noise isolation.

你无法改变事实。这是有原因的。以太网不适合人工智能也是有原因的。但你可以让以太网适合人工智能。在以太网行业中，这被称为超级以太网。所以大约三到四年后，超级以太网将问世，对人工智能会更好。但在那之前，它对人工智能不好。它是一个好的网络，但不适合人工智能。因此，我们扩展了以太网，我们为其添加了一些东西。我们称之为 Spectrum-X，基本上实现了自适应路由、拥塞控制和噪声隔离。

Remember, when you have chatty neighbors, it takes away from the network traffic. And AI, AI is not about the average throughput. AI is not about the average throughput of the network, which is what Ethernet is designed for, maximum average throughput. AI only cares about when did the last student turn in their partial product? It's the last person. A fundamentally different design point. If you're optimizing for highest average versus the worst student, you will come up with a different architecture. Does it make sense?

记住，当你有话多的邻居时，会减少网络流量。而 AI，AI 不是关于平均吞吐量的。AI 不是关于网络的平均吞吐量，以太网是为此设计的，最大平均吞吐量。AI 只关心最后一个学生何时提交了他们的部分产品？就是最后一个人。这是一个根本不同的设计点。如果你在优化最高平均值与最差学生之间的区别，你会得出不同的架构。这有意义吗？

Okay. And because AI has all reduce all to all, all gather, just look it up in the algorithm, the transformer algorithm, the mixture of experts algorithm, you'll see all of it. All these GPUs all have to communicate with each other and the last GPU to submit the answer holds everybody back. That's how it works. And so that's the reason why the networking is such a large impact.

好的。由于 AI 已经将所有内容减少到所有内容，所有内容聚集，只需在算法中查找，变压器算法，专家混合算法，你会看到所有这些。所有这些 GPU 都必须彼此通信，最后一个提交答案的 GPU 会拖慢所有人。这就是它的工作原理。这就是为什么网络对其影响如此之大的原因。

Can you network everything together? Yes. But will you lose 10%, 20% of utilization? Yes. And what's 10% to 20% utilization if the computer is $10,000? Not much. But what's 10% to 20% utilization if the computer is $2 billion? It paid for the whole network, which is the reason why supercomputers are paid -- are built the way they are. Okay.

你能把所有东西都连接在一起吗？可以。但你会损失 10%，20%的利用率吗？会。如果一台计算机价值$10,000，那么 10%到 20%的利用率算什么？不多。但如果一台计算机价值 20 亿美元，那么 10%到 20%的利用率就足以支付整个网络的费用，这也是超级计算机被建造的原因。好的。

And so anyways, I showed examples of all these different components and our company creates a platform and all the software associated with it, all the necessary electronics, and then we work with companies and customers to integrate that into their data center, because maybe their security is different, maybe their thermal management is different, maybe their management plane is different, maybe they want to use it just for one dedicated AI, maybe they want to rent it out for a lot of people to do different AI with.

所以无论如何，我展示了所有这些不同组件的示例，我们的公司创建了一个平台以及与之相关的所有软件，所有必要的电子设备，然后我们与公司和客户合作，将其集成到他们的数据中心中，因为也许他们的安全性不同，也许他们的热管理不同，也许他们的管理平面不同，也许他们只想用它来进行一个专用的人工智能，也许他们想要出租给很多人用来进行不同的人工智能。

The use cases are so broad. And maybe they want to build an on-prem and they want to run VMware on it. And maybe somebody just wants to run Kubernetes, somebody wants to run Slurm. Well, I could list off all of the different varieties of environments and it is completely mind blowing.

用例如此广泛。也许他们想要构建一个本地环境，并在其上运行 VMware。也许有人只想运行 Kubernetes，有人想运行 Slurm。嗯，我可以列出所有不同类型的环境，这完全让人难以置信。

And we took all of those considerations and over the course of quite a long time, we've now figured out how to serve literally everybody. As a result, we could build supercomputers at scale. But basically what NVIDIA does is build data centers. Okay. We break it up into small parts and we sell it as components. People think as a result, we're a chip company.

我们考虑了所有这些因素，并在相当长的时间内，我们现在已经弄清楚如何为每个人提供服务。因此，我们可以按规模构建超级计算机。但基本上，英伟达所做的是建立数据中心。好的。我们将其分解成小部分，并将其作为组件出售。人们认为，因此，我们是一家芯片公司。

The third thing that we did was we talked about this new type of software called NIMs. These large language models are miracles. ChatGPT is a miracle. It's a miracle not just in what it's able to do, but the team that put it so that you can interact with ChatGPT in very high response rate. That is a world class computer science organization. That is not a normal computer science organization.

我们做的第三件事是谈论这种新型软件，叫做 NIMs。这些大型语言模型是奇迹。ChatGPT 就是一个奇迹。它不仅仅在于它能做什么，还有团队将其打造成可以以非常高的响应速度与 ChatGPT 互动。这是一个世界级的计算机科学组织。这不是一个普通的计算机科学组织。

The OpenAI team that's working on this stuff is world class, is a world class team, some of the best in the world. Well, in order for every company to be able to build their own AI, operate their own AI, deploy their own AI, run it across multiple clouds, somebody is going to have to go do that computer science for them. And so instead of doing this for every single model, for every single company, every single configuration, we decided to create the tools and tooling and the operations and we're going to package up large language models for the very first time.

OpenAI 团队正在致力于这项工作，是世界一流的团队，是世界一流的团队，是世界上最好的团队之一。嗯，为了让每家公司都能够构建自己的 AI、运营自己的 AI、部署自己的 AI、在多个云上运行自己的 AI，必须有人为他们做计算机科学。因此，我们决定创建工具和工具以及操作，我们将首次打包大型语言模型。

And you could buy it. You could just come to our website, download it and you can run it. And the way we charge you is all of those models are free. But when you run it, when you deploy it in an enterprise, the cost of running it is $4,500 per GPU per year. Basically, the operating system of running that language model.

你可以购买它。您只需访问我们的网站，下载它，然后就可以运行它。我们向您收费的方式是所有这些模型都是免费的。但是，当您在企业中运行它时，运行成本是每个 GPU 每年 4500 美元。基本上，运行该语言模型的操作系统。

Okay. And so the per instance, the per-use cost is extremely low. It's very, very affordable. And -- but the benefit is really great. Okay. We call that NIMs, NVIDIA Inference Microservices. You take these NIMs and you're going to have NIMs of all kinds. You're going to have NIMs of computer vision. You're going to have NIMs of speech and speech recognition and text to speech and you're going to have facial animation. You're going to have robotic articulation. You're going to have all kinds of different types of NIMs.

好的。因此，每次使用的成本非常低。非常非常实惠。但好处真的很大。好的。我们称之为 NIMs，即 NVIDIA 推理微服务。您拿这些 NIMs，您将拥有各种类型的 NIMs。您将拥有计算机视觉的 NIMs。您将拥有语音和语音识别的 NIMs，文本转语音，您将拥有面部动画。您将拥有机器人关节。您将拥有各种不同类型的 NIMs。

These NIMs, the way that you would use it is you would download it from our website and you would fine tune it with your examples. You would give it examples. You say the way that you responded to that question isn't exactly right. It might be right in another company, but it's not right in ours. And so I'm going to give you some examples that are exactly the way we would like to have it. You show it your work products. This is the way -- this is what a good answer looks like. This is what right answer looks like, whole bunch of them.

这些 NIMs 的使用方式是您可以从我们的网站上下载它，并使用您的示例进行微调。您会给它示例。您会说您对那个问题的回答方式并不完全正确。在另一家公司可能是正确的，但在我们这里不正确。所以我会给您一些确切符合我们要求的示例。您展示它您的工作产品。这就是正确答案的样子。这就是正确答案的样子，有很多。

And we have a system that helps you curate that process that tokenize that, all of the AI processing that goes along with it, all the data processing that goes along with it, fine tuning that, evaluate that, guardrail that so that your AIs are very effective, number one, also very narrow.

我们有一个系统，可以帮助您策划这个过程，对其进行标记化，所有与之相关的人工智能处理，所有与之相关的数据处理，对其进行微调，评估其，设置防护栏，以便您的人工智能非常有效，第一，也非常狭窄。

And the reason why you want it to be very narrow is because if you're a retail company, you would prefer your AI just didn't pontificate about some random stuff, okay. And so whatever the questions are, it guardrails it back to that lane. And so that guard railing system is another AI. So, we have all these different AIs that help you customize our NIMs and you could create all kinds of different NIMs.

你希望它非常狭窄的原因是，如果你是一家零售公司，你会更喜欢你的人工智能不要胡言乱语，好吗。所以无论问题是什么，它都会将其限制在那个范围内。所以这个限制系统又是另一个人工智能。因此，我们有所有这些不同的人工智能来帮助您定制我们的 NIMs，您可以创建各种不同的 NIMs。

And we gave you some frameworks for many of them. And one of the very important ones is understanding proprietary data, because every company has proprietary data. And so we created a microservice called Retriever. It's state-of-the-art and it helps you take your database, which is structured or unstructured images or graphs or charts or whatever it is and we help you embed them.

我们为其中许多提供了一些框架。其中一个非常重要的框架是理解专有数据，因为每家公司都有专有数据。因此，我们创建了一个名为 Retriever 的微服务。它是最先进的，可以帮助您获取您的数据库，无论是结构化的还是非结构化的图像、图表或其他任何内容，我们都可以帮助您嵌入它们。

We help you extract the meaning out of that data. And then we take the -- it's called semantics and what that semantic is embedded in a vector that vector is now indexed into a new database called vector database, okay. And that vector database, then afterwards you can just talk to it. You say, hey, how many mammals do I have, for example. And it goes in there and says, hey, look at that. You got a cat, you have a dog, you have a giraffe.

我们帮助您从数据中提取含义。然后我们进行语义处理，将语义嵌入到一个向量中，这个向量现在被索引到一个名为向量数据库的新数据库中。然后您可以直接与这个向量数据库交互。比如，您可以说，嘿，我有多少哺乳动物。系统会回答，看看，你有一只猫，一只狗，一只长颈鹿。

This is what you have in inventory, in your warehouse you have, okay, so on and so forth, all right. And so all of that is called NeMo and we have experts to help you. And then we put our -- we put a canonical NVIDIA infrastructure we call DGX Cloud in all of the world's clouds. And so we have DGX Cloud in AWS, we have DGX Cloud in Azure, we have DGX Cloud in GCP and OCI.

这是您库存中拥有的东西，在您的仓库里，您拥有的，好的，等等，好的。 所以所有这些都被称为 NeMo，我们有专家来帮助您。 然后我们将我们的 - 我们在全球所有云中放置了一个规范的 NVIDIA 基础设施，我们称之为 DGX Cloud。 因此，我们在 AWS 中有 DGX Cloud，在 Azure 中有 DGX Cloud，在 GCP 和 OCI 中也有 DGX Cloud。

And so we work with the world's enterprise companies, particularly the enterprise IT companies and we create these great AIs with them, but when they're done, they can run in DGX Cloud, which means we're effectively bringing customers to the world's clouds. A platform like us, a platform company, brings system makers customers and CSPs are system makers.

因此，我们与世界上的企业公司合作，特别是企业 IT 公司，与他们一起创建这些出色的人工智能，但当它们完成时，它们可以在 DGX Cloud 中运行，这意味着我们有效地将客户带到世界的云端。像我们这样的平台公司为系统制造商带来客户，而 CSPs 是系统制造商。

They rent systems instead of sell systems, but they're system makers. And so we bring customers to our CSPs, which is a very sensible thing to do just as we brought customers to HP and Dell and IBM and Lenovo and so on and so forth and Supermicro and CoreWeave, so on and so forth, we bring customers to CSPs because a platform company does that. Does that make sense?

他们租用系统而不是出售系统，但他们是系统制造商。因此，我们将客户引荐给我们的 CSPs，这是一件非常明智的事情，就像我们曾经为惠普、戴尔、IBM、联想、超微和 CoreWeave 等公司带来客户一样，我们将客户引荐给 CSPs，因为一个平台公司就是这样做的。这样做有道理吗？

If you're a platform company, you create opportunities for everybody in your ecosystem. And so the DGX Cloud allows us to land all of these enterprise applications in the world CSPs. And they want to do it on-prem. We have great partnerships with Dell that we announced yesterday, HP and others, that you can land those NIMs in their systems.

如果您是一个平台公司，您为生态系统中的每个人创造机会。因此，DGX Cloud 让我们能够在全球的云服务提供商中部署所有这些企业应用程序。他们希望在本地进行部署。我们与戴尔、惠普等公司有着良好的合作关系，您可以将这些 NIM 部署在他们的系统中。

And then I talked about the next wave of AI, which is really about industrial AI. This -- that the vast majority of the world's industries, the largest in dollars, are heavy industries and heavy industries have never really benefited from IT. They've not benefited from a lot of the design and all of the digital.

然后我谈到了下一波人工智能，这实际上是关于工业人工智能。世界上绝大多数行业，以美元计算最大的行业是重工业，重工业从未真正受益于信息技术。他们没有从许多设计和所有数字化中受益。

It's called not digitization, but digitalization, putting it to use. They've not benefited from digitalization, not like our industry. And because our industry is completely digitalized, our technology advance is insanely great. We don't call it chip discovery. We call it chip design. Why do they call it drug discovery, like, tomorrow could be different than yesterday? Because it is.

它被称为数字化，而不是数字化，将其投入使用。 他们没有从数字化中受益，不像我们的行业。 因为我们的行业完全数字化，我们的技术进步是非常巨大的。 我们不称其为芯片发现。 我们称其为芯片设计。 他们为什么称其为药物发现，就像明天可能会不同于昨天一样？ 因为是这样。

And it's so much -- it's so complicated -- it's so complicated biology, it's so changed -- and the longitudinal impact is so great, because, as you know, life evolves at a different rate than transistors. And so therefore, cause and effect is harder to monitor because it happens over a large scale, large scale of systems and large scale of time. These are very complicated problems. Physics is very similar.

它是如此之多 - 它是如此复杂 - 它是如此复杂的生物学，它是如此改变 - 纵向影响是如此之大，因为，正如你所知，生命的演变速度与晶体管不同。 因此，因果关系更难监测，因为它发生在大规模、大规模系统和大规模时间上。 这些都是非常复杂的问题。 物理学非常相似。

Okay. Industrial physics is very similar. And so we finally have the ability using large language models, the same technologies. If we can tokenize proteins, if we could tokenize -- if we can tokenize words, tokenize speech, tokenize images, we can tokenize articulation. This is no different than speech, right?

好的。工业物理学非常相似。因此，我们最终有能力使用大型语言模型，相同的技术。如果我们可以对蛋白质进行标记，如果我们可以对单词进行标记，对语音进行标记，对图像进行标记，我们可以对表达进行标记。这与语音没有什么不同，对吧？

If we can understand its meaning and we can connect it to other modalities then we can do generative AI. So I just explained very quickly that 12 years ago I saw it, our company saw it with ImageNet. The big breakthrough was literally 12 years ago.

如果我们能理解它的含义并将其与其他模态连接起来，那么我们就可以做生成式人工智能。所以我很快地解释了 12 年前我看到它，我们公司与 ImageNet 一起看到了它。这个重大突破确实是 12 年前。

We said, interesting, but what are we actually looking at? Interesting, but what are you looking at? ChatGPT, I would say, everybody should say interesting, but what are we looking at? What are we looking at? We are looking at a computer software that can emulate you -- emulate us.

我们说，有趣，但我们实际上在看什么？有趣，但你在看什么？ChatGPT，我会说，每个人都应该说有趣，但我们在看什么？我们在看什么？我们正在看一个可以模拟你 - 模拟我们的计算机软件。

By reading our words, it's emulating the production of our words. Why -- if you can tokenize words and if you could tokenize articulation, for example, why can't it imitate us and generalize it in a way that ChatGPT has. So the ChatGPT moment for robotics has got to be around the corner. And so we want to enable people to be able to do that. And so we created this operating system that enables these AIs to be able to practice in a physically based world and we call it Omniverse.

通过阅读我们的文字，它正在模拟我们的文字的产生。为什么——如果你可以对单词进行标记化，如果你可以对表达进行标记化，例如，为什么它不能模仿我们并以 ChatGPT 的方式进行概括。因此，机器人的 ChatGPT 时刻已经临近。因此，我们希望使人们能够做到这一点。因此，我们创建了这个操作系统，使这些人工智能能够在基于物理的世界中进行实践，我们称之为 Omniverse。

Omniverse is not a tool. Omniverse is not even an engine. Omniverse are APIs, technology APIs that supercharge other people's tools. And so I'm super excited about the announcement with Dassault. They're using -- they're connecting to Omniverse API to supercharge 3DEXCITE. Microsoft is connected it to Power BI.

Omniverse 不是一个工具。 Omniverse 甚至不是一个引擎。 Omniverse 是 API，技术 API，可以为其他人的工具提供强大支持。因此，我对与达索的宣布感到非常兴奋。他们正在使用 Omniverse API 来强化 3DEXCITE。微软将其连接到 Power BI。

Rockwell has connected it to their tools for industrial automation. Siemens has connected to their, so it's a bunch of APIs that is physically based and it produces image or articulation and it connects a whole bunch of different environments. And so these APIs are intended to supercharge third party tools. And I'm super delighted to see the adoption across it, particularly in industrial automation. And so those are the five things that we did.

Rockwell 已将其连接到他们的工业自动化工具。西门子已将其连接到他们的工业自动化工具，因此这是一堆基于物理的 API，它产生图像或表达，并连接了许多不同的环境。因此，这些 API 旨在为第三方工具提供强大支持。我非常高兴看到它在各个领域的采用，尤其是在工业自动化领域。这就是我们所做的五件事。

I'll do this next one very quickly. I'm sorry I took longer than I should, but let me do this next one really quickly. Look at that. All right. So this chart, don't over stare at it, but it's basically, it communicates several things. On top are developers. NVIDIA is a market maker, not share taker. The reason for that is everything we do doesn't exist when we started doing it. There is no such -- you just go up and down. In fact, even in originally 3D computer games didn't exist when we started working on it.

我会很快完成下一个。抱歉我花的时间比应该的长，但让我很快完成下一个。看看这个。好的。所以这个图表，不要盯着它看，但基本上，它传达了几个事情。顶部是开发者。NVIDIA 是市场创造者，而不是份额占有者。这是因为我们所做的一切在我们开始做时并不存在。没有这样的——你只是上上下下。事实上，即使最初的 3D 电脑游戏在我们开始研究时也不存在。

And so we had to go create the algorithms necessary. Real time ray tracing did not exist until we created it. And so all of these different capabilities did not exist until we created it. And once we created it, there are no applications for it. So we had to go cultivate and work with developers to integrate this technology we have just created so that applications could be benefited by it.

因此，我们不得不去创建必要的算法。直到我们创造出实时光线追踪技术之前，它并不存在。因此，所有这些不同的功能直到我们创造出来之前也都不存在。一旦我们创造出来，就没有应用程序可以使用它。因此，我们不得不去培育并与开发人员合作，将我们刚刚创造的技术整合进去，以便应用程序可以从中受益。

I just explained that for Omniverse. We invented Omniverse. We didn't take anything from anybody, didn't exist. And in order for it to be useful, we now have to have developers, Dassault, Ansys, Cadence, so on and so forth. Does that make sense? Rockwell, Siemens.

我刚刚解释了 Omniverse。我们发明了 Omniverse。我们没有从任何人那里拿走任何东西，也不存在。为了让它有用，我们现在必须有开发人员，达索系统，安捷伦，凯登斯等等。这有道理吗？洛克韦尔，西门子。

We need the developers to take advantage of our APIs, our technologies. Sometimes they're in the form of an SDK. In the case of Omniverse, I'm super proud that it's in the form of cloud APIs, because now it's so easy to use that you could use it in both ways, but APIs are much, much easier to use, okay. And we host Omniverse in the Azure cloud. And notice whenever we connect it to a customer, we create an opportunity for Azure.

我们需要开发人员利用我们的 API、我们的技术。有时它们以 SDK 的形式存在。就 Omniverse 而言，我非常自豪它以云 API 的形式存在，因为现在使用起来非常容易，你可以两种方式使用，但 API 更容易使用，好吧。我们将 Omniverse 托管在 Azure 云中。请注意，每当我们将其连接到客户时，我们为 Azure 创造了机会。

So Azure is on the foundation, their system provider. Back in the old days, system providers used to be OEMs and they continue to be, but system providers on the bottom, developers on top. We invent technology in the middle. The technology that we invent happens to be chip last.

所以 Azure 是基础，他们的系统提供商。在过去，系统提供商过去是 OEM，他们仍然是，但系统提供商在底部，开发人员在顶部。我们在中间发明技术。我们发明的技术恰好是芯片最后。

It's software first. And the reason for that is without a developer, there will be no demand for chips. And so NVIDIA is an algorithm company first and we create these SDKs. They call them DSLs, domain specific libraries. SQL is a domain specific library. You might have heard of Hadoop is a domain specific library in storage computing.

这是软件优先。原因是没有开发人员，就不会有对芯片的需求。因此，英伟达首先是一家算法公司，我们创建这些 SDK。他们称之为 DSL，领域特定库。SQL 是一个领域特定库。你可能听说过 Hadoop 是一个领域特定库在存储计算中。

NVIDIA's cuDNN is potentially the most successful domain specific library short of SQL the world has ever seen. cuDNN is the domain specific library. It's computation engine library for deep neural networks. Without DNN, none of them would have been able to use CUDA. So DNN was invented.

NVIDIA 的 cuDNN 可能是世界上除了 SQL 之外最成功的领域特定库。cuDNN 是领域特定库。它是用于深度神经网络的计算引擎库。没有 DNN，它们中的任何一个都无法使用 CUDA。因此 DNN 被发明了。

Real time ray tracing optics, which led to RTX, makes sense. And we have hundreds of domain specific libraries. Omniverse is a domain specific library. And these domain specific libraries are integrated with developers on the software side, which then when the applications are created and there's demand for that application, creates opportunities for the foundation below. We are market makers, not share takers. Does that make sense?

实时光线追踪光学技术，导致了 RTX，这是有道理的。我们拥有数百个领域特定的库。Omniverse 是一个领域特定的库。这些领域特定的库与开发人员在软件方面集成在一起，当应用程序被创建并且对该应用程序有需求时，为底层基础创造机会。我们是市场创造者，而不是份额占有者。这有道理吗？

And so what's the takeaway? The takeaway is you can't create markets without software. It has always been the case. That has never changed. You could build chips to make software run better, but you can't create a new market without software. What makes NVIDIA unique is that we're the only chip company I believe that can go create its own market and notice all the markets we're creating.

那么结论是什么？结论是没有软件就无法创造市场。这一直是事实。这一点从未改变。你可以制造芯片来让软件运行更好，但没有软件就无法创造新市场。英伟达独特之处在于，我相信我们是唯一能够创造自己市场并注意到我们正在创造的所有市场的芯片公司。

That's why we're always talking about the future. These are things that we're working on. We really -- nothing would give me more joy to work with the entire industry to create the computer aided drug design industry, not drug discovery industry, drug design industry.

这就是为什么我们总是在谈论未来。这些是我们正在努力的事情。我们真的——没有什么能比与整个行业合作，共同打造计算机辅助药物设计行业更让我感到快乐，而不是药物发现行业，而是药物设计行业。

We had to do drug designed the way we do drug chip design not chip discovery. And so I expect every single chip next year to be better than the one before, not as if I'm looking for truffles, which is discovery. Some days are good, some days are less good.

Okay, all right. So we have developers on top. We have our foundation on the bottom. The developers want something very, very simple. They want to make sure that your technology is performing, but they have to solve the problem, that they couldn't solve any other way. But the most important thing for a developer is installed base. And the reason for that is they don't sell hardware, their software doesn't get used if nobody has the hardware to run it.

好的，好的。所以我们有开发人员在顶部。我们在底部有我们的基础。开发人员想要的是非常非常简单的东西。他们想要确保您的技术运行正常，但他们必须解决这个问题，他们无法用其他方式解决。但对于开发人员来说，最重要的是已安装的基础。原因是他们不销售硬件，如果没有人有硬件来运行它，他们的软件就无法使用。

Okay. So what developers want is installed base that has not changed since the beginning of time, is has not changed now. Artificial intelligence, if you develop artificial intelligence software and you want to deploy so that people could use it, you need installed base.

好的。开发人员想要的是自始至终没有改变的安装基础，现在也没有改变。如果您开发人工智能软件并希望部署以便人们使用，您需要安装基础。

Second, the systems companies, the foundation companies they want killer apps. That's the way -- that's the reason why killer app word existed because where there is a killer app, there is customer demand, where there is customer demand, you can sell hardware.

其次，系统公司、基础公司都希望有杀手级应用。这就是为什么有“杀手级应用”这个词存在的原因，因为有了杀手级应用，就会有客户需求，有了客户需求，就可以销售硬件。

And so, it turns out this loop is insanely hard to kick-start. And how many accelerated computing platforms can you really, really build? Can you have an accelerated computing platform for generative AI as well as industrial robotics, as well as quantum as well as 6G as well as weather prediction as well.

因此，结果表明这个循环非常难以启动。您真的可以构建多少加速计算平台？您是否可以为生成式人工智能、工业机器人技术、量子技术、6G 技术以及天气预测等领域构建加速计算平台。

And you can have all these different versions because some of it is good at fluids. Some of it's good at particle. Some of it is good at biology. Some of it is good at robotics. Some of it is good at AI. Some of it is good at SQL. The answer is no. You need a general -- sufficiently general purpose accelerated computing platform. Just as the last computing platform was insanely successful because they ran everything.

你可以拥有所有这些不同版本，因为其中一些擅长流体。其中一些擅长粒子。其中一些擅长生物学。其中一些擅长机器人技术。其中一些擅长人工智能。其中一些擅长 SQL。答案是否定的。你需要一个通用的——足够通用的加速计算平台。就像最后一个计算平台之所以如此成功，是因为它们运行了一切。

Now NVIDIA is taken us a long time, but we basically run everything. If your software is accelerated, I am very certain, it runs on NVIDIA. Does that makes sense? Okay. If you have accelerated software, I am very, very certain it runs on NVIDIA. And the reason for that is because it probably ran on NVIDIA first.

现在 NVIDIA 已经花了我们很长时间，但基本上我们运行一切。如果您的软件加速了，我非常肯定，它是在 NVIDIA 上运行的。这有道理吗？好的。如果您有加速软件，我非常非常肯定它是在 NVIDIA 上运行的。这是因为它很可能首先在 NVIDIA 上运行。

Okay. All right. So this is the NVIDIA architecture. I spoke about whenever I give keynotes, I tend to touch on all of them, different pieces of it, something that -- some new things that we did in the middle, in this case, Blackwell. I spoke about there were so many good stuff and you really have to go to our tox, looks like a 1000 tox. 6G research, how 6G going to happen? Of course, AI. And why do you use the AI for? Robotic MIMO.

好的。好的。所以这是 NVIDIA 架构。每当我发表主题演讲时，我倾向于涉及所有这些不同部分，一些新的东西，这次是在中间做的，Blackwell。我谈到了有很多好东西，你真的必须去我们的 tox，看起来像一个 1000 tox。6G 研究，6G 将如何发生？当然，人工智能。你为什么使用人工智能？机器人 MIMO。

Why is MIMO so pre-installed meaning that, why does the algorithm come before the site. We should have site-specific MIMO just like Robotic MIMO. And so, reinforcement learning and the deals with the environment and so 6G of course is going to be software-defined, of course, it's going to be AI.

为什么 MIMO 如此预先安装意味着，为什么算法在站点之前出现。我们应该像机器人 MIMO 一样具有特定于站点的 MIMO。因此，强化学习和处理环境，当然，6G 将是软件定义的，当然，它将是人工智能。

Quantum Computing, of course, we should be a great partner for the quantum computing industry. How else are you going to drive a Quantum Computer? To have the world's fastest computer sitting next to it.

量子计算，当然，我们应该是量子计算行业的重要合作伙伴。否则，你怎么能驱动量子计算机呢？要让世界上最快的计算机坐在旁边。

And how are you going to stimulate a Quantum Computer, emulate the Quantum Computer? What is the programming model for Quantum Computer? You can't just program a Quantum Computer all by itself. You need to have classical computing sitting next to it. And sort of quantum would be kind of a quantum accelerator.

你打算如何激发量子计算机，模拟量子计算机？量子计算机的编程模型是什么？你不能只是单独为量子计算机编程。你需要旁边有经典计算机。而且量子会成为一种量子加速器。

And so that -- who should go do that, well we've done that and so we work with all the industry on that. So across the board, some really, really great stuff. I wish I could have covered, we could have a whole keynote just on all that stuff. But we cover the whole gamut. Okay. So that was kind of yesterday. Thank you for that.

所以 -- 谁应该去做那个，嗯，我们已经做了，所以我们与所有行业合作。因此，总体而言，一些真的非常棒的东西。我希望我能够涵盖，我们可以就所有这些东西做一个完整的主题演讲。但我们涵盖了整个范围。好的。所以那就是昨天的事情。谢谢你。

Question-and-Answer Session

问答环节

A - Colette Kress

Okay. We have them going around and we'll see if we can grab your questions.

好的。我们让他们四处走动，看看能否回答您的问题。

Jensen Huang 黄仁勋

That was the question that I'm sure, first question goes. If you could have -- done the keynote in 10 minutes, why didn't just do yesterday in 10 minutes? Good question.

那就是我确定的问题，第一个问题。如果你可以在 10 分钟内完成主题演讲，为什么昨天不在 10 分钟内完成呢？好问题。

Ben Reitzes 本·赖茨斯

Yeah. Hi, Jensen. 是的。嗨，詹森。

Jensen Huang 黄仁勋

Hi. 嗨。

Ben Reitzes 本·赖茨斯

Ben Reitzes with Melius Research. Nice to see you. Thanks for being here.

本·赖茨斯与梅利乌斯研究。很高兴见到你。谢谢你在这里。

Jensen Huang 黄仁勋

Thank you, Ben. 谢谢你，本。

Ben Reitzes 本·赖茨斯

It's a big thrill, I think for all of us. So I wanted to ask you a little bit more about your vision with software. You are creating industries. You have a full-stack approach. It's clear, your software makes your chips run better. Do you feel that your software business over the long term could be as big as your chip businesses? How do you look at -- if we look in 10 years are you -- and you're not a chip company, but what do you think, you look like given what you're seeing with the momentum in software and how you're building these industries. It would seem like you're going to be a lot more.

这对我们所有人来说都是一个巨大的兴奋。所以我想再问你一点关于你对软件愿景的看法。你正在创造产业。你采用了全栈方法。很明显，你的软件让你的芯片运行得更好。你觉得你的软件业务从长远来看可能会像你的芯片业务一样大吗？你如何看待——如果我们往后看 10 年，你——你不是一家芯片公司，但你认为，根据你对软件动能和你如何打造这些产业的看法，你会是什么样子。看起来你会更加强大。

Jensen Huang 黄仁勋

Yeah. Thank you, Ben. I appreciate that. First of all, I appreciate all of you coming. This is a very, very different type of event as you know. Most of the talks are software talks, and they're all computer scientists, and they're talking about algorithms. What NVIDIA -- the NVIDIA software stack is about two things. It's either algorithms that help the computer run better, TensorRT-LLM. It's an insanely complicated algorithm, and it explores the computing space in a way that most compilers never have to do. And TensorRT-LLM can't even be built without a supercomputer. And it's very likely that TensorRT in the future, TensorRT-LLM in the future, actually just have to run on a supercomputer all the time and in order to optimize AIs for everybody's computer. And so that optimization problem is very, very complicated. So that would be an example of software that we create, the optimization, the runtime. The second software we create is whenever there's an algorithm where the principled algorithm is well known. For example, Navier-Stokes, however --Schrodinger's equation, however, maybe the expression of it in a supercomputing or accelerated computing or real-time way ray tracing is a great example. Real-time way has never been discovered. Does that make sense? Okay. And so, as you know, Navier-Stokes is insanely complicated algorithm. And to be able to refactor that in a way that can run in real-time is insanely complicated as well and requires a lot of invention and some of the inventions, some of our computer scientists in our company have Oscars. There's award-winning computer scientists because they've solved these problems at such a large scale that you use it for movies. And their inventions are, their algorithms are, their data structures are computer science in itself. Okay. And so we'll dedicate ourselves to these two layers. And then when you package it -- all back in the old days, that's useful for entertainment, media entertainment, science, so on and so forth. But today, because AI has brought this technology so close to application, simulating molecules used to be a thing that you do in universities. Now you can do that at work. So as we now reformulate all of these algorithms for the consumption of enterprise, it becomes enterprise software. Enterprise software like nobody's ever seen before. We call them -- we're going to put them in NIMs, these packages. We'll have hundreds of them, and we'll manufacture these things and support them and maintain them and keep them performant and so on, to support customers with it. And so we'll produce NIMs at a very large scale, is my guess. And this is going to be, we call that underneath the entire bucket of software, we call NVIDIA AI Enterprise. A NIM is basically an AI in a microservice for enterprise. And so my expectation is that this is going to be a very large business, and this is the part of the industrial revolution. If you saw that, there's the IT industry today, SAPs and great companies, ServiceNow's and Adobe's and Autodesk and Canes, that layer, that's today's IT industry. That's not where we're going to play. We're going to play on the layer above. That layer above is a bunch of AIs and these algorithms, really, we are the right company to go build them. And so we'll build some with them, we'll build some ourselves, but we'll package them up and deploy it at enterprise scale. Okay. And so I appreciate you asking the question. And while she's walking there. Go ahead. Yeah.

是的。谢谢你，本。我很感激。首先，我感激你们所有人的到来。正如你们所知，这是一种非常非常不同类型的活动。大多数讨论都是关于软件的，都是计算机科学家，他们在谈论算法。NVIDIA - NVIDIA 软件堆栈涉及两个方面。要么是帮助计算机更好运行的算法，TensorRT-LLM。这是一个极其复杂的算法，它以大多数编译器从未尝试过的方式探索计算空间。而 TensorRT-LLM甚至无法在没有超级计算机的情况下构建。很可能在未来，TensorRT-LLM将不得不始终在超级计算机上运行，以优化每个人的计算机的人工智能。因此，这个优化问题非常非常复杂。这将是我们创建的软件的一个例子，优化，运行时。我们创建的第二个软件是每当有一个已知良好的原则算法的算法时。 例如，Navier-Stokes，然而 - 薛定谔方程，也许在超级计算或加速计算或实时光线追踪的表达方式是一个很好的例子。实时方式从未被发现过。这有意义吗？好的。所以，正如你所知，Navier-Stokes 是一个极其复杂的算法。能够以实时方式重构它同样也是极其复杂的，并且需要大量的发明和一些发明，我们公司的一些计算机科学家获得了奥斯卡奖。有些是获奖的计算机科学家，因为他们解决了这些问题，这些问题在很大程度上被用于电影。他们的发明，他们的算法，他们的数据结构本身就是计算机科学。好的。所以我们将致力于这两个层面。然后当你打包它时 - 在过去，这对娱乐、媒体娱乐、科学等方面很有用。但是今天，因为人工智能将这项技术带得如此接近应用，模拟分子曾经是你在大学里做的事情。现在你可以在工作中做到这一点。 因此，当我们现在重新制定所有这些算法以供企业使用时，它就变成了企业软件。企业软件就像以前从未见过的那样。我们称它们为 - 我们将把它们放在 NIM 中，这些软件包。我们将拥有数百个，我们将制造这些东西并支持它们、维护它们并保持它们的性能等，以支持客户使用。因此，我猜我们将以非常大的规模生产 NIM。这将是我们称之为整个软件桶底下的 NVIDIA AI Enterprise。NIM 基本上是企业中的微服务人工智能。因此，我期望这将是一个非常庞大的业务，并且这是工业革命的一部分。如果你看到了，今天的 IT 行业，SAP、伟大的公司、ServiceNow、Adobe、Autodesk 和 Canes，那一层，那就是今天的 IT 行业。那不是我们要参与的地方。我们要参与的是上面的那一层。那一层是一堆人工智能和这些算法，实际上，我们是去构建它们的正确公司。 所以我们将与他们一起建立一些，我们将自己建立一些，但我们将打包它们并以企业规模部署。好的。所以我感谢你提出这个问题。当她走到那里时。继续。是的。

Vivek Arya

Hi. Vivek Arya from Bank of America Securities. Thank you, Jensen. Thank you Colette for the presentation. So Jensen my question is perhaps a little more near to medium term, which is just the size of the addressable market, because your revenues have gotten big so quickly. And when I look at how much they represent as a percentage of the spending of some of your large customers they are like 30%, 40%, 50%, sometimes more, but when I look at how much money they are generating from generative AI is like less than 10% of their sales. So, how long can this gap persist? Right. And then more importantly, are we kind of midway through how much of their spending can be spent on your products? So just I think in the past you have given us kind of a trillion-dollar market, going to $2 trillion. If you could just educate us on how large the market is? And where are we in that adoption curve based on how much it can be -- based on how much it's being monetized in the near-to-medium term?

嗨。我是来自美国银行证券的 Vivek Arya。谢谢，Jensen。感谢 Colette 的演示。所以 Jensen，我的问题可能更接近中期，即可寻址市场的规模，因为您的收入增长得如此迅速。当我看到它们在一些大客户支出中所占比例时，它们有时达到 30％、40％、50％，甚至更多，但当我看到它们从生成式人工智能中产生的收入占其销售额的比例不到 10％。那么，这种差距会持续多久呢？对。更重要的是，我们是否已经过了大客户在您产品上的支出能达到多少的中间阶段？所以我认为在过去您曾告诉我们一个万亿美元的市场，将增长到 2 万亿美元。如果您能告诉我们市场有多大？以及根据在中期内能够实现多少货币化来看，我们在采用曲线中处于什么位置？

Jensen Huang 黄仁勋

Okay. I'm going to first give you the super-condensed version, and I'll come back and work it out. Okay. So the answer for how big the market is? How big we can be has to do with the size of the market and what we sell. Remember, what we sell is a data center. I just broke it into parts. But in the end, I sold the data center. Notice that the last image you saw at the keynote, it's a reminder of what we actually sell. We showed a bunch of chips. But remember, we don't really sell that. The chips don't work all by themselves. You can buy the chips, but they don't work. You need to build them into our system. And most importantly, the system software and the ecosystem stack is really complicated. And so NVIDIA builds entire data centers for AI. And we just break it up into parts of that. It fits into your company. So that's number one. What do we sell? And what is the opportunity? The opportunity for the world today, the data center size is $1 trillion. Right. And it's a $1 trillion worth of installed, $250 billion a year. We sell an entire data center in parts and so our percentage of that $250 billion per year is likely a lot, lot, lot higher than somebody who sells a chip. It could be a GPU chip or CPU chip or networking chip. That opportunity hasn't changed from before. But what NVIDIA makes is an accelerated computing platform data center scale. Okay. And so our percentage of $250 billion will likely be higher than the past. Now, second question. How sustainable is it? There are two answers for that. One reason that you buy NVIDIA is for AI. If you just build TPUs, if your GPU is only used for one application, then you have to hang your hat on a 100% of that. What can you monetize of AI today? Token generation returns. However, if your value proposition is that AI token generation but that AI training the model and very importantly, reducing the cost of expense of computing, accelerated computing, sustainable computing, energy-efficient computing that's what NVIDIA does for a living at its core. It's just we did it so well that generative AI was created. Okay. And now people forgot that it's a little bit like our first application was computer graphics. And the first application was games. We did that so well, we did it so passionately people forgot, we are accelerated computing company. They thought, hey, you're a gaming company, and a whole generation of young people grew up. And once they learn, they use RIVA 128 and they went to college with GeForce, and then when they finally became an adult, they thought you were a gaming company. And so -- we just do -- we do accelerated computing so well. We do AI so well, people think that that's all we do. But accelerated computing is a trillion -- it's $250 billion a year. $250 billion a year should go to accelerated computing with or without AI, just for the sake of a sustainable computing, just to process SQL, which is, as you guys know, one of the largest consumption of computing in the world. Okay. So I would say $250 billion a year should go to accelerated computing no matter what. And then on top of there is generative AI. How sustainable do I think generative AI is going to be? You know how I feel about it. I think we're going to be generating words, images, videos, proteins, chemicals, kinetic action, manipulation. We're going to be generating forecasts. We're going to be generating bill plans. We're going to be generating bill of materials, we're going to be generating list goes on.

好的。我将首先给你一个超级压缩版本，然后再回来详细解释。好的。那么市场有多大的答案？我们可以有多大与市场规模和我们销售的产品有关。记住，我们销售的是数据中心。我只是把它分成了几部分。但最终，我卖掉了数据中心。请注意，在主题演讲中你看到的最后一个图像，它提醒了我们实际销售的东西。我们展示了一堆芯片。但请记住，我们实际上并不销售那些芯片。芯片本身无法独立工作。你可以购买芯片，但它们无法工作。你需要将它们构建到我们的系统中。最重要的是，系统软件和生态系统堆栈非常复杂。因此，英伟达为人工智能构建整个数据中心。我们只是把它分成了几部分。它适合你的公司。这就是第一点。我们销售什么？机会是什么？今天的机会，数据中心规模是 1 万亿美元。是的。这是价值 1 万亿美元的已安装设备，每年 2500 亿美元。我们将整个数据中心分成几部分销售，因此我们每年 2500 亿美元的份额可能比销售芯片的人高得多。 这可能是 GPU 芯片、CPU 芯片或网络芯片。这个机会与以前没有变化。但 NVIDIA 所做的是一个加速计算平台数据中心规模。好的。因此，我们在 2500 亿美元中的比例可能会高于过去。现在，第二个问题。这是可持续的吗？对此有两个答案。购买 NVIDIA 的一个原因是为了人工智能。如果你只建造 TPU，如果你的 GPU 只用于一个应用程序，那么你必须全力以赴。你今天可以从人工智能中获得什么收益？代币生成回报。然而，如果你的价值主张是 AI 代币生成，但是 AI 训练模型并且非常重要的是，降低计算费用、加速计算、可持续计算、节能计算，这就是 NVIDIA 在其核心所做的。我们做得如此出色，以至于生成式人工智能被创造出来。好的。现在人们忘记了，这有点像我们的第一个应用是计算机图形。第一个应用是游戏。我们做得如此出色，我们如此热情，人们忘记了，我们是一家加速计算公司。 他们认为，嘿，你是一家游戏公司，整整一代年轻人成长起来了。一旦他们学会了，他们用 RIVA 128，然后带着 GeForce 去上大学，最终成年后，他们认为你是一家游戏公司。所以 - 我们只是 - 我们做加速计算做得很好。我们做 AI 做得很好，人们认为那是我们唯一做的事情。但加速计算是一个兆级市场 - 每年 2500 亿美元。每年 2500 亿美元应该用于加速计算，无论是否涉及 AI，仅仅为了可持续计算的缘故，仅仅为了处理 SQL，正如你们所知，这是世界上最大的计算消耗之一。好的。所以我会说，每年应该有 2500 亿美元用于加速计算，无论如何。然后在这之上是生成式 AI。我认为生成式 AI 会有多可持续？你知道我是怎么想的。我认为我们将会生成文字、图像、视频、蛋白质、化学品、动力学行为、操作。我们将会生成预测。我们将会生成账单计划。我们将会生成物料清单，我们将会生成等等。

Stacy Rasgon

Hi, Jensen, Colette. Thanks. It's Stacy Rasgon, Bernstein Research. I wanted to ask about the interplay between CPUs and GPUs. Most of the benchmarks, if not all of them, that you showed yesterday, were really around the Grace Blackwell system that had, I guess, two GPUs and one CPU sort of doubled the CPU per GPU ratio versus Grace Hopper. You didn't talk a lot about benchmarks relative to the standalone GPUs. Is this a shift? Are you guys looking for much more CPU content, I guess, in these AI servers going forward? And then how do I think about the interplay between the ARM CPUs that you're developing and x86 seems like you're putting a little less emphasis on the x86 side of things going forward.

嗨，詹森，科莱特。谢谢。我是斯泰西·拉斯贡，伯恩斯坦研究。我想问一下关于 CPU 和 GPU 之间的相互作用。昨天你展示的大部分基准测试，如果不是全部，都是关于 Grace Blackwell系统的，我想，有两个 GPU 和一个 CPU，CPU 每 GPU 比例是 Grace Hopper 的两倍。你没有谈论很多关于独立 GPU 的基准测试。这是一个转变吗？你们是否在未来寻求更多的 CPU 内容，我想，在这些 AI 服务器中？然后，我该如何考虑你们正在开发的 ARM CPU 和 x86 之间的相互作用，似乎你们在未来对 x86 方面的重视稍微减少了一些。

Jensen Huang 黄仁勋

Yeah, Stacy. Appreciate the question. You the -- there is actually zero concern about either one of them. I think x86 and ARM are both perfectly fine for data centers. There's a reason why Grace is built, the way it is. Grace is built in such a way, the benefit of ARM is that we could mold the NVIDIA system architecture around the CPU. So that we can create this thing called chip to chip, the NVLink that connects between the GPU and the CPU. We can make the two sides coherent, meaning, when the CPU touches a register it invalidates the same register on the GPU side. As a result, the two sides can work together on one variable coherently. You can't do that today between x86 and peripherals and so we were able to solve some problems that we couldn't solve otherwise. And as a result, Grace Hopper is insanely great for CAE applications which is multi-physics. Some of it is running on CPUs, some of it is on GPUs. It's insanely great for different combinations of CPU and GPUs. So that we can have very large memories associated with each maybe one GPU or two GPU coherently. And so we can solve some of these problems, data processing, for example, insanely great on Grace Hopper. Okay. And so it's just harder to solve not because the CPU itself but because we couldn't adopt the system. Second, the reason why I showed I will say that there was one chart where I showed Hopper versus Blackwell on x86 systems B100, B200 and then also GB200 which is the Grace Blackwell. The benefit of Blackwell in that case wasn't because the CPUs better. It's because in the case of Grace Blackwell we were able to create a larger NVLink domain. And that larger NVLink domain is really, really important for the next generation of AI. The next three years, the next three -- five years, which is, as far as we can see right now. If you really want a good inference performance, you're going to need NVLink. That was the message, I was trying to deliver. And we're going to talk more about this. It's abundantly clear now, these large language models, they're never going to fit on one GPU. Okay. That's not the point, anyways. And in order for you to be sufficiently responsive and have high throughput to keep the cost down, you need a lot more GPUs than what you even fit in. And in order to have a lot of GPUs working together without the overhead, the IO overhead getting in the way you need NVLink. NVLinks benefit and inference every -- always thought NVLinks benefit is in training. NVLinks benefit and inference is off the charts. That's the difference between 5X and 30X that was another 6X, it's all NVLink. NVLinks in the new Tensor Core, excuse me. Yeah, okay. And so the Grace gives us the ability to architect a system exactly as we needed and it's harder to do it with x86. That's all. But we support both. We'll have two versions of both. And in the case of B100 it just slides into where H100 and H200 goes into. And so the adoption of transition for Hopper to Blackwell is instantaneous. The moment it's available you just slide it in and then you can figure out what to do about the next data center. Okay. So we get the benefit of extremely excellent performance at its limit of the architecture as well as easy-peasy transition.

是的，Stacy。感谢提问。实际上，对它们两者都没有任何担忧。我认为 x86 和 ARM 对数据中心都非常适用。Grace 被构建的原因是有道理的。ARM 的好处在于我们可以围绕 CPU 塑造 NVIDIA 系统架构。这样我们就可以创建这个称为芯片对芯片的东西，即连接 GPU 和 CPU 之间的 NVLink。我们可以使两侧一致，也就是说，当 CPU 触及一个寄存器时，会使 GPU 侧相同的寄存器无效。因此，两侧可以在一个变量上协同工作。今天在 x86 和外围设备之间无法做到这一点，因此我们能够解决一些我们无法解决的问题。结果，Grace Hopper 对于多物理量的 CAE 应用非常出色。一些应用在 CPU 上运行，一些应用在 GPU 上运行。对于 CPU 和 GPU 的不同组合非常出色。这样我们可以使每个 GPU 或两个 GPU 关联的内存非常大。因此，我们可以解决一些问题，例如数据处理，在 Grace Hopper 上非常出色。好的。 因此，解决问题变得更加困难，不是因为 CPU 本身，而是因为我们无法采用这个系统。其次，我展示的原因是，有一个图表显示了 Hopper 与 x86 系统 B100、B200 以及 GB200（即 Grace）之间的对比。在这种情况下的好处并不是因为 CPU 更好，而是因为在 Grace 的情况下，我们能够创建一个更大的 NVLink 领域。这个更大的 NVLink 领域对于下一代人工智能非常重要。接下来的三年，接下来的三至五年，就目前而言，我们可以看到。如果您真的想要良好的推理性能，您将需要 NVLink。这是我试图传达的信息。我们将会更多地讨论这个问题。现在非常明显，这些大型语言模型永远不会适合一个 GPU。好吧。无论如何，为了让您具有足够的响应能力和高吞吐量以降低成本，您需要比您能够容纳的 GPU 更多的 GPU。 为了让许多 GPU 一起工作而没有额外开销，IO 开销不会成为障碍，您需要 NVLink。NVLink 的好处和推理每一次--总是认为 NVLink 的好处在训练中。NVLink 在推理中的好处是非常突出的。这就是 5 倍和 30 倍之间的差异，另外还有 6 倍，这都是 NVLink。新的 Tensor Core 中有 NVLink，对不起。是的，好的。所以 Grace 给了我们按照我们需要的方式设计系统的能力，用 x86 做这件事更难。就是这样。但我们支持两者。我们将有两个版本。在 B100 的情况下，它只需插入到 H100 和 H200 的位置。因此，从 Hopper 过渡到Blackwell的采用是瞬间的。一旦可用，您只需将其插入，然后可以考虑下一个数据中心的情况。好的。因此，我们既获得了极其出色的性能优势，又实现了轻松的过渡。

Stacy Rasgon

Thank you. 谢谢。

Matt Ramsay 马特·拉姆齐

Hey there. It's Matt Ramsay from TD Cowen. Hey, Jensen, Colette. Thank you. Good morning for doing this. I wanted to -- Jensen for you to comment on a couple topics that I've been noodling on. One of which is NIMs that you guys talked about yesterday, it seems like a vertical-specific accelerant for people to get into AIE and onboard customers more quickly. I wonder if you could just give us an overview of how your company is going at broader enterprise and just what different vehicles there are for people to onboard into AI? The second topic is on power. My team has been spending a good bit of time on power. I'm trying to decide if I should spend more time there or less. Some of the systems you introduced yesterday are up to 100 kilowatts or more. I know that scale of computing couldn't be done without the integration that you guys are doing, but also we are getting questions on power generation at the macro-level, power delivery to the cabinet at that density. I just would love to hear your thoughts about how your company is working with the industry to power these systems. Thanks.

嘿，你好。我是 TD Cowen 的马特·拉姆齐。嗨，詹森，科莱特。谢谢你们。感谢你们早上做这件事。我想让——詹森，你来评论一下我一直在思考的几个话题。其中之一是你们昨天谈到的 NIMs，似乎是一个垂直特定的加速剂，让人们更快地进入 AIE 并快速接入客户。我想知道你能否给我们一个关于你们公司如何在更广泛的企业中发展以及人们如何接入 AI 的概述？第二个话题是关于电力。我的团队一直在花费大量时间在电力上。我正在考虑是否应该在这方面花更多时间还是更少时间。你们昨天介绍的一些系统的功率高达 100 千瓦甚至更高。我知道这种规模的计算是无法做到没有你们正在进行的集成的，但同时我们也在宏观层面上对电力生成、以及在这种密度下将电力传送到机柜上有疑问。我很想听听你们如何与行业合作来为这些系统提供电力的想法。谢谢。

Jensen Huang 黄仁勋

Okay. I'll start with the second first. Power delivery, 100 kilowatts as you know for computer is a lot, but 100 kilowatts is a commodity, you guys know that, right. The world needs a lot more than 120 kilowatts. And so the absolute amount of power is not an issue. The delivery of the power is not an issue. And the physics of delivering the power is not an issue. And cooling 120 kilowatts is not an issue. We can all agree on that. Okay. And so none of this is a physics problem. None of this requires invention. All of it require supply chain planning. Makes sense. So that's the way. And how big of a deal is supply chain planning? A lot. I mean, we take it very seriously. And so we think about supply chain planning for all the time and you got to go at, the reason why we have great partnerships with. If you go -- I think if you look at Vertiv, I think the front pages of paper that we wrote together. So Vertiv and NVIDIA engineers working on cooling systems. Okay. And so Vertiv is very important in the supply chain of designing liquid cooled and otherwise data centers. We have great partnerships with Siemens. We have great partnerships with Rockwell, Schneider for all the reasons. This is exactly the same as having great partnerships with TSMC and Samsung and SPIL and Wistron and so on and so forth. And so we're going to have to go -- our company supply chain relationships are quite broad and quite deep. And thus the fact that we build our own data centers, really help that. We've been building supercomputers now for quite some time. This is not our first time. Our first supercomputer was DGX-1 in 2016 that kind of puts in perspective. And we've built one every year and this year we're building several. And so the fact that we're building it, it gives us tactile sensation of who we're working with, who are the best and we do it for that very reason, one of the reasons for that. NIMs. There are two onboard. Two-ways to onboard into enterprise. There is the most impactful way. And then there's the other way. Okay. They're both important. The other -- I'll start with the other. The other way is that we are going to create these NIMs. We are going to put it on our website. And we're going to go through GSIs and a lot of solution providers and they're going to help companies turn these NIMs into applications. And that's going to have a whole thing. That's going to have a whole thing, okay. And so that go-to-market includes large GSIs and smaller specialized GSIs and so on and so forth okay. We have lots of partnerships in that area. The other area that I think it's really quite exciting. And I think that this is really where big action is going to happen is the trillion dollars of enterprise companies in the world. They create tools today. In the future they're going to offer you tools plus copilots. Remember, the single most pervasive tool in the world is Office. And now copilots for office. There is another tool that is super important to NVIDIA Synopsys, Cadence, Ansys. We would like to have copilots for all of them. Notice we were building copilots for our own tools. We call them ChipNeMo. And ChipNeMo is super smart. And ChipNeMo now understands NVIDIA Lingo, NVIDIA Chip Talk and it knows how to program NVIDIA programs. And so every engineer that we hire the first thing we're going to tell them, here's ChipNeMo, and then there's the bathroom, and then there's the cafeteria, and so, in that order. And so they will be productive right away whether you lunch, they could ChipNeMo could be doing some stuff. And so that just gives you an example. But we have copilots are being built on top of our own tools all over the place. Most companies probably can't do this, and we can teach the GSIs to do this, but in the area of these tools Cadence and others, they're going to build their own copilots. And they will rent them out as hire them out as engineers. I think they're sitting on a goldmine. SAP is going to do that. ServiceNow is going to do that and they're very specialized copilots. They understand languages like -- in the case of SAP, ABAP is that right, which is a language that only an SAP lever would love and as you know, ABAP is a very important language for the world's ERP systems. Every company runs on it. We use ABAP. And so now they have to go create a Chat ABAP and that Chat ABAP, just like ChipNeMo or ChatUSD that we created for Omniverse and so Siemens will do that, Rockwell will do that, so on so forth. Does that makes sense? And that I think is another way, you get to enterprise and that ServiceNow is going to do that. Lots and lots of copilots they're building. And that's how they can create another industry on top of their current industry, it's almost like an AI workforce industry. Yeah. I am super excited about the partnerships we have with all of them. Just I'm so excited for them. Every time I see them, I just -- I tell them, anywhere you're sitting on a goldmine, you're sitting on a goldmine. I mean I'm so excited for them.

好的。我会先从第二个开始。电力传输，正如你们所知，对于计算机来说 100 千瓦是很多的，但是 100 千瓦是一种商品，你们知道的，对吧。世界需要的远远超过 120 千瓦。因此，电力的绝对数量不是问题。电力的传输不是问题。传输电力的物理学也不是问题。冷却 120 千瓦也不是问题。我们都可以达成一致。好的。因此，这些都不是物理问题。这些都不需要发明。所有这些都需要供应链规划。有道理。这就是方法。供应链规划有多重要？非常重要。我的意思是，我们非常认真对待它。因此，我们一直在考虑供应链规划，你必须去做，我们与伟易升有着很好的合作关系的原因。如果你去看一下，我认为如果你看一下我们一起写的论文的封面。所以伟易升和英伟达的工程师们正在研究冷却系统。好的。因此，伟易升在设计液冷和其他数据中心的供应链中非常重要。我们与西门子有着很好的合作关系。我们与罗克韦尔、施耐德也有着很好的合作关系，出于各种原因。 这与与台积电、三星、SPIL、Wistron 等伙伴关系密切一样。因此，我们必须继续前进——我们公司的供应链关系非常广泛且深入。因此，我们自建数据中心的事实确实有所帮助。我们已经建造超级计算机相当长时间了。这不是我们第一次。我们的第一台超级计算机是 2016 年的 DGX-1，这让人有所了解。我们每年都会建造一台，今年我们将建造几台。因此，我们正在建造它，这让我们有了与谁合作、谁是最好的的实际感觉，我们出于这个原因之一才这样做。NIMs。有两种内置方式。两种进入企业的方式。有一种方式最具影响力。然后还有另一种方式。好的。它们都很重要。另一种——我将从另一种方式开始。另一种方式是我们将创建这些 NIMs。我们将把它放在我们的网站上。我们将通过 GSIs 和许多解决方案提供商，他们将帮助公司将这些 NIMs 转化为应用程序。这将带来整个事情。 这将是一个整体的事情，好的。因此，市场推广包括大型 GSIs 和较小的专业 GSIs 等等。我们在这个领域有很多合作伙伴关系。我认为另一个领域真的很令人兴奋。我认为这才是大动作将会发生的地方，那就是全球万亿美元的企业公司。他们今天创造工具。将来，他们将为您提供工具和副驾驶。请记住，世界上最普遍的工具是 Office。现在还有 Office 的副驾驶。对于 NVIDIA Synopsys、Cadence、Ansys 等公司来说，另一个非常重要的工具是副驾驶。我们希望为它们所有的工具提供副驾驶。请注意，我们正在为我们自己的工具构建副驾驶。我们称它们为 ChipNeMo。ChipNeMo 非常聪明。ChipNeMo 现在了解 NVIDIA 术语、NVIDIA 芯片语言，并且知道如何编程 NVIDIA 程序。因此，我们雇用的每位工程师，我们都会告诉他们的第一件事是，这是 ChipNeMo，然后是洗手间，然后是餐厅，按照这个顺序。因此，无论是吃午饭还是工作，他们都可以立即提高生产力。 这只是一个例子。我们的合作伙伴正在利用我们自己的工具在各个地方构建。大多数公司可能做不到这一点，我们可以教 GSIs 这样做，但在 Cadence 和其他工具领域，他们将构建自己的合作伙伴。他们将把它们出租或雇佣为工程师。我认为他们坐拥一座金矿。SAP 将会这样做。ServiceNow 将会这样做，它们是非常专业的合作伙伴。他们了解诸如 SAP 的语言--在 SAP 的情况下，ABAP 是正确的，这是一种只有 SAP 杠杆会喜欢的语言，正如您所知，ABAP 是世界 ERP 系统中非常重要的语言。每家公司都在使用 ABAP。现在他们必须去创建一个 Chat ABAP，就像我们为 Omniverse 创建的 ChipNeMo 或 ChatUSD 一样，因此西门子将这样做，洛克韦尔将这样做，等等。这样做有道理吗？我认为这是另一种方式，您可以进入企业，ServiceNow 将会这样做。他们正在构建大量的合作伙伴。 这就是他们如何在当前行业的基础上创造另一个行业，几乎就像是一个人工智能劳动力行业。是的。我对我们与他们所有人的合作伙伴关系感到非常兴奋。我为他们感到如此兴奋。每次见到他们，我就会告诉他们，你们坐在一个金矿上，你们坐在一个金矿上。我的意思是我为他们感到如此兴奋。

Tim Arcuri 蒂姆·阿库里

Jensen, hi. It's Tim Arcuri at UBS. I had a question also about the TAM and it's more greenfield versus brownfield, because up until now H100 was pretty much all greenfield. So people weren't taking A100s and ripping them out and replacing them with H100s, could B100 be the first time, where you see some brownfield upgrades where we go in and we rip out A100s and we replace them with B100s? So that may be the TAM, if the $1 trillion goes to $2 trillion, you have a four-year replacement cycle. You're talking about $500 billion, but much of that growth comes from upgrading the existing installed base. Wondering if you can comment on that.

詹森，你好。我是瑞银的蒂姆·阿库里。我也有一个关于 TAM 的问题，主要是关于绿地和棕地的区别，因为直到现在 H100 基本上都是绿地。所以人们并没有拿走 A100 然后用 H100 替换它们，B100 可能是第一次，我们会看到一些棕地升级，我们会拆除 A100 然后用 B100 替换它们吗？所以如果 TAM 从 1 万亿美元增长到 2 万亿美元，你有一个四年的更换周期。你在谈论 5000 亿美元，但其中很大一部分增长来自于升级现有的安装基础。想知道你是否能对此发表评论。

Jensen Huang 黄仁勋

Yeah, really good question. Today, we are upgrading the slowest computers in the data center, which will be the CPUs. And so that's what should happen. And then eventually you'll get around to the Amperes and then you get around to the Hoppers. I do believe that in five, six, seven, eight years, you're going to give you --we're going to be in -- picker you're out there I'm not picking one. I'm just saying in the outer years, you're going to start seeing replacement cycles, obviously, of our own infrastructure. Yeah, but, I wouldn't think that that's the best utilization of capital at the moment. Amperes are super productive as you know.

是的，真的是一个很好的问题。今天，我们正在升级数据中心中最慢的计算机，即 CPU。这就是应该发生的事情。然后最终你会开始使用 Amperes，然后再开始使用 Hoppers。我相信在五、六、七、八年内，我们会开始看到明显的基础设施更换周期。是的，但是，我认为目前这并不是最佳的资本利用方式。正如你所知，Amperes 非常高效。

Brett Simpson 布雷特·辛普森

Yeah. Hi, Jensen. It's Brett Simpson here at Arete Research, and thanks for hosting a great event this last couple of days. My question was on Inference. I wanted to get your perspective on -- you put up some good performance numbers with the B100 in terms of how Inference compares with H100. How -- what's the message you're giving to customers on cost of ownership around this new platform? And how do you think it's going to compare with ASICs or other Inference platforms in the industry? Thank you.

是的。嗨，詹森。我是 Arete Research 的布雷特·辛普森，感谢您在过去几天举办了一场很棒的活动。我的问题是关于推理的。我想听听您的看法——您用 B100 展示了一些很好的性能数据，关于推理与 H100 相比如何？您对客户在这个新平台的拥有成本有什么建议？您认为它将如何与 ASIC 或其他行业推理平台相比？谢谢。

Jensen Huang 黄仁勋

I think for language models, large language models Blackwell with the new transformer engine and NVLink is going to be very, very, very hard to overcome. And the reason for that is the dimensionality of the problem is so large. And TensorRT-LLM this exploration tool, this optimization compiler that I talked about. The architecture underneath the Tensor Cores are programmable. NVLink allows you to connect up whole bunch of GPUs working in tandem with very, very low overhead, basically no overhead. Okay. And so as a result, 64 GPUs is the same as one programmatically. It is incredible. And so when you have 64 GPUs without overhead without this NVLink overhead, if you have to go over the network like Ethernet, it's over. You can't do it. You just wasted everything. And because they all have to communicate with each other, it's called all2all. Whenever all have to communicate each other the slowest link is the bottleneck, right. It's no different than having a city on one side of the river having a city on the other side of the river that bridge that's it. That's the throughput doing -- that defines the throughput. Okay. And that bridge will be Ethernet. On one side is NVLink, on the other side is NVLink Ethernet in the middle makes no sense. So we have to turn that into NVLink. And now we have all of the GPUs working together generating tokens one at a time. Remember the tokens cannot be -- it's not as if you splat out a token because tokens the transformer has to generate the tokens one at a time in sequence. And so this is a very complicated parallel computing problem, okay. And so I think the -- I think Blackwell has raised the bar a lot. Just mountains. Utterly mountains, ASIC or otherwise.

我认为对于语言模型来说，具有新的变压器引擎和 NVLink 的大型语言模型将非常非常非常难以克服。原因在于问题的维度非常大。TensorRT-这个探索工具，这个我谈到的优化编译器。张量核心下面的架构是可编程的。NVLink 允许您连接整个一堆 GPU，以非常非常低的开销，基本上没有开销。好的。因此，64 个 GPU 与一个程序性相同。这是令人难以置信的。因此，当您有 64 个 GPU 没有这种 NVLink 开销时，如果您必须通过像以太网这样的网络进行通信，那就完了。您什么都做不了。您只是浪费了一切。因为它们都必须彼此通信，这被称为全互连。每当所有人都必须彼此通信时，最慢的链接就成了瓶颈，对吧。这与在河的一边有一个城市，在河的另一边有一个城市没有什么不同，那座桥就是它。这就是吞吐量的定义。好的。那座桥将是以太网。 一边是 NVLink，另一边是 NVLink 以太网，中间没有意义。所以我们必须将其转换为 NVLink。现在我们有所有的 GPU 一起工作，一次生成一个令牌。请记住，令牌不能--不是你一下子就生成一个令牌，因为变压器必须按顺序逐个生成令牌。所以这是一个非常复杂的并行计算问题，好吧。所以我认为--我认为Blackwell已经大大提高了标准。只是山。完全是山，无论是 ASIC 还是其他。

C.J. Muse

Hello, Jensen and Colette. C.J. Muse with Cantor. Thank you for hosting this. And it's great to see you both. Question on your pricing strategy. Historically, you talked about the more you buy, the more you save. But it sounds like initial pricing on Blackwell is coming in at perhaps maybe a lower premium than the productivity that you're offering. So, curious, as you think about maybe razor, razor blade and selling software and the full system, how that might cause you to kind of evolve your pricing strategy and how we should think about kind of normalized margins within that construct? Thank you.

你好，詹森和科莱特。C.J. Muse 与康特公司。感谢你们的款待。很高兴见到你们。关于你们的定价策略有个问题。历史上，你们提到购买越多，节省越多。但听起来Blackwell的初始定价可能比你们提供的生产力更低。因此，好奇的是，当你们考虑剃刀、剃须刀片和销售软件以及整个系统时，这可能会如何促使你们改进定价策略，以及我们应该如何考虑在这个框架内的标准化利润率？谢谢。

Jensen Huang 黄仁勋

The pricing that we create always starts from TCO. I appreciate that comment, C.J. We always come from TCO. However, we also want to have the TCO not of the main body of customers. And so when the customers -- when you only have one particular domain of customers, let's say, molecular dynamics, then if it's only one application, then you set the TCO based on that one application. It could be a medical imaging system. And all of a sudden, the TCO is really very, very high, but the market size is quite small. In every single generation that goes by, our market size is growing, isn't that right? And we want to make the entire market be able to afford Blackwell. And so in a way, it's kind of a self-carrying problem. As we solve for the TCO for a much larger problem -- larger market then some customers would get too much value, if you will. But that's okay. But you're making the business simpler, having one basic product and you're able to support a very, very large market. Now, over time, if the market were to bifurcate, then we can always segment, but that's we're nowhere near that today. And so I think we have the opportunity to create a product that delivers extraordinary value for many and extremely good value for all. And that's our purpose.

我们制定的定价始终从 TCO 开始。我很感激您的评论，C.J.我们始终从 TCO 出发。但是，我们也希望客户的 TCO 不是主体。所以当客户——当您只有一个特定领域的客户，比如说，分子动力学，那么如果只有一个应用程序，那么您可以根据该应用程序设置 TCO。它可能是医学成像系统。突然之间，TCO 真的非常非常高，但市场规模相当小。在每一代中，我们的市场规模都在增长，对吧？我们希望整个市场都能负担得起Blackwell。所以在某种程度上，这是一种自我承载的问题。当我们解决更大问题的 TCO 时——更大的市场，那么一些客户可能会获得太多价值，如果您愿意的话。但没关系。但是，通过简化业务，只有一个基本产品，您就能支持一个非常非常大的市场。随着时间的推移，如果市场分化，那么我们总是可以分割，但是我们今天还远远没有达到那个地步。 所以我认为我们有机会创造一个为许多人提供非凡价值，对所有人都提供极好价值的产品。这就是我们的目的。

Joseph Moore 约瑟夫·摩尔

Hi. Joe Moore from Morgan Stanley. It seems like the most impressive specs that you showed were around GB200, which you just described as a function of having that bigger NVLink domain. Can you contrast what you're doing with GB200 with what you did with GH200? And why you think it could be a much bigger product this time around?

嗨。摩根士丹利的乔·摩尔。你展示的最令人印象深刻的规格似乎是围绕 GB200 的，你刚刚描述它是拥有更大的 NVLink 域的一个功能。你能对比一下你在 GB200 上的做法和你在 GH200 上的做法吗？为什么你认为这次可能会是一个更大的产品？

Jensen Huang 黄仁勋

Oh, great question. The simple answer is GH200, 100, 200, Grace Hopper, before it could really take off significantly, Grace Blackwell is already here. And Grace Hopper had the additional burden that Hopper didn't have. Hopper fit right into where Ampere left off. A100s went to H100s, they're going to go to B100s, so on and so forth. And so that particular chassis or that particular use case is fairly well established and we'll just keep on moving. Software is built for it. People know how to operate it so on and so forth. Grace Hopper is a little different and it addressed a new class of applications that we didn't address very well before. And I was mentioning some of it earlier. Multi physics problems with a CPU and GPU was having to work closely together, very large data sets, so on and so forth. Difficult to paralyze, for example, those kind of problems Grace Hopper was really good for. And so we started developing software for that. My recommendation for most customers is, at this point, just gear for Grace Blackwell and I have given them that recommendation. And so everything that they do with Grace Hopper will be completely architecturally compatible. That's the wonderful thing. And so, whatever they have, whatever they buy is still fantastic, but I would recommend that they put all their energy into Grace Blackwell because it's so much better.

哦，好问题。简单的答案是 GH200，100，200，Grace Hopper，在它真正起飞之前，Grace Blackwell已经在这里了。Grace Hopper 还有一个额外的负担，Hopper 没有。Hopper 正好适应了 Ampere 的延续。A100s 转向 H100s，它们将转向 B100s，依此类推。因此，那个特定的机箱或特定的用例已经相当成熟，我们将继续前进。软件已经为此构建。人们知道如何操作等等。Grace Hopper 有点不同，它解决了我们以前没有很好解决的一类应用。我之前提到了一些。CPU 和 GPU 一起工作的多物理问题，非常庞大的数据集等等。例如，难以并行化的问题，Grace Hopper 非常擅长。因此，我们开始为此开发软件。我对大多数客户的建议是，在这一点上，只需为 Grace Blackwell做好准备，我已经给过他们这个建议。因此，他们在 Grace Hopper 上所做的一切都将完全符合架构。 这就是美妙之处。所以，无论他们拥有什么，无论他们购买什么，仍然是很棒的，但我建议他们把所有精力投入到 Grace Blackwell，因为它更好。

Unidentified Analyst 未知分析师

Jensen, Collete, thanks for having us here today. I want to ask a question on robotics. It seems like every time we come back to GTC, you sneak something at the end. And in a couple years, we go, wow, he has been talking about that for a while. I heard this week you guys mentioned that robotics may be getting close to its ChatGPT moment. Can you describe what that means and where you start to see that robotics evolution kind of like our day to day lives? That would be super helpful. Thank you.

詹森，科莱特，感谢今天邀请我们来这里。我想问一个关于机器人技术的问题。每次我们回到 GTC，你似乎总是在最后偷偷摸摸地加入一些东西。几年后，我们会惊叹，哇，他一直在谈论这个。我听说这周你们提到机器人技术可能接近其 ChatGPT 时刻。你能描述一下这意味着什么，以及你开始看到机器人技术演变如何影响我们的日常生活吗？那将非常有帮助。谢谢。

Jensen Huang 黄仁勋

Okay, several things. First of all, I appreciate that. I showed Earth-2, two years ago. And two years later, we have this new algorithm that is able to do regional weather prediction at 3 kilometers. The supercomputer you need to do that is 25 times larger, excuse me, 25,000 times larger than the one that you currently use to do weather simulations at NOA and in Europe and so on and so forth. 3 kilometer resolution is very high resolution, if you will, right above your head, okay. And weather simulation also requires a whole lot of what is called ensembles because the world looks chaotic and you want to simulate a lot of distribution, sample a lot of different parameters, a lot of different perturbations, and try to figure out what is that distribution and that the middle of that distribution likely is going to be the weather pattern. Well, if it takes that much energy just to do it one time, they're not going to do it more than one time. But in order to predict where weather is going to be a week from now, especially extreme weather that can change so dramatically, you're going to need a lot of what they call members, a lot of ensemble members, a lot of samplings. And so you're basically doing -- we're basically doing weather simulation 10,000 times, okay. And because we train an AI to understand physics and it's physically possible and it can't hallucinate, so it has to understand the laws of physics and such. And so two years ago, I showed it today and we connected into the most trusted source of weather in the world, the weather company. And so we're going to help people do regional weather all over the world. If you're a shipping company and you need to know weather conditions. If you're an insurance company, you need to know weather conditions. If you're in the Southeast Asia region, you have so many hurricanes and typhoons and things like that, you need some of this technology. And so we're going to help people adapt it for their region and their use case. Well, I did that a couple of years ago. The ChatGPT moment kind of works like this. Take a step back and ask yourself what happened with ChatGPT? The technology is insanely great, okay. It's really incredible. But there's several things that happened. One, it learned from a whole lot of human examples. We wrote the words, right? It was our words. So it learned from our human examples and it generalized it. So it's not repeating back the words. So it can understand the context and it can generate a regional form. It understood the context meanings that it adapted to itself, okay, or it adapted to the current circumstance, the context. And then the third thing is, it could now generate original tokens. Now I'm going to take everything back into tokens. Forget words, just tokens now. Use all the same words that I just use, but replace words with tokens. If I could just figure out how to communicate with this computer, what this token means? Okay, if I can just tokenize this. Just as when you do speech recognition, you tokenized my sound, my voice. Just as when we reconstructed proteins, we tokenized the amino acids. You can tokenize almost everything. You can digitize a simple way of representing each chunk of the data, okay. So once you can tokenize it, then you can learn it. We call it learning the embeddings of it, the meanings of it. And so if I can tokenize motion, okay, the world and I can generalize and I can tokenize articulation, kinematics, and I can learn and generalize it and then generate just. I just did the ChatGPT moment, how is it any different? The computer doesn't know. Now, of course, the problem space is a lot more complicated because it's physical things. So, you need this thing called alignment. And what was the great invention of ChatGPT, reinforcement, learning human feedback alignment. Is that right? So, it would try something. You say no, that's not as good as this. It would try something else. You said, no, that's not as good as this. Human feedback, reinforcement learning and it keeps -- it takes that reinforcement and improves itself. And so what is Omniverse for? Well, if it's in a robot, then how would you do feedback? And what is feedback about? It's physical feedback, physics feedback. It generalized -- it generated a movement to go pick up a cup, but it tipped a cup over. It needs the reinforcement learning to know when to stop. Does that make sense? And so that feedback system is not human. That feedback system is physics. And that physics simulation feedback is called Omniverse. So Omniverse is reinforcement learning, physical feedback, which grounds the AI to the physical world, just as reinforcement learning human feedback grounds the AI to human values. Are you guys following me? I just described two completely different domains using exactly the same concepts. And so what I've done is I've generalized general AI. And by generalizing it, I can reapply it somewhere else. And so we made this observation some time ago and we started preparing for this. And now you're going to find that Isaac Sim, which is a gym on top of Omniverse is going to be super, super successful for just about anybody who is doing these robotic systems. We've created the operating system for robots. I'm sure there's a corporate answer for all the questions you guys ask, but unfortunately, I only know how to answer the one geek way.

好的，几件事情。首先，我很感激。两年前，我展示了 Earth-2。两年后，我们有了这个新算法，可以在 3 公里的范围内进行区域天气预测。你需要的超级计算机比你目前用来进行 NOA 和欧洲等地天气模拟的计算机大 25 倍，对不起，是大 25000 倍。如果你愿意的话，3 公里的分辨率是非常高的分辨率，就在你头顶上方。天气模拟还需要大量的所谓集合，因为世界看起来混乱，你想模拟很多分布，对很多不同的参数进行抽样，对很多不同的扰动进行抽样，然后尝试弄清楚那个分布是什么，那个分布的中间可能是天气模式。嗯，如果仅仅做一次就需要那么多能量，他们不会做超过一次。但是为了预测一周后天气会在哪里，尤其是那种可能发生剧烈变化的极端天气，你需要很多所谓的成员，很多集合成员，很多抽样。 所以基本上你在做的是 - 我们基本上在做天气模拟 10,000 次，好吧。因为我们训练了一种人工智能来理解物理学，它是物理上可能的，它不能产生幻觉，所以它必须理解物理定律等等。两年前，我展示了它，今天我们将其连接到世界上最值得信赖的天气信息来源，即天气公司。因此，我们将帮助人们在全球范围内进行区域天气预报。如果你是一家航运公司，需要了解天气条件。如果你是一家保险公司，需要了解天气条件。如果你在东南亚地区，有很多飓风和台风等等，你需要一些这种技术。因此，我们将帮助人们为他们的地区和使用情况调整它。嗯，我几年前就做到了。ChatGPT 的时刻有点像这样。退一步，问问自己 ChatGPT 发生了什么？这项技术非常出色，好吧。它真的很不可思议。但发生了几件事情。首先，它从大量的人类示例中学习。我们写了这些话，对吧？这是我们的话。 因此，它从我们人类的示例中学到了，并且进行了泛化。所以它不是简单地重复单词。它可以理解上下文，并生成一个区域形式。它理解了上下文的含义，并适应了自身，或者适应了当前的情况，上下文。然后第三件事是，它现在可以生成原始令牌。现在我要把所有东西都转换成令牌。忘记单词，现在只有令牌。使用我刚才使用的所有相同单词，但用令牌替换单词。如果我能弄清楚如何与这台计算机进行通信，这个令牌意味着什么？好的，如果我能够对其进行令牌化。就像当你进行语音识别时，你对我的声音进行了令牌化。就像当我们重构蛋白质时，我们对氨基酸进行了令牌化。你几乎可以对一切进行令牌化。你可以将数据的每个块数字化，这是一种简单的表示方式。一旦你对其进行令牌化，那么你就可以学习它。我们称之为学习其嵌入，其含义。所以如果我能对运动进行令牌化，世界和我可以进行泛化，我可以对关节运动、运动学进行令牌化，学习并泛化，然后生成。 我刚刚做了 ChatGPT 的时刻，它有什么不同？计算机不知道。现在，当然，问题空间更加复杂，因为涉及到物理事物。所以，你需要这个叫做对齐的东西。ChatGPT 的伟大发明是什么，强化学习人类反馈对齐。是这样吗？所以，它会尝试一些东西。你说不，这不如这个好。它会尝试其他东西。你说，不，这不如这个好。人类反馈，强化学习，它会接受这种强化并改进自己。那么 Omniverse 是用来做什么的？嗯，如果它是一个机器人，那么你如何进行反馈？反馈是关于什么的？它是物理反馈，物理反馈。它通用——它生成了一个动作去拿起一个杯子，但是把杯子打翻了。它需要强化学习来知道何时停止。这有意义吗？所以这个反馈系统不是人类的。这个反馈系统是物理的。而那个物理模拟反馈被称为 Omniverse。 所以 Omniverse 是强化学习、物理反馈，将 AI 接地到物理世界，就像强化学习人类反馈将 AI 接地到人类价值观一样。你们跟上我了吗？我刚刚用完全相同的概念描述了两个完全不同的领域。所以我所做的是泛化了通用人工智能。通过泛化，我可以将其重新应用到其他地方。我们一段时间前就注意到了这一点，开始为此做准备。现在你们会发现，Isaac Sim，这是 Omniverse 之上的一个健身房，对于任何从事这些机器人系统的人来说都会非常成功。我们已经为机器人创建了操作系统。我相信你们问的所有问题都有一个公司的答案，但不幸的是，我只知道如何用一种极客的方式回答。

Atif Malik 阿蒂夫·马利克

Hi. I am Atif Malik from Citigroup. I have a question for Colette. Colette in your slides, you talked about availability for the Blackwell platform later this year. Can you be more specific? Is that the October quarter or the January quarter? And then on the supply chain, readiness for the new products is the packaging, particularly on the B200 CoWoS-L and how you are getting your supply chain ready for the new products?

嗨。我是花旗集团的 Atif Malik。我有一个问题要问 Colette。Colette，在你的幻灯片中，你谈到了Blackwell平台今年晚些时候的可用性。你能更具体一些吗？是十月季度还是一月季度？然后在供应链方面，新产品的准备情况是包装，特别是 B200 CoWoS-L，以及你如何让供应链为新产品做好准备？

Colette Kress 柯莱特·克雷斯

Yeah, so let me let me start with your second part of the question, talking about the supply-chain readiness. That's something that we've been working well over a year getting ready for these new products coming to market. We feel so privileged to have the partners that work with us in developing out our supply chain. We've continued to work on resiliency and redundancy. But also, you're right, moving into new areas, new areas of CoWoS, new areas of memory, and just a sheer volume of components and complexity of what we're building. So that's well on its way and will be here for when we are ready to launch our products. So there is also a part of our supply chain as we talked earlier today, talking about the partners that will help us with the liquid cooling and the additional partners that will be ready in terms of building out the full of the data center. So this work is a very important part to ease the planning and the processing to put in all of our Blackwell different configurations. Going back to your first part of the question, which is when do we think we're going to come to market? Later this year, late this year, you will start to see our products come to market. Many of our customers that we have already spoken with talked about the designs, talked about the specs, have provided us their demand desires. And that has been very helpful for us to begin our supply chain work, to begin our volumes and what we're going to do. It's very true though that on the onset of the very first one coming to market, there might be constraints until we can meet some of the demand that's put in front of us. Hope that answers the question.

是的，让我从你问题的第二部分开始，谈论供应链准备情况。这是我们一直在努力工作的事情，已经有一年多的时间为新产品上市做准备。我们感到非常荣幸能与合作伙伴一起开发我们的供应链。我们继续致力于弹性和冗余性。但同时，你说得对，我们正在进入新领域，CoWoS 的新领域，记忆的新领域，以及我们正在构建的组件数量和复杂性。所以这方面已经在顺利进行，当我们准备好推出产品时，这方面也会准备就绪。正如我们今天早些时候所讨论的，我们的供应链中还有一部分，谈论将帮助我们进行液冷和其他合作伙伴将准备好建设完整数据中心的合作伙伴。这项工作是减轻规划和处理负担的非常重要的一部分，以适应我们的 1001 种不同配置。回到你问题的第一部分，我们认为我们什么时候会上市？今年晚些时候，今年晚些时候，你将开始看到我们的产品上市。 我们已经与许多客户交谈过，他们谈到了设计，谈到了规格，向我们提供了他们的需求愿望。这对我们开始供应链工作，开始我们的产量以及我们将要做的事情非常有帮助。尽管在第一个产品上市之初，可能会有一些限制，直到我们能够满足一部分提出的需求。希望这回答了问题。

Jensen Huang 黄仁勋

Yeah, That's right. And just remember that Hopper and Blackwell, they're used for people's operations and people need to operate today. And the demand is so great for Hoppers. They -- most of our customers have known about Blackwell now for some time, just so you know. Okay, so they've known about Blackwell. They've known about the schedule. They've known about the capabilities for some time. As soon as possible, we try to let people know so they can plan their data centers and notice the Hopper demand doesn't change. And the reason for that is they have an operations they have to serve. They have customers today and they have to run the business today, not next year.

是的，没错。只要记住，Hopper 和Blackwell是用于人们的操作的，人们今天需要进行操作。对于 Hopper 的需求非常大。我们大多数客户现在已经知道Blackwell有一段时间了，只是让你知道。好的，所以他们知道Blackwell。他们知道时间表。他们知道一段时间以来的能力。我们尽快让人们知道，这样他们就可以规划他们的数据中心，并注意到 Hopper 的需求不会改变。原因是他们有一个必须服务的操作。他们今天有客户，他们必须今天经营业务，而不是明年。

Atif Malik 阿蒂夫·马利克

Okay. 好的。

Pierre Ferragu 皮埃尔·费拉古

Pierre Ferragu, New Street Research. So, like a geeky question on Blackwell and --

皮埃尔·费拉古，新街研究。所以，就像一个关于Blackwell的怪胎问题--

Jensen Huang 黄仁勋

Thank you. 谢谢。

Pierre Ferragu 皮埃尔·费拉古

The two dyes and the 10 terabytes between the two dyes, can you tell us about how you achieve that? How much work you've put over the years into being able to achieve that technically like from a manufacturing standpoint? And then how you see the future in your roadmap, looking further away? Do you think we're going to see more and more dyes getting together into a single package? So that's one side of my question, which is more like on the chip and the architecture. And the other side is, you must be seeing like all these models that are like Sam Altman said, behind the veil of ignorance. And so can you tell us about what you see and how you see the next generation of models influencing your architecture? And so what's the direction of travel for GPU architecture for data center AI?

两种染料和两种染料之间的 10 TB，您能告诉我们您是如何实现的吗？您在多年来为了在技术上实现这一点而付出了多少努力，比如从制造角度来看？然后您如何看待未来在您的路线图中，看得更远？您认为我们会看到越来越多的染料聚集到一个单一的包装中吗？这是我问题的一面，更像是关于芯片和架构。另一方面是，您一定会看到像 Sam Altman 所说的那样，所有这些模型都隐藏在无知的面纱后面。那么您能告诉我们您看到了什么，以及您如何看待下一代模型如何影响您的架构吗？那么对于数据中心 AI 的 GPU 架构的发展方向是什么？

Jensen Huang 黄仁勋

Yeah, I'll start with the second. This is one of the great things about being the platform where all AI research is done. And so we get the benefit of seeing everything that's coming down the pike. And, of course, all next generation models are intended to push the limits of current generation systems to its limit. And so large context windows, for example, insanely large context windows, state space vectors, synthetic data generation, essentially models talking to themselves, reinforcement learning, essentially AlphaGo of large language models, Tree Search. These models are going to have to learn how to reason and do multipath planning. And so instead of one shot, it's a little bit like us thinking we have to work through our plan. And that planning system, that reasoning system, multistep reasoning systems could be quite abstract and the path could be quite long, just like playing go. And so -- but the constraints are much, much more difficult to describe. And so this whole area of research is super, super exciting. The type of systems that we're going to see in the next several years, a couple of two, three years, is unimaginable compared to today for the reasons I described. There are some concern about the amount of Internet data that's available for training these models, but that's just not true. 10 trillion tokens is great, but don't forget, synthetic data generation, models talking to each other, reinforcement learning, the amount of data you're going to be generating, it's going to take two computers to train each other. Today we have one computer training on data. Tomorrow it's going to be two computers, just -- right? Don't forget. Remember, AlphaGo. It's multiple systems competing against -- playing against each other, okay, so that we could do that as quickly as possible. So some really exciting ground-breaking work around the corner. All right. The one thing that we're certain is that the scale of these -- the scale of our GPUs, they want to be even bigger. The SerDes of our company is world class. NVIDIA's SerDes are absolutely the world's best. The data rate and the energy consumed, the data rate, the picojoule per bit in our company is unbelievably good. It is the reason why we were able to do NVLink. Remember, NVLink was because we could not make a chip big enough and so we connected eight of them together. This is in 2016. We're on NVLink Gen 5. The rest of the world doesn't even have NVLink Gen 1 yet. NVLink Gen 5 allows us to connect 576 chips together. They are together as far as I'm concerned. The data center is so big, does it have to be this close together? No, not at all. And so it's okay to split them up 576 ways. And the SerDes are so low energy anyways. Now we could make even closer chips. Now, the reason why we want that is because then the software cannot tell the difference. When you break up chips, the algorithm should be build the largest chip that lithography can make and then put multiple of them together. However, whatever technology is available to do so. But you start by building the largest chip ever. Otherwise, why didn't we do multichip back in the old days? We just kept pushing, right, monolithic as far as. And the reason for that is because the data rate on chip and the energy on chip allows for the programming model to be as uniform as possible. You don't have these things called, speaking of geeking out NUMA, non-uniform memory access, right? So you don't have NUMA behavior. You don't have weird cache behavior. You don't have memory locality behavior, which causes the programs to work differently depending on the nodes that the systems they run on. We want our software to run exactly the same wherever they are. And so you start with the biggest chip possible. That's the first Blackwell dye. We connect the two of them together. The technology 10 terabytes per second is insane. Nobody's ever seen 10 terabytes per second link before. That's 10 terabytes per second and it obviously consumes very little power, otherwise it would be nothing but that link. And so we -- you had to solve that number one. The second thing you had to solve was the question that came up before was CoWoS. It's the largest CoWoS in the world, because the first generation CoWoS was already the largest CoWoS in the world. Now the second generation is even larger. The benefit that we have is we're not surprised this time. The volume ramp demand happened fairly sharply last time, but this time we've had plenty of visibility. And so Colette is absolutely right. We've worked with the supply chain, worked with TSMC very closely. We are geared up for an exciting ramp.

是的，我会从第二个开始。这是作为所有 AI 研究都在进行的平台的伟大之处之一。因此，我们有幸看到即将到来的一切。当然，所有下一代模型的目标是将当前一代系统的极限推到极限。例如，大上下文窗口，疯狂大的上下文窗口，状态空间向量，合成数据生成，基本上是模型自言自语，强化学习，基本上是大型语言模型的 AlphaGo，树搜索。这些模型将不得不学会如何推理和进行多路径规划。因此，与其一次性，这有点像我们认为我们必须通过我们的计划。那个规划系统，那个推理系统，多步推理系统可能相当抽象，路径可能相当长，就像下围棋一样。因此——但是约束条件要难得多，难得多。因此，这整个研究领域非常、非常令人兴奋。我们将在接下来的几年中看到的系统类型，两三年，与我描述的原因相比，是无法想象的。 有人担心可用于训练这些模型的互联网数据量，但这并不是真的。100 万亿令牌很棒，但不要忘记，合成数据生成，模型之间的对话，强化学习，您将要生成的数据量，将需要两台计算机相互训练。今天我们有一台计算机在训练数据。明天将会是两台计算机，对吧？不要忘记。记住，AlphaGo。这是多个系统相互竞争，相互对抗，好吗，这样我们可以尽快实现。所以一些真正令人兴奋的突破性工作即将到来。好的。我们确定的一件事是，这些 GPU 的规模，它们想要更大。我们公司的 SerDes 是世界一流的。英伟达的 SerDes 绝对是世界上最好的。数据速率和能耗，我们公司的每比特皮焦耳是令人难以置信的好。这就是为什么我们能够做 NVLink。记住，NVLink 是因为我们无法制造足够大的芯片，所以我们将它们连接在一起。这是在 2016 年。 我们正在使用 NVLink Gen 5。全世界甚至还没有 NVLink Gen 1。NVLink Gen 5 允许我们将 576 个芯片连接在一起。就我而言，它们是在一起的。数据中心如此之大，是否必须这么靠近？不，一点也不。所以将它们分成 576 部分是可以的。而 SerDes 的能耗非常低。现在我们甚至可以制造更接近的芯片。我们想要这样做的原因是软件无法区分。当你分解芯片时，算法应该构建出最大的芯片，然后将多个芯片组合在一起。无论采用何种技术。但首先要建造有史以来最大的芯片。否则，为什么我们不在过去做多芯片呢？我们一直在不断推进，对吧，就像单片一样。这样做的原因是因为芯片上的数据速率和能量允许编程模型尽可能统一。你不会遇到所谓的 NUMA，即非一致性内存访问，对吧？因此，你不会遇到 NUMA 行为。你也不会遇到奇怪的缓存行为。 您没有内存局部性行为，这会导致程序在运行的系统节点不同时表现不同。我们希望我们的软件在任何地方都能完全相同地运行。因此，您从可能的最大芯片开始。那是第Blackwell个染料。我们将它们连接在一起。每秒 10TB 的技术是疯狂的。以前从未见过每秒 10TB 的链接。那是每秒 10TB，显然消耗非常少的功率，否则它将只是那个链接。因此，我们——您必须解决第一个问题。您必须解决的第二个问题是之前提出的 CoWoS 问题。这是世界上最大的 CoWoS，因为第一代 CoWoS 已经是世界上最大的 CoWoS。现在第二代更大。我们的好处是这次我们不会感到惊讶。上次需求量迅速增长，但这次我们有足够的可见性。因此，Colette 绝对是正确的。我们与供应链合作，与 TSMC 密切合作。我们已经为激动人心的增长做好了准备。

Aaron Rakers 亚伦·雷克斯

This will be the last question then.

这将是最后一个问题了。

Jensen Huang 黄仁勋

Bummer. Come on. 糟糕。加油。

Aaron Rakers 亚伦·雷克斯

Wow, thank you. Aaron Rakers at Wells Fargo. I really appreciate all this detail. I'm actually going to dovetail off this last comment, because today you started the conversation by talking a little bit about Ethernet and how Ethernet with Ultra.

哇，谢谢。威尔斯法戈的亚伦·雷克斯。我真的很感激这些细节。实际上，我要接着上一个评论，因为今天你开始谈论以太网和以太网与超级的一点。

Jensen Huang 黄仁勋

I love Ethernet. 我爱以太网。

Aaron Rakers 亚伦·雷克斯

Yeah. So I want to understand a little bit, NVLink, 576 GPUs now interconnected together. This idea of the fabric architecture, where does that play relative to the evolution of Ethernet, your Spectrum 4 product, this move to 800 gig? I'm just trying to understand the interplay between those and whether or not you see NVLink competing with Ethernet in those environments.

是的。所以我想了解一点，NVLink，现在 576 个 GPU 相互连接在一起。这种面向织物的架构的概念，在以太网的演进中扮演着什么角色，您的 Spectrum 4 产品，这种转向 800 吉比特的举措？我只是想了解它们之间的相互作用，以及您是否认为在这些环境中，NVLink 会与以太网竞争。

Jensen Huang 黄仁勋

No. First, the algorithm is actually very simple. First, build the largest dye you possibly can. So big that if you added one more transistor, it would literally fall on the ground. That's algorithm number one. And look at the chips that we build. They're literally the largest. They're radicle limits. Number two. If possible, connect the two of them -- connect two of them together. You're not going to connect four of them together. That's not going to happen. But if you can connect two of them together and that's the Blackwell invention. We now know how to build dice that big. But beyond that, you're going to have all kinds of weird NUMA effects and locality effects. You might as well go to NVLink. And so once you get to NVLink, the question is -- and of course, we're in Gen 5. If you don't have NVLink, then you're kind of stuck. Okay, you can't build systems like this. But if you have NVLink, then the next part is build NVLink as large as you can, modulated by power and cost. And that's the reason why NVLink is direct connect. It's direct drive, not because optical transceivers are out of fashion. Optical? Are you kidding me? We love optical. We need optical. We're going to use tons of optical. But you should build the NVLink as large as you can using copper, because you could save a lot of power, you could save a lot of money. You can make it scalable, sufficiently scalable. Now, you've got one giant chip, 576 GPU chip effectively. But that's only 576 GPU chips. That's not enough. And so we're going to have to connect multiple of them. The next click after that, the best thing you have is InfiniBand. The second best you have is Ethernet with an augmented computing layer on top of it we call Spectrum X, so that we can control the traffic that's in the system, so that we don't have these long tails. Remember, as I said, the last one to finish determines the speed of the computer. This is not an average throughput. This is not like all of us individually are accessing hyperscale and our average throughput is good enough. This is literally the last person who finishes that partial product, who finishes that tensor. Everybody else is waiting on them. I don't know who it is in this room. That's going to be the last, but we're going to hope that that person doesn't hold up right. And so we're going to make sure that that last one is -- we push everything to the middle. We only want one answer. It all shows up at the right time. Okay. And so that's the second best. And then you scale that out as much as you can and that's going to need optics and so on and so forth. There's a place for all of it. I think if anybody's concerned about optics, don't be concerned. We're -- I think the demand for optics is very, very high. Demand for repeaters is very, very high. We didn't change anything about that. All we did was we made computers larger, we made GPUs larger. Can we take one more question? This is so much fun.

首先，算法实际上非常简单。首先，尽可能构建最大的骰子。要大到如果再增加一个晶体管，它就会直接掉在地上。这就是算法一。再看看我们构建的芯片。它们确实是最大的。它们是根本限制。第二。如果可能的话，将其中两个连接在一起。你不会将四个连接在一起。那是不可能的。但如果你能将两个连接在一起，那就是Blackwell的发明。我们现在知道如何构建那么大的骰子。但除此之外，你将会遇到各种奇怪的 NUMA 效应和局部效应。你可能会转向 NVLink。所以一旦你使用 NVLink，问题就是——当然，我们现在是第五代。如果你没有 NVLink，那么你就会陷入困境。好吧，你无法构建这样的系统。但如果你有 NVLink，那么接下来的部分就是尽可能大地构建 NVLink，受功耗和成本的调节。这就是为什么 NVLink 是直接连接的原因。它是直接驱动，不是因为光学收发器过时了。光学？你在开玩笑吗？我们喜欢光学。我们需要光学。 我们将使用大量的光学。但是你应该尽可能使用铜建立 NVLink，因为你可以节省大量电力，你可以节省大量金钱。你可以使其可扩展，足够可扩展。现在，你有一个巨大的芯片，有效地有 576 个 GPU 芯片。但那只是 576 个 GPU 芯片。这还不够。因此，我们将不得不连接多个。在那之后，你拥有的最好的选择是 InfiniBand。第二好的选择是 Ethernet，顶部有一个增强计算层，我们称之为 Spectrum X，这样我们就可以控制系统中的流量，这样我们就不会有这些长尾巴。记住，正如我所说的，最后一个完成的人决定了计算机的速度。这不是平均吞吐量。这不像我们每个人都在访问超大规模，我们的平均吞吐量就足够好。这实际上是最后一个完成部分产品的人，完成那个张量的人。其他人都在等待他们。我不知道在这个房间里是谁。那将是最后一个，但我们希望那个人不要拖延。 所以我们要确保最后一个是这样的 - 我们把所有东西都推到中间。我们只想要一个答案。所有东西都会在正确的时间出现。好的。这就是第二好的。然后你尽可能扩展它，这将需要光学器件等等。每样东西都有它的位置。我认为如果有人担心光学器件，不要担心。我认为对光学器件的需求非常非常高。中继器的需求也非常非常高。我们没有改变任何东西。我们只是把计算机做得更大，把 GPU 做得更大。我们可以再回答一个问题吗？这太有趣了。

Will Stein 威尔·斯坦

One last question from the buy-side, Jensen. You've talked a lot about -- oh, I'm sorry.

买方最后一个问题，詹森。你谈了很多关于 - 哦，对不起。

Jensen Huang 黄仁勋

Where is he? Oh, there is. Hey, Will.

他在哪里？哦，他在那里。嘿，威尔。

Will Stein 威尔·斯坦

Yeah, hey. 是的，嘿。

Jensen Huang 黄仁勋

Hi, Will. 嗨，威尔。

Will Stein 威尔·斯坦

Sovereign AI. Is there a way to sort of understand like what you're going to do for the United Arab Emirates? That would be one question. And I guess my second question is, like, I'm going to go home. I'm going to see my 91-year-old mother. How can I try to explain to some 91 year old what accelerated computing? I guess, I've got a good answer of the first question. I'll figure out the second one. Thanks.

主权 AI。有没有一种方法可以了解你将为阿拉伯联合酋长国做些什么？这将是一个问题。我想我的第二个问题是，我要回家了。我要见我的 91 岁的母亲。我怎样才能试图向一个 91 岁的人解释什么是加速计算？我想，我已经得到了第一个问题的好答案。我会想出第二个问题的答案。谢谢。

Jensen Huang 黄仁勋

Okay. Yeah, I don't know what you were going to say on the second one, but on the second one, I would say use the right tools for the right job. And right now, general purpose computing, you're using the same tool for every single job. Literally what you have is a screwdriver and you're using it from the moment you woke up to the moment you go to bed. And so you start with you brushing your teeth with a screwdriver. It probably works. I haven't tried it, but it probably works. And so you just use that one tool the whole day. Now, of course, because you're going to use that one tool for the whole day, over time humans have gotten pretty smart. And so we made that general purpose tool. And so now the screwdriver has brushes on it, it's got hair on it. So then it becomes useful for all kinds of stuff. And you could also use it to clean the bathroom and all that kind of stuff. And so one tool. Was that the answer you were going to give? All right. So, we created basically two tools. We said that the CPU is incredibly good at sequential things and what it's not good at is parallel things. Now, the parallel things, the weird thing is this, for most applications, let's say XL, the parallel part is not very much. That's the reason why CPUs are really the best processor for XL. For your web browser, except for graphics that we came along later, most web browsers are largely single threaded. Java is largely single threaded. And so for many applications of personal computing is largely single threaded and then CPU is really quite ideal. And then all of a sudden, there's this new application that came along, computer graphics, video games, where literally 1% of the code is 99% of the runtime. Do you guys understand what I'm saying? 1% of the code is 99% of the runtime. And the reason for that is because it's computing the pixels one at a time. So 1% of the code is 99% of the runtime. And we said, look at that. How interesting. Why don't we take, go create something that's insanely good at 1% of the runtime, meaning it's bad at 99% of the runtime. Excuse me, bad at 99% of the code. It's good at 1% of the code. And we go create applications or find applications where that 1% of the code is 99% of the runtime. Molecular dynamics, medical imaging, seismic processing, artificial intelligence, makes sense. That's why accelerated computing, data processing, so on and so forth where 1% of the code is 99% of the runtime. And that's the reason why we get such great speed up. All right.

好的。是的，我不知道你在第二个问题上要说什么，但是在第二个问题上，我会说要用对的工具做对的工作。现在，通用计算，你在每一个工作中都在使用同样的工具。从你醒来到睡觉的时刻，你手里拿的就是一个螺丝刀。所以你从刷牙开始就用螺丝刀。可能会有效果。我没试过，但可能有效。所以你整天只用那一个工具。当然，因为你整天只用那一个工具，随着时间的推移，人类变得相当聪明。所以我们制造了那个通用工具。现在螺丝刀上有刷子，上面有头发。所以它变得适用于各种事情。你也可以用它来清洁浴室等等。所以一个工具。这是你要给的答案吗？好的。所以，我们基本上创造了两个工具。我们说 CPU 在顺序事物上非常擅长，而在并行事物上不擅长。 现在，关于并行的事情，奇怪的是，对于大多数应用程序，比如 XL，并行部分并不多。这就是为什么 CPU 实际上是 XL 的最佳处理器的原因。对于您的网络浏览器，除了后来出现的图形之外，大多数网络浏览器基本上是单线程的。Java 基本上是单线程的。因此，对于许多个人计算应用程序来说，基本上是单线程的，CPU 实际上非常理想。突然之间，出现了这种新的应用程序，计算机图形，视频游戏，其中实际上 1%的代码占了 99%的运行时间。你们明白我在说什么吗？1%的代码占了 99%的运行时间。原因是因为它逐个像素地计算。所以 1%的代码占了 99%的运行时间。我们说，看看那个。多有趣啊。为什么不创造出一些在 1%的运行时间上非常出色的东西，也就是说，在 99%的代码上表现不佳。对 99%的代码表现不佳，对 1%的代码表现出色。我们创造应用程序或找到应用程序，其中 1%的代码占了 99%的运行时间。 分子动力学，医学成像，地震处理，人工智能，都有意义。这就是为什么加速计算，数据处理等等，其中 1%的代码占据了 99%的运行时间。这就是我们获得如此快速提升的原因。好的。

Colette Kress 柯莱特·克雷斯

Sovereign AI 主权人工智能

Jensen Huang 黄仁勋

Sovereign AI. Every country has their own natural resource and that natural resource is called their intelligence. It's in their language. India has their own language. They have many of them, lots of different dialects. They have their own language, their sensibility, their culture, their history. It belongs to them. And a lot of it is in their national archives and is digitized. It's not actually on the Internet. It belongs to them. They ought to take that and go create their own sovereign AI. And they believe the same. Sweden is the same way. Japan is going to do the same. You name it. Companies -- countries all over the world realize that this is their natural resource and they shouldn't let it just be used by anybody to then import their natural resource back to them in an automated way by paying somebody else. Don't let their data go out for free and import AI. They now realize it ought to be the other way around that they should keep their own data and then export AI. And so export the AI of Korea, export the AI of Malaysia, export the AI of, you name it, Middle East countries. So, we have export control limitations on our products. And in most of the areas, the answer is it's not export control. And if there's any export control, we can still work with the US government and make sure that the export is going to be fine. But we, number one, just make sure that we're compliant with export control and in some countries we have to offer degraded products or -- I didn't say that right or lower specification products. And -- but anyways, number one, just be compliant with export controls and help countries around the world to be able to do this. It's a very big market. Yeah, it's a very big market. There are going to be AIs that are going to be trained and continuously refined for just about every culture in the world. 

主权人工智能。每个国家都有自己的自然资源，这种自然资源被称为他们的智慧。它存在于他们的语言中。印度有自己的语言。他们有很多，许多不同的方言。他们有自己的语言、感性、文化、历史。这属于他们。其中很多内容存储在他们的国家档案中并已数字化。实际上并不在互联网上。这属于他们。他们应该利用这些内容，创造自己的主权人工智能。他们也这样认为。瑞典也是如此。日本也将采取同样的做法。你可以说其他国家也会这样。世界各地的公司和国家都意识到这是他们的自然资源，他们不应该让它被任何人使用，然后以自动化的方式通过支付他人将其自然资源重新引入。不要让他们的数据免费流出并引入人工智能。他们现在意识到应该反过来，他们应该保留自己的数据，然后出口人工智能。因此，出口韩国的人工智能，出口马来西亚的人工智能，出口中东国家的人工智能，你可以说其他国家。因此，我们对产品实施出口管制限制。 在大多数地区，答案是不受出口管制。如果有任何出口管制，我们仍然可以与美国政府合作，确保出口顺利进行。但我们首先要确保遵守出口管制，在一些国家我们不得不提供降级产品或者更低规格的产品。但无论如何，首要任务是遵守出口管制，并帮助世界各国能够做到这一点。这是一个非常大的市场。是的，这是一个非常大的市场。将会有人工智能被训练和不断完善，以适应世界各地几乎每种文化。

Jensen Huang 黄仁勋

Thank you. Do you guys? No, no. Thank you very, very much. I appreciate -- Colette and I appreciate all of your support and interest in the company. And this is really quite an extraordinary time. It's not usual that we get to live through a time like this where the single most important instrument of society is being reinvented after 60 years, that a new way of doing software has emerged. And you know that software is one of the most important technologies that humanity has ever created and that you're in the beginning of a new industrial revolution. And so the next 10 years, you definitely don't want to miss. All right. Thank you very much.

谢谢。你们呢？不，不。非常非常感谢。我感激——Colette 和我感激你们对公司的支持和关注。这真的是一个非常特殊的时刻。我们很少有机会经历像这样的时刻，社会最重要的工具在 60 年后被重新发明，出现了一种新的软件开发方式。你们知道，软件是人类创造的最重要的技术之一，你们正处在新工业革命的开端。所以接下来的 10 年，你们绝对不想错过。好的。非常感谢。