---
title: "黄仁勋 × BG2：AI 收入撑得起资本开支吗？2027 年瓶颈、奇点与开源闭源"
title_original: "Jensen Huang | Does AI Revenue Justify Capex, '27 AI Bottlenecks, Singularity, Open vs Closed & more"
guest: "Jensen Huang（黄仁勋）"
guest_role: "NVIDIA 创始人兼 CEO"
host: ["Brad Gerstner", "Clark Tang"]
show: "BG2 Pod"
date: 2026-10-15
duration: "1:39:53"
language: en
source: "https://www.youtube.com/watch?v=eS7qQ3WJSrw"
source_platform: "YouTube"
page: "https://ai-daily-cn.pages.dev/watch/bg2-jensen-huang/"
tags: ["研几", "研几AI", "访谈"]
---

# 黄仁勋 × BG2：AI 收入撑得起资本开支吗？2027 年瓶颈、奇点与开源闭源

嘉宾 **Jensen Huang（黄仁勋）** · NVIDIA 创始人兼 CEO · 主持 Brad Gerstner、Clark Tang · BG2 Pod · 2026-10-15 · 时长 1:39:53 · [YouTube 原片](https://www.youtube.com/watch?v=eS7qQ3WJSrw) · [研几 · AI 访谈页](https://ai-daily-cn.pages.dev/watch/bg2-jensen-huang/)

## 核心总结

黄仁勋在与 BG2 播客的对谈中判断，AI 算力需求远未见顶，推理与智能体将带来 10 亿倍算力增长，而 2027 年前后土地、电力、厂房等物理建设将成为最大瓶颈；同时他反驳了“AI 收入撑不起资本开支”的担忧，认为不必把数万亿美元基建回报押在两家前沿实验室身上。

- **推理与智能体推高算力需求**：黄仁勋重申，思考、规划、反思和工具使用比单纯复述知识消耗更多算力，而模型变强后会被雇去做更多工作，两者叠加；他把过去一年 harness（智能体外壳）、记忆与工具使用视为关键突破，编程只是“最容易的第一个入口”，下一步将扩展到设计、创意营销和科学领域。

- **Navier-Stokes 与智能成本骤降**：他提到 OpenAI 用约 1 万个智能体协同解决 Navier-Stokes，算力成本约 400 万美元，并预计未来几年同类成本再降 5–10 倍；训练 GPT-1 级别模型的成本已从约 1000 万–2000 万美元降到约 100 美元，智能成本将像摩尔定律一样快速下降，而需求会远远补上。

- **2027 年瓶颈在土地、电力、厂房**：面对明年上线约 43 吉瓦新 AI 算力的预测，黄仁勋认为需求更大，但土地许可、发电、建筑工人和供应链（如基板、硅光）是主要障碍；他提到市场讨论约 50 吉瓦已宣布芯片对应可能只有 35–40 吉瓦的土地、电力和外壳能上线，并称英伟达每吉瓦对应的收入已从约 180 亿美元升到 400 多亿美元。

- **AI 收入不必只靠两家实验室**：他把英伟达业务分为三块——经典计算与数据处理（约三分之一，年增 30%–40%）、前沿实验室（约四分之一到三成）、以及 NeoCloud、主权云、企业和数千家初创公司（约三分之一）；并指出 80% 的初创公司使用开源模型，开源模型在垂直领域解锁新机会，而非与前沿实验室正面竞争。

- **资本开支回报与“泡沫”担忧**：他承认今年前沿实验室合计收入约 2000 亿美元、需明年翻倍到 4000 亿、后年到 8000 亿才能接近数万亿美元资本开支的回报，但强调不应把全部基建压力外推到 OpenAI 和 Anthropic 身上；摩尔定律终结意味着想要两倍算力就要付两倍钱，到本十年末全球 AI 基建可能达每年 3–4 万亿美元。

- **安全与监管框架已具操作性**：黄仁勋把美国超级智能框架描述为“三条腿的凳子”——商务部对前沿模型审查、白宫超级智能协议下的内外审计与董事会监督、以及 SI Force 的行业协调；他用“安全气囊与刹车也是汽车技术”的类比反驳能力与安全对立的说法，并称对齐永远无法完美，因此要靠零信任、最小权限、带外监控和关停能力来遏制自主系统。

## 金句

- 「我们需要的算力将是不久前的十亿倍」 —— Jensen Huang · [8:28](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=508s)
  > the amount of compute that we're going to need is a billion times more than where we were just recently
- 「这就是为什么我们说伟大的战略家喜欢战略选择权。」 —— Jensen Huang · [9:53](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=593s)
  > That's the reason why we say great strategists love strategic optionality.
- 「如果不安全、不可靠，功能就永远没有机会展现。」 —— Jensen Huang · [39:49](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2389s)
  > if it's not safe, it's not secure, the functionality will never get a chance to speak.
- 「人们看到几个实验室，但我看到更多，我看到五六个实验室」 —— Jensen Huang · [1:04:06](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3846s)
  > people see a couple of labs, but I see more than that I see five six labs
- 「它们会取得巨大的成功。但它们不必接近一万亿美元才算成功。」 —— Jensen Huang · [1:12:25](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4345s)
  > They're going to be magnificently successful. But they don't have to be anywhere near a trillion dollars to be successful.

## 精华片段

1. [8:21–10:20](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=501s) 为什么算力需求要再涨十亿倍
2. [39:38–40:57](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2378s) 安全与能力是一回事
3. [1:02:17–1:04:40](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3737s) 英伟达的三块生意：没有 AI 实验室也不是零
4. [1:12:13–1:12:42](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4333s) 别把压力都压在两家实验室身上

## 文字稿（中英对照）

### [0:00](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=0s) 开场
*Intro*

[0:00](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=0s) 这对我们来说是一个生存性问题。我们完全可以停下来，但如果我们停下来，今天我们就输了。功能性、能力、安全性、保障性，它们都处在完全相同的位置。头号优先事项。这是我们第三次一起上播客。在过去10天里，你出席了白宫为习主席举办的国宴。你参加了白宫峰会，我们在那里达成了关于超级智能的协议。就在昨天，你和你的父母一起在那里，总统向你颁发了国家科学奖章。你是桌上唯一的投资者。

> This is an existential issue for us. We could easily stop, but if we stop, today we lose. Functionality, capability, safety, security, they're all at exactly the same place. Priority number one. It's our third pod back together. In the last 10 days, you've been at the Xi State Dinner at the White House. You were at the White House Summit where we did the accord on super intelligence. Just yesterday, you were there with your parents. when the president presented you with the National Medal of Science. You were the only investor at the table.

### [0:20](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=20s) 在白宫的一周与National Medal of Science
*A Week at the White House & the National Medal of Science*

[0:42](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=42s) 你是桌上唯一的投资者，现在我们回想起来。是的。嗯，告诉我，昨天和你的父母一起在那里，让我震惊的是，台上六个人中有五个都出生在美国以外。那是什么感觉？嗯，首先，这本身就是极大的荣誉。对。这也是对公司所做贡献的体现。

> You were the only investor at the table, now that we reflected on it. Yeah. Well, tell me, being there with your parents yesterday, it struck me five out of the six people on stage were all born outside of the United States. How did that feel? Well, first of all, it's a great honor all by itself. Right. And it's a reflection of the contribution that the company has made.

[1:10](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=70s) 显然，我们发明了一种新的计算模型。这可能是当今世界上应用最广泛的计算模型。它是唯一一种存在于每个云中的计算架构。它存在于每家计算机公司。它被用于每个科学领域。它被游戏玩家使用。它被创作者和构建者使用。它被科学家使用。它为智能提供动力。我的意思是，如果你退一步看公司所取得的成就，我们最初是为游戏玩家制造图形芯片，当然我仍然为此感到自豪，我们现在已经出货了15亿、16.7亿个单位。那是多少代……我的意思是，我培养了代代科学家、孩子和设计师。所以我认为影响是，我们在这里，我们是支撑社会智能的基础设施。因此，代表公司接受这个奖项对我来说是一个极其自豪的时刻。

> And obviously, we invented a new computing model. This is probably the most pervasive computing model in the world today. It's the only computing architecture that's in every cloud. It's in every computer company. It's used in every field of science. It's used by gamers. It's used by creators and builders. And it's used by scientists. It powers intelligence. I mean, if you take a step back and you look at what the company has accomplished, we started out building a graphics chip for gamers, and I'm still proud of that, of course, and we now shipped 1.5, 1.67 billion units. That's how many generations of... I mean, I raised generations of scientists and kids and designers. And so I think the impact, and here we are, we're the infrastructure that powers the intelligence of society. And so it's an incredibly proud moment for me to receive the award on behalf of the company.

[2:09](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=129s) 现在对我的父母来说，我想让他们去，我给了他们真的24小时的时间，他们不怎么旅行。他们不怎么旅行。我爸爸甚至没有一套西装。哇。是的。所以他跑出去买了一套西装。他看起来很棒。他们看起来很棒。那是一套全新的西装，他出现了。我不知道他有一套西装。我以为他只会穿件衬衫。他出去买了一套西装。我妈妈买了一条新裙子，你知道，真的在24小时内。是的。哦，我知道这是怎么回事。哦我的天。太不可思议了。所以我们让他们上了……我们为他们包了一架飞机，让他们到那里，他们见到总统非常高兴。

> Now for my parents, I wanted them to go and I gave them literally 24 hours at a time and they don't travel that much. They don't travel that much. And my dad didn't even have a suit. Wow. Yeah. So he ran out and bought a suit. He looked great. They looked great. It was a brand new suit and he showed up. I didn't realize he had a suit. I thought he was just going to wear some shirt. And he went out and bought a suit. My mom got a new dress, you know, literally in the 24 hours. Yeah. Oh, I know how it works. Oh my gosh. It's incredible. And so we got them on a... We chartered a plane for them, got them there, and they were so happy to meet the president.

[2:55](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=175s) 特朗普总统扶着我妈妈，因为她走路困难，他们俩一起走下舞台。他扶着她，陪她走下舞台。他们见到她时无比、无比自豪。但对他们来说，这是他们的美国梦。你知道，我爸爸，我还记得，我想我当时大概五岁左右。他刚从美国出差回来。他在一家空调公司工作。你们当时住在台湾？我们当时，我们刚搬到泰国。泰国。

> And President Trump walked my mom because she has a hard time walking, and the two of them were walking off stage. And he held her and walked her down the stage. They were just incredibly, incredibly proud to meet her. But for them, it's their American dream. You know, my dad, I still remember, I think I was like five years old or something like that. And he had just come back from the United States on a business trip. He was working for an air conditioning company. And you were living in Taiwan? We were at the time, we had just moved to Thailand. Thailand.

[3:30](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=210s) 哦，不，不，我们正要搬到泰国。我们当时在台湾。他回来了。他在美国为一家叫开利的公司工作，他们的空调公司。他来这里是培训。在世界所有地方中，他们把他送到了纽约。所以你可以想象一个20多岁，我想是20多岁中后期，从未离开过台湾的孩子，去了美国，偏偏是纽约市。他看到了未来的未来。他回来时惊呆了。

> Oh, no, no, we were just about to move to Thailand. And we were in Taiwan. He had come back. And he was in the United States working for a company called Carrier, their air conditioning company. And he was here for training. And of all the places in the world, they landed him in New York. And so you could just imagine a kid in the late 20s, I guess mid-late 20s, and never been outside of Taiwan, goes to United States, New York City of all places. And he saw the future of the future. And he came back mind blown.

[4:06](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=246s) 他在机场买的所有东西，都带回来了。他带了巧克力。他带了，你知道过去那种管装奶酪吗？有点像白奶酪。奇怪的奶酪。它装在管子里。太棒了。他带回了各种奇怪的东西。他说，这是……我们有一天应该去那里。他们一有机会，就把我和我哥哥送到了美国，这是他的梦想。我认为没有多少人意识到这一点。你比你的父母先来到美国。

> All the things that he bought at the airport, brought it back. He brought chocolate. He brought, you know that cheese that you have in a tube back in the old days? It's kind of like white cheese. Strange cheese. It comes in a tube. That's great. And he brought back all kinds of strange things. And he says, this is... We should go there someday. And the first opportunity they had, they sent my older brother and I to United States and it was his dream. I don't think many people realize that. You came to the United States ahead of your parents.

[4:43](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=283s) 他们几年后也来了，但这让我恍然大悟，昨天台上六个人中有五个……世界上其他地方没有这样的。原因是我們吸引世界上最优秀的人。我们是他们想要来建立生活的地方，因为有机会提供她所提供的自由、它所提供的机会、发明和创业精神以及敢做敢为的态度。在美国，一切皆有可能。我们从“当然可能”开始。我们当然能做到。而不是，你知道，我经常旅行，在其他地方，当然是“你做不到”。我们把这视为理所当然。但这里我们创造的文化所带来的精神、能量，怎么强调都不为过。所以我认为，知识产权流入这个国家的潮流，是世界前所未见的。

> They joined a few years later, but it just hit me like a ton of bricks that five out. Out of the six people on that stage, yesterday, were unlike anything else in the world. And the reason for that is we attract the world's best. We are the place where they want to come and build their lives because of the opportunity to provide the freedom that she offers, the opportunity that it provides, the spirit of invention and entrepreneurialism and the can-do attitude. In the United States, everything is possible. We start with, of course it's possible. Of course we could do that. Instead of, you know, I do a lot of travel, and of course you can't do that. Of course, you know, in other places. And we take it for granted. But the spirit, the energy that comes around with the culture that we've created here, it just can't be understated. And so I think the flow, the inflow of intellectual property into this country is unlike anything else the world's ever seen.

### [5:06](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=306s) 美国的超级力量
*America’s Superpower*

[6:12](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=372s) 我们想要这样。我们希望世界上所有伟大的头脑都来这里，带来他们伟大的想法，在这里实现它们，在这里实现它们，让我们保持这个飞轮运转。这是一个微妙的系统。我明白，这是一个微妙的平衡。我明白，这是一个微妙的平衡。所以我们必须合法地做。我们必须正当地做。但我们确实想这样做。是的，当然。所以我想从高层开始，先谈谈我们现在的位置，然后我想深入一点，谈谈安全和监管以及如何做好这件事，我们在算力建设方面的进展，最后我们将深入一点谈谈英伟达的增长以及那里发生了什么。但你一直很擅长预测未来，对吧？两年前，我们坐在这里，你说市场大幅低估了未来五年所需的算力建设规模。你说不会是1万亿，至少会是2万亿的累计算力建设，当然现在我们已经超过了。在所有人都在考虑之前，你说推理，对，算力将增长10亿倍，因为推理时的推理和思考。

> And we want that. We want all of the world's great minds to come here, bring their great ideas, make them, realize them here, realize them here and let us keep that flywheel going. It's a delicate system. It's a delicate balance I get. It's a delicate balance I get. And so we have to do it lawfully. We have to do it rightfully. But we surely want to do it. Yeah, for sure. So I want to start high level just kind of where we are then I want to dig down a little bit Into the safety and regulation and getting this right Where we are in terms of the compute build out and then finally we're gonna finish up Think you know drilling in a little bit to Nvidia's growth and and and and what's going on there But you've been really good at calling the future, right? Two years ago We were sitting here and you said that the market was dramatically under-est the amount of compute build-out that was going to be required and was going to happen over the course of the next five years. You said it's not going to be a trillion, it's going to be two trillion at a minimum in terms of cumulative compute build-out, and of course now we're over that. You said before everybody was thinking about it that inference, right, compute was going to 1 billion X because of inference time reasoning and thinking.

### [6:37](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=397s) 进入奇点：我们现在在哪里？
*Entering the Singularity: Where Are We Now?*

[7:23](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=443s) 然后去年，当然，在我们看到Opus 4.6之前，在我们看到Claude Code之前，你说我们正在进入智能体时代，我们将释放很多这些能力。当然，我们看到了Opus 4.6和Claude Code真正起飞。我们现在在哪里？我们有Elon和Sam还有——我说过的话只有三件成真了。还有另外200件。你为什么不继续呢？让那成为播客吧。所以几周后的投资者日是关于进入奇点，这种加速。我们有Elon、Sam和Demis都说，嘿，我们处在山麓或正在进入奇点。我们有，你知道，模型解决纳维-斯托克斯方程，如你所知。所以当你看着你的水晶球时，我们在哪里？我们是在拐点吗？我们是在加速进入超级智能的不同阶段吗？

> And then last year, of course, before we saw Opus 4.6, before we saw Claude Code, You said we're entering the age of agents where we're going to unlock a lot of these capabilities. Of course, we saw Opus 4.6 and Claude Code really took off. Where are we now? We have Elon and Sam and- There's only three things that I said that came true. There's another 200. Why don't you just keep on going? Let that be the podcast. So our investor day in a couple of weeks is on entering the singularity, this acceleration. We have Elon and Sam and Demis all say, hey, we're in the foothills or entering the singularity. We have, you know, model solving Navier Stokes, as you know. So when you look into your crystal ball, where are we? And are we inflecting? Are we accelerating, you know, into a different phase of super intelligence?

[8:21](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=501s) 让我们来拆解一下。首先，我们需要的算力将是不久前的十亿倍，这是相当合乎逻辑的。原因是，思考和解决问题需要更多的推理，而不是重复某件事、反刍你记住的东西。所以当你使用 ChatGPT 时，你给它一个提示，大部分知识都编码在一个大型语言模型中。不幸的是，当然，它会产生幻觉。所以下一次点击是生成式的，但它也基于事实依据或受其条件限制。是的。所以它必须出去做搜索，做一个快速搜索，带着事实依据回来，然后从那个条件出发，用它已经学到的预训练知识来增强，然后它会生成新的答案，不管是什么。

> And so let's unpack that. First of all, it's fairly logical why the amount of compute that we're going to need is a billion times more than where we were just recently, … last year. And the reason for that is because thinking and problem solving takes a lot more, you inferences than repeating something, regurgitating something that you memorized. So when you're a ChatGPT,, you give it a prompt, most of that knowledge was encoded into a large language model. Unfortunately, of course, it hallucinated. And so the next click was a generative, but it was also grounded or conditioned by ground truth. Yes. And so it would have to go out and do a search, do a quick search, comes back with the ground truth, and then from that conditioning, augmented with its own what it already learned pre-training, and it would generate new whatever the answer is.

[9:24](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=564s) 嗯，我们真正想做的很多有趣的事情，一步一步推理，想出几个选项，计划，你可能会决定，让我们看看这些事情如何展开。然后根据所有这些答案，我反思迄今为止最好的答案，迄今为止最好的路径。我可能会选择一两个或三个，展开它们，修剪其余的，然后继续这样做。人们，我们这样做。这就是为什么我们说伟大的战略家喜欢战略选择权。一方面，你不想太快修剪东西，但另一方面你也必须优化你的资源。嗯，AI正在这样做。所以它们保持战略选择权，展开，修剪，反思，修剪，等等。所有这些思考，所有这些探索都需要更多的算力。

> Well, most of the things that we want to really do that are interesting, to reason step by step, to come up with several options, plan, and you might decide, let's see how these things all roll out. And then based on the answers of all of those, I reflect on the best answer so far, the best path so far. I might choose one or two or three of them, roll them out, prune the rest of it, and then just keep on doing this. And people, we do this. That's the reason why we say great strategists love strategic optionality. You don't want to prune things too fast, on the one hand, but you also have to optimize your resources. Well, AIs are doing that. So they're keeping their strategic optionality, they're rolling things out, they're pruning, they're reflecting, they're pruning, so on and so forth. All of that thinking, all of that exploration just takes a lot more compute.

[10:21](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=621s) 而且不仅如此，因为它现在能做一些有产出的事情，我们就会把它用在更多事情上。所以这两者会叠加：它做自己的工作时会消耗更多算力，但它做得很好。因此，我们会雇它做更多工作。没错。所以这一切都会叠加。到这个时候，我想说，去年最大的观察是：一方面，我们希望模型越来越好。我们仍然希望模型越来越好。模型的缩放定律会继续下去。然而，给大语言模型配上一个harness（智能体外壳），基本上就是它的外骨骼，你知道，就像大脑的身体，让它能够管理工作记忆、管理知识。你怎么从工作记忆到短期记忆、到长期记忆，再到基本上构建你的知识图谱，这正是我们所有人都在做的事。我们会想，嘿，你知道，我学到了什么，它如何融入我过去所知的一切？还有你知道我们怎么睡觉，我们做压缩，那些说不通的事情，或者不重要的，我们最终会忘掉，但那些重要的事情，不知怎么地会被重新连接起来。所以我们在离线做这种事。于是我们有了工作记忆，我们有离线知识提取和创造、知识图谱构建。

> And not on top of that, because it's able to now do productive things, we're going to use it for more things. And so the compounding of both, it uses more compute to do its job, but it's doing a good job. And therefore, we're going to hire it to do more jobs. Exactly. And so that all compounds. At this point, I would say that the big observation from last year is that on the one hand, we want the models to get better and better. And we still want the models to get better. And the scaling laws of models are going to continue. However, the attachment of a harness, basically the exoskeleton of the large language model, You know, kind of the body of the brain that allows it to manage working memory, manage knowledge. How do you go from working memory to short term, to long term, to basically formulating your knowledge graph, which is what we all do. We kind of think about, hey, you know, what did I learn and how does it fit into everything that I know in the past? And you know how we go to bed, we do compaction, things that don't make sense, we kind of or don't matter, we kind of eventually forget, but the things that kind of matter, somehow it gets reconnected. And so we're doing that kind of offline. And so we got working memory, we have offline knowledge extraction and creation, knowledge graph creation.

[11:41](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=701s) 整个记忆这部分是件大事。另一件事当然是工具使用。我们已经教会这些工具、这些AI如何与人类有史以来最重要的工具交互。这跟机器人没什么不同，它来到我们家，更可能使用我家里所有的工具，而不是从零开始去造。所以，Spice仍然会有用，Verilog会有用，模拟器会有用，设计编译器、Photoshop、AutoCAD，还有Katia，以及所有Siemens工具。我们一直用来运营公司、建设公司、设计东西、操作东西的所有这些不同工具，都会成为智能体使用的绝佳工具。但智能体必须学会如何使用它们。所以这两个领域，知识和工作记忆，以及工具使用，在过去一年里是巨大的突破。

> That whole part of memory is a big deal. The other thing is, of course, tool use. And we've taught these tools, these AIs, how to interact with the most important tools that the humans ever made. It's no different than if you're a robot, it comes into our house, it's more likely you're going to use all the tools that I have in the house than to go off and build it from scratch. So, Spice is still going to be useful, and Verilog will be useful, and simulators will be useful, and design compilers, and Photoshop, and AutoCAD, and the So, Katia, and all the Siemens tools. All of these different tools that we've been using to run our company, you build our companies, design things, operate things. They're all going to be great tools for the agents to use. But the agents had to learn how to use it. And so these two areas, knowledge and working memory and tool use, huge unlocks in the last year.

[12:39](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=759s) 所以现在突然之间，你可以把这些大语言模型封装在它们的harness或外骨骼里，你可以把它们放进你的公司，把它们连接到信息、连接到你所拥有的事实依据，把它们连接到它们可以使用的工具，然后突然之间它们就开始做惊人的事情。坦率地说，最容易的部分是编程。这算是通往有用工作的第一个入口。在很多方面，编程容易，是因为客户和它们使用的工具都相当习惯于那些有巨大用途但并非极其易用的东西。所以在编程的早期，部署并不是极其容易，但软件工程师找到了办法。所以现在我认为我们的下一个跨越就是现在，你们可能已经在玩Astra，OpenAI的Astra，它真的很棒，我们在所有这些不同地方都在用它。所以现在这些智能体的下一阶段，它们能够做设计工作和大量创意营销工作，我们用它来做，你拍一张照片交给Astra。它真的能把它重建成CAD。

> So now all of a sudden, you could put these large language models encapsulated in their harness or exoskeleton, and you You can put them into your company, connect them to the information, the ground truth that you have, connect them to the tools that they can use, and all of a sudden they start doing amazing things. The easiest part, frankly, was coding. That's kind of the first on-ramp to useful work. In a lot of ways, coding is easy because both the customer and the tools that they use are fairly accustomed to things that has tremendous use but are not extremely easy to use. And so in the early days of coding, it wasn't extremely easy to deploy, but software engineers found a way. And so now I think the next click for us is now, and you guys have probably been playing with Astra, OpenAI's Astra, it's really fantastic and we're using it in all these different places. And so now the next phase of these agents, they're able to do design work and a lot of creative marketing work, and we're using it to go, you take a picture and give it to Astra. It literally recreates it into CAD.

[13:59](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=839s) 它把它重建成CAD，这太不可思议了，因为我们需要CAD，这样我们才能与CAD交互，做微调、做调整。所以Astra真的能给我们最初的初始实现。生产力高得惊人。你拿Astra和最新版本，把它应用到所有这些不同领域。我认为除了软件工程，我们还会看到设计师、建造者、创作者，以及整个领域，科学家和所有这些都会被解锁。当然，挑战在于，我们如何安全、可靠地把智能体部署到全球的企业中。这又完全是安全和保障的同一个问题。

> It recreates it into the CAD, which is incredible, because we need the CAD so that we can interact interact with the CAD and make fine tuning and adjust. And so Astra could really give us the first initial implementation. Incredibly productive. And you take Astra and the latest release and you apply it to all of these different domains. I think that besides software engineering, we're going to see designers and builders and creators and this whole area, scientists and all of this unlocking. Now, the challenge, of course, is how do we deploy agents safely and securely into the world's enterprise? This is exactly the same question again with safety and security.

[14:48](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=888s) 我们会谈到这个。对，我们会聊到。让我印象深刻的是，这一切都发生在今年，而OpenAI和Anthropic各自只有，我不知道，两三个吉瓦的算力。所以相对于即将到来的算力，这是相当小的量。所以，我感兴趣的是，经济通常的运作方式，或者人们建模经济、建模公司的方式，是线性减速。他们不太擅长建模加速。但这里我们有这个双指数。我们有大量新算力上线。如果OpenAI和Anthropic明年的算力是，我不知道，3到5倍，这是否顺理成章？我们解决了Navier-Stokes。我们会回看这件事说，那只是很小的算力和智能，是我们从此往后能拥有的最差的智能，却解决了一个千禧年难题。

> We'll talk about it. Yeah, we'll come in on that. What I'm struck by is that all of this occurred this year with OpenAI and Anthropic having, I don't know, two, three gigawatts of compute each. So relatively small amount of compute relative to what's coming. And so again, what I'm interested in is the way the economy generally works or people model the economy or they model a company is linear deceleration. They're not very good at modeling acceleration. But here we have this double exponential. We have massive amount of new compute coming on board. If OpenAI and Anthropic are, I don't know, 3 to 5 Xing their compute next year, does it stand to reason? We solved Navier-Stokes. We'll look back at this and say that was a tiny amount of compute and intelligence, the worst intelligence we're ever going to have from here forward, solving a millennium problem.

[15:45](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=945s) 当你看到明年即将上线的东西时，你预期我们会在长时间运行的智能体问题上取得重大突破，让我们能使用这些智能体集群吗？嗯，Navier-Stokes确实是个很好的例子。记住，OpenAI并不是凭空解决Navier-Stokes的。他们是在许多伟大科学家、许多伟大数学家和许多伟大先前工作的基础上解决的。之前我说过，我们正处于许多不同领域的临界点，因为我们在Navier-Stokes上正处于那个面、那个点，大量基础工作已经完成。如果我们能把AI的穷举式探索应用到先前已完成的工作上，我们很快就会看到一个又一个突破。但要做到这一点，我们确实需要大量算力。原因在于，大量的探索是必要的。当然。如果只是某人解决某个基本原理理论就能发现突破，如果它不需要对空间、搜索空间——不是太空空间，而是搜索空间——进行大量探索，如果它不需要这些，那早就有人找到了。对吧。你知道，人们相当聪明。所以我认为有许多这样的问题就处在悬崖边上，真的只需要那种大规模的智能探索来把我们带到那里。

> When you look at what's coming online next year, do you expect that we're going to have major breakthroughs in long running agentic problems that we can use these swarms of agents on? Well, Navier-Stokes is really a great example. Remember, OpenAI did not solve Navier-Stokes from nothing. They solved it from the foundation of a lot of great scientists, a lot of great mathematicians, and a lot of great previous work. Earlier, I said, we're at the tipping point of a lot of these different fields because we're at that face, that point with Navier-Stokes with a great deal of foundational work already been done. If we could just apply the exhaustive exploration of AI to the previous done work, we're going to see one breakthrough after another breakthrough rather quickly. But we do need a lot of compute to do that. And the reason for that is because a large amount of exploration was necessary. For sure. If it was simply somebody resolving some fundamental first principle theory in order to discover the breakthrough, if it didn't require a vast amount of exploration of space, search space, not space space, but search space, if it didn't require that, then somebody would have found it earlier. Right. You know, people are pretty smart. And so I think that there are many problems like this that are just on a precipice and really wants to just have that massive amount of intelligence exploration to get us there.

### [15:58](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=958s) Navier-Stokes与智能成本的下降
*Navier-Stokes & the Falling Cost of Intelligence*

[17:24](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1044s) 在未来几年，我的预期是智能的成本会相当快地下降。原因很明显，我们正在上线更多算力。这些模型越来越聪明。当你更聪明时，你可以更高效。所以这两件事结合起来，智能的成本会下降，这当然正是他们想要的。这跟摩尔定律没什么不同。如果你回到30多年前，我们有三十年的计算成本每年下降一半。我们希望智能成本每年下降2倍或20倍，因为对它的消费会远远弥补这一点。没错。远远弥补。所以世界上有一大片巨大的空白想要被填补，机会的真空，我们只需要把成本降下来。关于Navier-Stokes，让我印象深刻的是，他们有1万个智能体协同工作。在我看来，这听起来像一家超大规模公司，基本上把所有智能体指向一个目标协同工作。

> In the coming years, my expectation is that the cost of intelligence would fall fairly quickly. And the reason for that is because obviously we're bringing more compute online. These models are getting smarter. When you're smarter, you could be more efficient. And so the combination of these two things, the cost of intelligence will come down, which is of course what they want. And this is no different than Moore's law. If you go back 30-some odd years, we have three decades of computing cost decreasing by a factor of two every single year. And we want intelligence costs to decrease by a factor of 2 or 20 every single year because the consumption of it is going to more than make up for it. That's right. More than make up for it. And so there's a massive empty space in the world that wants to be filled, empty vacuum of opportunity and we just got to get the cost down. I guess what stands out to me about Naviosh Soaks is that they had 10,000 agents working together. It looks like a, to me that sounds like a hyperscale company, basically all pointing agents to work together towards a goal.

[18:40](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1120s) 我认为现在早期阶段，很多是并行搜索探索。但我认为目标是更多这样的智能体能够为共同目标协同工作，我们能够解决更多问题。按你的说法，我认为我们的数学计算表明，解决Navier-Stokes这个实例大概花了400万美元的算力。未来几年做类似事情的成本还会再降5倍、10倍，对吧？所以我们将能够做的事情数量在迅速扩大。我们的雄心会冲上云霄。最初的GPT大概花了，我们随便挑个数，比如说也许2000万美元。如果你现在想训练GPT-1，大概要花多少？一千？一百美元？也许？大概一百美元。好。所以你从2000万美元，就算1000万美元吧，降到一百美元。我们在10年里做到了。对。

> And I think in the early days right now, it's a lot of paralyzed search exploration. But I think the goal is that more of these agents can work together to a common goal and that we can solve a lot more. And to your point, I think our math says it probably cost about $4 million of compute to solve this instantiation of Navi Stokes. The cost to do something like that over the next few years is going to drop by another 5X, 10X, right? So the number of things that we're going to be able to do is rapidly expanding. Our ambitions going to go sky high. The original GPT probably cost, let's pick a number, say, maybe $20 million to do. If you wanted to train GPT-1 now, it'll probably cost you what? A thousand? A hundred dollars? Maybe? It's like a hundred bucks. Okay. And so you went from $20 million, let's say $10 million to a hundred bucks. And we did that in 10 years. Yeah.

[19:44](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1184s) 好。所以我完全预期，与此同时，我们的雄心变得如此之大，我们现在训练的模型，3万亿、10万亿、30万亿。我们现在谈论的是巨型模型，我们希望它如此，所以我们对技术能力的雄心上升了。与此同时，我们对它的使用从十年前到现在大概上升了，我不知道，一百万倍。所以我不认为这个趋势会停止。我认为对智能有相当贪得无厌的需求，而你越聪明、越有智能，你就有越多应用智能的想法。

> Okay. And so I fully expect, meanwhile, our ambition got so large that the models that we're training now, 3 trillion, 10 trillion, 30 trillion. We're talking about gigantic models now and we want it to be and so our ambition for for the capabilities that technology has gone up Meanwhile our use of it has gone up between ten years ago to now probably I don't know a million times and so I Don't think this trend is gonna stop I think this there's a there's a fairly insatiable demand for for intelligence and the more in the smarter you are the more intelligent you are, the more ideas of applying intelligence you have.

[20:27](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1227s) 我这一生学得越多，就越意识到还有更多事情要做。这是因为每当你学到新东西，就会打开一扇新的大门。每当你学到新东西，就会打开一扇新的大门。而每当你打开一扇新的大门，你就会看到更多机会，看到更多挑战，看到更多问题。所以这就是未来。人们说，我们会有更多时间。我完全不同意。我认为对大多数人来说，这种智能将释放更多机会，而他们会想要去追求，因为他们有雄心壮志。

> The more I've learned in life, the more I realize there's more to do. And that's because every time you learn something, it opens a new door. Every time you learn something, it opens a new door. And every time you open a new door, you see more opportunity, you see more challenges, you see more problems. And so that's the future. People say, we're going to have more time. I completely disagree. I think for most people, this intelligence is going to unlock more opportunities, and they're going to want to pursue it because they're ambitious.

### [20:45](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1245s) 我们将如何利用富足？
*What Will We Do with Abundance?*

[20:59](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1259s) 给我举个例子，一个家庭、一个父亲、一个儿子、一个丈夫、一个妻子、一个母亲，谁不想要？谁没有雄心让家人拥有更多？他们只是没有机会，而这些东西将打开机会。如果你回到1930年，发现就是目的，对吧？它深深编码在，我认为，我们的DNA中。而我比以往任何时候都更努力地工作，也比以往任何时候都更享受乐趣。不是因为我必须像我的祖父母那样，对吧？我这么做是因为我想做。

> Give me an example of a family, a father, a son, a husband, a wife, a mother, who doesn't want it, who don't have ambition for their families to have more. They just don't have the opportunity, and these things are going to open up opportunity. If you go back to 1930 Discovery is purpose, right? It's deeply encoded in, I think, our DNA. And I'm working harder than I ever have, having more fun than I ever have. Not because I have to, like my grandparents did, right? I'm doing it because I want to.

[21:55](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1315s) 所以我在这一点上同意你。我对这个想法很着迷，你知道，奇点假设认为，我们现在有所有这些智能体可以协同处理长期运行的任务，它们可以持续专注于任务。我喜欢你提出的‘让彼此保持任务’、作为一个团队协作的想法。嘿，你跑偏了。没错，我喜欢你的想法，我们在所有已经出现的东西之上构建，对吧？所以如果我把完全自动驾驶作为一个例子，我们谈论了十年，却真的没有实现，对吧？但突然间，就像你上Twitter，每个人都在体验特斯拉FSD，体验Waymo，我们做到了。有人问我，今天早上开车过来怎么样？我说，我不知道，因为我上了车，启动了FSD，直到我们开进停车库我都没注意。我一路上都在工作。但如果我在科学领域思考这一点，将会出现的突破，在制药和遗传学领域，我们已经铺设轨道30年了。我们将在许多重大疗法上取得突破，造福世界。当然，ChatGPT已经在帮助父母规划他们的一天，解决孩子的健康问题，帮助他们获得更好的教育，不过请稍微跟我们谈谈，为什么它如此难以突破，Jensen。我们有所有这些正在改善生活、近在眼前的东西。所以这让我有点好奇，为什么它没有突破。

> And so I agree with you on that. I'm fascinated with this idea, you know, the singularity postulates that, you know, we've kind of now have all these agents that can work together on long running tasks, they can stay on tasks. I love your keep each other on task working as a team. Hey, you're off. That's right I love your idea that you know We build on on top of all the things that have come right and so if I think about full self-driving as an example We talked about it for a decade and we really didn't get there, right? But all of the sudden like you go on Twitter everybody experiencing Tesla FSD experiencing Waymo We're here. Somebody asked me. How is the drive here this morning? I said, I don't know because I got in my car and I hit FSD and I didn't pay any attention until we pulled into the parking garage. I was working the whole way here. But if I think about that in the fields of science, the breakthroughs that are going to come, in the fields of pharma and genetics, we've been laying down the tracks for 30 years. We're going to break through on a lot of massive cures for the world. Of course, ChatGPT, is already helping parents plan their day and solve their kids' health issue and helping them get better educated, talk to us a little bit though as to why that's having such a hard time breaking through, Jensen. We have all of these things that are out there improving lives and that are right around the corner. And so that's kind of curious to me as to why that's not breaking through.

[23:23](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1403s) 我想，让我试着回答一下。我们最终不会真正知道。我和很多人谈过。他们告诉了我他们的想法。当然，其中一个原因是，AI或超级智能是软物质。你看不见它。所以全世界都在谈论这项技术，但你无法看见，无法触摸。即使他们已经在从中受益，当他们搜索时，已经在受益。当他们购物时，在受益。当他们使用Netflix时，在受益。你去沃尔玛，也在受益。FedEx能把东西送到我们手中，这一切都在利用超级智能和加速计算。但我们并不擅长讲述这个故事。我经常想——生活就是，他们从中受益。它是无形的。对。所以，如果我们关掉数据中心，关掉超级智能一个周末，我想每个人都会注意到。

> I guess, let me take a shot at it. And we'll ultimately not really know. I've spoken to a lot of people. They told me what they thought. I mean, one of the things, of course, is AI or superintelligence is soft matter. You can't look, you can't see it. And so you've got the whole world talking about this technology, but you can't see You can't touch it. And even though they're already benefiting from it, when they do search, they're already benefiting from it. When they go shopping, they're benefiting from it. When they're using Netflix, they're benefiting from it. You go to Walmart, you're benefiting from it. The fact that FedEx is getting stuff to us. It's all leveraging super intelligence and accelerated compute today. But we don't do a very good job telling that story. I often think- Life is, they're benefiting from it. It's invisible. Right. So, if we shut off the data centers and shut off super intelligence for a weekend, I think everybody would notice.

### [24:27](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1467s) 物理AI与机器人技术
*Physical AI & Robotics*

[24:26](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1466s) 我知道你有一个关于机器人技术的特别问题。是的。我想你今年多次特别提到，机器人的ChatGPT时刻已经到来。我还认为你也说过，这可能是你的下一个1000亿美元市场。所以我想，你今年看到了什么拐点，表明这近在眼前？是的，我低估了那1000亿美元，因为今年前六个月，我们在物理AI上就做了100亿美元，物理AI基本上就是，你知道，以某种方式与物理世界交互的AI，它可以是光子计数CT机，可以是超声，可以是新成像雷达，下一代5G不仅理解无线电波，还将理解各种信号。它是一个传感网络。当然，还有机器人、自动驾驶汽车、制造系统、视觉检测、地理空间卫星，对世界成像并将其重建为3D。

> I know you had a question about robotics in particular. Yeah. I guess you've notably said a few times this year that the ChatGPT, moment for robotics is here. And I think you've also said that it could be your next $100 billion market. So I guess what are you seeing now in the inflection point in this year that this is around the corner? Yeah, I understated the hundred billion dollars because we're already the first six months of this year We did ten billion dollars in physical AI is basically, you know AI that's somehow interacting with the physical world and it could be photon photon counting CT machines could be ultrasound. It could be new imaging radars Next generation of 5g not only does it understand radio waves is going to understand signals of all kinds. It's a sensing network Of course, robotics, self-driving cars, manufacturing systems, visual inspections, geospatial satellites, imaging the world and reconstructing it into 3D.

[25:29](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1529s) 所以基本上，如果你看整个电磁频谱，从可见光到不可见光谱，都有AI在运作。所以，我们正在整个范围内开展工作。作为AI，有感知部分，有推理部分，然后还有规划和执行部分。当你最终来到规划和执行，那就是自动驾驶汽车、机器人等等。在仓库配送车辆中，拾放机器人你不需要那么精确地编程，因为它自己学习，你知道，一直到人们谈论的消费机器人，进入你家，整理你的床等等，或者医疗机器人、医院机器人，有很多很好的例子，精确地将你的处方送达患者，拥有机器人护理，真正24/7每一秒都在保持患者安全。

> And so there's basically, if you look at the entire electromagnetic spectrum, there's AI stuff going on from visible to invisible spectrums. And so, and we're doing work across that entire thing. And as an AI, there's the perception part of it, there's the reasoning part of it, and then there's the planning and actuation part of it. When you finally come to planning and actuation, then that's self-driving cars and robotics and things like that. In warehouse delivery vehicles, pick and place robots that you don't have to program You know as precisely because it learns by itself, you know all the way to people talking about consumer robots You know coming into your house, you know making your bed and things And so so or health care robots hospital robots lots of really good examples there You know precisely delivering Exactly your prescription to the patient Having having robotic care literally 24 7 every single second to keep patients safe.

[26:38](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1598s) 所以有大量很好的例子，整个手术室将完全机器人化。所以大量伟大的技术正在部署。所以如果你看这整个事情，我们是世界上最大的物理AI公司。我们在前六个月做了100亿美元。去年我们做了100亿美元。那是我们第一个每年100亿美元。前一年我们做了50亿。是的，好的。所以从50亿到100亿，假设我们做了200亿。是的，谁知道呢？所以也许100亿有点低估了。而且我认为推理部分，归根结底，对我们重要的一切最终都触及物理世界，而我们生活在物理世界中，所以我认为将软件计算智能带入一个历史上主要是机械、电气的行业，这是一个巨大的机会。

> And so just lots and lots of great examples, entire surgical rooms that are gonna be completely robotic. And so lots of great technology being deployed. And so if you look at this whole thing, we're the world's largest physical AI company. We did $10 billion in the first six months. We did $10 billion in the last year. And that was our first $10 billion a year. The year before that we did five. Yeah, okay. And so when five ten, let's say we did 20. Yeah Who knows? So maybe the hundred was a little understated and And I think the reasoning part of it is is everything that matters to us in the final analysis touches the physical world and and we live in the physical world and so I I think the the market opportunity for bringing software computing intelligence into an industry that largely historically was mechanical, electrical, that's a big opportunity.

[27:36](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1656s) 我的意思是，看看Elon对汽车行业所做的，显然把它变成了一个计算行业、一个技术行业。所以我认为每家汽车公司都将成为技术公司，每家运输和接收公司、物流公司都将成为技术公司。我为FedEx感到非常兴奋，他们显然将成为AI公司，显然将成为技术公司。沃尔玛，显然将成为技术公司、AI公司。所以为他们感到非常兴奋。整个电信行业将再次被重塑为技术行业。他们很久以前是技术行业。他们变成了服务行业。他们将再次成为技术行业。我很高兴与AT&T、Verizon以及欧洲和世界各地的公司合作，帮助他们重塑自我。所以你看看每一个触及物理世界的行业。我，有一个用AI重塑它们的伟大想法，你知道，我们，我刚刚在华盛顿，奇怪的是，你知道人们把我们归入芯片。是的，你知道，所以我们实际上在仓储和物流。是的，我们在医疗保健。是的，我们在汽车。我们在电信行业。是的，每个行业。是的，所以我们在照明。我们在，我们在生命科学，你知道。所以所有这些，娱乐，显然很长一段时间，NVIDIA实际上主要是一家娱乐公司。

> I mean look at obviously what Elon has done to the automotive industry, turned it into essentially a computing industry, a technology industry. So I think every car company is gonna be a technology company, every shipping and receiving company, logistics company is gonna be a technology company. I'm super excited for FedEx, they're obviously gonna be an AI company, they're obviously gonna be a technology company. Walmart, obviously going to be a technology company, an AI company. So super excited for them. The entire telecommunications industry will be reinvented again into a technology industry. They were a technology industry a long time ago. They became a services industry. They're going to be a technology industry again. And I'm delighted to work with AT &T and Verizon and Europe around the world as they reinvent themselves. And so you take a look at every one of these industries that touch the physical world. And I there's a great idea for reinventing them with AI, you know, we we I was just in Washington weird, you know people put us in chips. Yes, you know and and so we're actually in Warehouse and logistics. Yes, we're in healthcare. Yeah, we're in automotive. We're in teleconvery industry Yeah, every industry. Yeah, and and so so we're in light. We're we're in we're in life sciences, you know So all these entertainment, obviously for a long time, NVIDIA was really vastly an entertainment company.

[28:59](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1739s) 所以每一个垂直领域，NVIDIA都有庞大的技术栈深入其中，我们深度融入生态系统，在那里有出色的合作伙伴关系。所以有一天，人们可能不再把我们归类为芯片公司。策略将是把汽车领域的剧本，我想，插入到每一个行业中。是的，这是最终的贡献。但在每一个领域中，构建物理AI平台需要三台计算机。所以你需要一台计算机来创建大脑。那就是训练系统。但非常重要的是，这也是你们将看到实验室也会这样做的地方，我们需要构建互联网的数字孪生，这样他们就不必把AI智能体放到互联网上去学习如何使用它。就像我们创建Omniverse作为仓库、建筑、城市和物理世界的数字孪生一样，我们把机器人放入数字孪生。数字孪生学习如何成为机器人。

> So each one of these verticals, NVIDIA has a massive stack that's inside of it and we're very deep in the ecosystem and we have excellent partnerships there. And so one of these days, people might not categorize us as a chip company anymore. The playbook would be to take the alpomio automotive, I guess, playbook and insert it into each one of these industries. Yeah, that's the final contribution. But in each one of these, it takes three computers to build a physical AI platform. So you need a computer to create the brain. That's the training system. But very importantly, and this is where you're going to see the labs also do this, they We need to build a digital twin of the internet so that they don't have to put their AI agents out into the internet to learn how to use it. Just like we created Omniverse as the digital twin of warehouses and buildings and cities and the physical world, we put our robot into the digital twin. The digital twin learns how to be a robot.

[30:06](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1806s) 然后我们部署到最后一台计算机，也就是比如机器人、汽车或手术器械等内部的那台计算机。所以它需要三台机器人、三台计算机和三套完整的栈。我们为每一套都构建完整的栈。我们是世界上唯一这样做的公司。然后，如果你是一家想要端到端构建的公司，你可以使用我们两三台或一台计算机，我们不在乎。所以每台计算机都有自己的软件栈。我们在所有垂直领域都这样做。但当然，这些能力是双刃剑，这就是为什么我们听到很多人谈论安全。你一直对末日论叙事持高度批评态度。只有负责任地构建技术，你才能达到那里。没错。所以第一件事就是负责任地构建技术。注意‘构建’这个词。这是一个工程词汇。它关乎技术，关乎工程。

> Then we deployed into the last computer, which is the computer that's inside, for example, a robot or a car or a surgical instrument or whatever it is. And so it takes three robots, three computers, and three complete stacks. And we build complete stacks for each one of them. We're the only company in the world that does that. And then if you're a company that wants to build it end to end, you could use two or three or one of our computers, we don't care. And so each one of a computer has its own software stack. We do the same across all of our verticals. But of course, these capabilities cut both ways, which is why we hear a lot of people talking about safety. You've been highly critical of the doomer narrative. You can only get there if you build the technology responsibly. That's right. And so the first thing is build the technology responsibly. Notice the word build. It's an engineering word. It's about technologies, about engineering.

### [30:36](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1836s) 负责任地建设
*Building Responsibly*

[31:01](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1861s) 工程师、科学家、研究人员，我们必须安全、负责任地构建这项技术。问题是，你能做到吗？答案是绝对可以。因为如果你做不到，那你根本就不应该构建它。当然。所以我们要么必须相信我们能，要么就不应该构建它。所以我相信我们能。让我和你一起推理一下。第一部分是，有人说AI是不可控的，因为不可能完美对齐。对齐对所有人来说意味着这个。你告诉某人，我要你做这个。所以任务很明确。任务很明确。

> Engineers, scientists, researchers, we have to build this technology safely, responsibly. And the question is, can you? And the answer is absolutely yes. Because if you can't, then you shouldn't build it at all. Of course. And so we either have to believe we can or we shouldn't build it. And so I believe we can. And so let me just reason about it with you. So the first part of it is they say, some people say that AI is uncontrollable because it's impossible to align perfectly. Alignment means this for all the audience. You tell somebody, I want you to do this. And so the mission is clear. The mission is clear.

[31:47](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1907s) 对你来说很明确，因为你用英语说了。所以可能是，我要你解这个二次方程。所以任务是解二次方程。对齐是你怎么做？有很多方法。比如，你背下来了。对吧？所以答案就是，你知道，不管是什么。或者你在教科书上看到了答案，所以你查了一下。对。对吧？你参考了一个参考。或者你走到一个抽屉前，你知道抽屉里有所有答案。你打开抽屉，挑出答案。有人称之为作弊。我称之为打开抽屉并检索答案。这就是计算机所做的。有很多不同的方法可以解决那个问题。当然，那个非常明确。这是另一个。在生活中取得成功。任务说起来容易。事实上，我们一直对彼此说。你知道，我们对孩子说，对朋友说，让我们尽力为社会做贡献。那么现在问题是，怎么做？

> And it's clear to you because you said it in English. And so it could be, I want you to solve this quadratic equation. And so the mission is to solve the quadratic equation. The alignment is how do you do it? There are many ways to do it. For example, you memorized it. Okay? And so the answer is, you know, whatever it is. Or you saw the answer in a textbook, so you looked it up. Right. Okay? You refer to a reference. Or you went to a drawer and you know that the drawer had all the answers. You opened the drawer, you picked out the answer. Some people call that cheating. I call that opening the drawer and retrieving the answer. And that's what computers do. There's a lot of different ways you could solve that problem. Now, of course, that one is very explicit. Here's another one. Be successful in life. The mission is easy to say. And in fact, we say it to each other all the time. You know, we say to our children, we say to our friends, let's do our best to contribute to society. Well, now the question is how?

[32:49](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=1969s) 有太多方法。你如何合理地对其，每个人的文化和背景，有些人认为为社会做贡献意味着慷慨、善良。有些人可能意味着为社会做贡献赚很多钱然后做慈善。有些人可能，等等等等。不是吗？是的。对吧？有些人可能只是认为独自建立一家公司或独自做某事就足以对社会做贡献，创造一件美丽的艺术品，或者所有这些都成立。正确。所以问题是，对齐是什么？这取决于你的背景、你的文化，也许与你住在哪里、你来自哪里有关。有无数种方式。所以我的观点是，对齐将永远是一个问题。如果我们依赖完美对齐来确保安全，那么我们永远无法达到。所以答案是这个。

> There are so many ways. How do you reasonably align and everybody's culture and background and some people think that contributing to society means to be generous, to be kind. Some people might mean that contributing to society make a lot of money and then be philanthropic. Some people might, so on and so forth. Isn't that right? Yes. Okay? And some people might just think that building a company alone or doing something alone is a sufficient contribution to society, creating a beautiful piece of art, or all of it is true. Correct. And so the question is, what's the alignment? And it just depends on your background, your culture, maybe it has something to do with where you live, where you came from. There's just an infinite number of ways. And so my point is this, alignment is going to be a forever problem. If we depend on perfect alignment in order to ensure safety, then we're never going to get there. And so the answer is this.

[33:46](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2026s) 我们对待的方式，我们思考婴儿的方式，以及我们养育孩子时的做法，还有，你知道，我们养宠物或者不管是什么，还有员工。你从隔离约束开始。你给他们权利，给他们使用工具、获取信息的权限，给他们任何责任、能力、资金、资源的权限。你根据你要求他们做的工作来给予他们权限。你一开始是零信任。你雇一个人的时候，一开始就是零信任。你有一个一岁大的婴儿时，你把他们放进游戏围栏里，你就是假定零信任。所以零信任，你给他们最低限度的权利，然后随着你了解得更多，你扩大那个边界。所以每当我们有一个新的 AI，我们都必须假定同样的事情。每当你测试一个 AI，零信任，最小边界，最小权利，然后随着你测试并给自己建立信心，你再扩大它。所以隔离约束，非常严格的隔离约束，对任何自主系统都非常重要。第二件要做的事，监控。

> The way we treat, the way we think about babies and when we were raising kids and, you know, we have pets or whatever it is, employees. You start with containment. You give them the rights that the access to tools, the access to information, the access to whatever responsibility capabilities that they have, funds, resources. You give them access based on the job that you ask them to do. You start out with zero trust. When you hire somebody, you start out with zero trust. When you have a baby that's one-year-old, you put them in a playpen, you assume zero trust. And so you zero trust, you give them minimum rights, and you expand that boundary as you learn more. And so whenever we have a new AI, we have to assume the same thing. Whenever you're testing an AI, zero trust, minimum boundaries, minimum rights, and you expand it as you test and give yourself confidence. So containment, very tight containment is really important to any autonomous system. Second thing you do, monitoring.

[34:50](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2090s) 监控必须是带外的。原因是，如果你是一个聪明人，嘿，听着，我要坐在这里看着你工作，确保你做对了。对吧？对吧？我要坐在这里看着你，确保你正确服务客户，并且工作足够努力。对吧？那个自主系统在你不在看的时候会表现得不同。所以Brad在房间里和我在一起吗？是的，没错。没错。

> And the monitoring has to be out of band. And the reason for that is because if you're an intelligent person, hey, listen, I'm gonna sit here and watch you do your work to make sure you're doing it right. Right. Okay? I'm gonna sit here and watch you make sure that you're serving customers properly and that you're working hard enough. Okay? That autonomous system is gonna behave differently than when you're not watching. So Brad's in the room with me or not? Yeah, exactly. Exactly.

[35:17](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2117s) 所以监控必须是带外的，并且必须有能力观察自主系统正在做的事情是否与政策不一致或与对齐不一致。并且你有能力升级或关闭它。直接关掉。所以这类能力必须存在。那是工程。那是技术。这是可能的。我们最近宣布了一个叫Open Shell的东西。我们现在有超过100家公司加入我们。我们在世界各地有大量的活动，因为每个人都想要同样的东西。从最高层开始，当我在白宫环顾桌子时，桌子周围的每个人，他们的父母，他们有孩子，我认为这个国家有一种观点，认为一些超级智能的领导者只是在YOLO，并不真正关心。他们鲁莽。他们只想抓住所有的钱。然而当我环顾那张桌子时，我在那里看到的没有一个人想要鲁莽，想要YOLO。事实上，那里的每个人都想在快速和安全之间找到正确的平衡。那是总统的指示。我们稍微辩论了一下那意味着什么。

> And so the monitoring has to be out of band and it has to have the ability to observe that the autonomous system is doing something that was inconsistent with the policy or inconsistent with alignment. And you have the ability to either escalate or shut it down. Just turn off. And so those kind of capabilities have to exist. That's engineering. That's technology. And it's possible. And we announced it recently called Open Shell. We have over 100 companies now joining us. We've got lots and lots of activity around the world because everybody wants the same thing. Starting at the highest level, when I looked around the table at the White House, everybody around that table, their parents, they have kids, I think there's this view in the country that there are some leaders of super intelligence that are just yoloing, that don't really care. They're being reckless. They just want to grab all the money. And yet when I looked around that table, There wasn't a person in there that I saw that wants to be reckless, that wants to YOLO. In fact, everybody there wanted to find the right balance between fast and safe. That was the president's directive. We debated a little bit what that meant.

[36:29](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2189s) 关于如何实施有不同的版本。你会对美国公众说什么，他们担心我们有一个鲁莽的总统或鲁莽的领导人？你看到任何证据吗？你有孩子。你担心你通过超级智能安全和保障所创造的未来吗？技术对吧，是AI的一部分对吧？有一种奇怪的表述，认为AI技术和安全与保障是正交的。是的，有点像安全气囊和刹车系统，对吧，以及雷达、摄像头、ABS、车道保持，所有这些技术不知何故独立于汽车。你们明白我在说什么吗？对汽车唯一重要的是引擎。事实上，整辆车都是汽车技术的一部分。当我说我们需要加速AI技术时，我指的是我们需要加速安全技术。我们必须加速安全技术。它们都必须被加速。如果你现在问我，我想每个人都会说，让我们加速安全。让我们加速安全技术。没有人说，不，让我们放慢那个。

> There are different versions of how to implement that. What would you say to the American public who fears that we have a reckless president or reckless leaders? Did you see any evidence of that? You have kids. Do you worry about the future that you're bringing into existence with super intelligence safety and security? Technology right are part of AI right? There's some weird formulation that AI Technology and safety and security are thought orthogonal. Yeah, it's kind of like the airbags and the and the braking system Right and the radars of cameras the ABS, the lane keeping, all of that technology is somehow independent of the car. Do you guys understand what I'm saying? The only thing that matters to the car is the engine. In fact, the whole car is part of the car's technology. When I say we need to accelerate AI's technology, what I mean by that is we need to accelerate the security technology. We have to accelerate the safety technology. They all have to be accelerated. If you ask me right now, I think everybody would say, let's accelerate security. Let's accelerate safety technology. Nobody says, no, let's pace that.

### [36:46](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2206s) 安全、保障与安全气囊的类比
*Safety, Security & the Airbags Analogy*

[37:44](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2264s) 我们应该加速那个。这都是同一件事的一部分。你不能将能力和功能与安全和保障分开。它们是一体的。当你提供服务时，没有人会说，我真的很喜欢那个功能，在- 对。我们把它当作一件事，整个行业是一件事，叫他的新好朋友Dario。他提到了他。他说，你知道，向我解释你的版本，快快快快快快。你说谁赢，AI赢。

> We should accelerate that. It's all part of the same thing. You can't separate capability, functionality from safety and security. They go as one. And when you deliver the service, nobody's going to say, I really like the the functionality of in- Right. And we treat it as all one thing, and the industry's all one thing, called on his new best friend Dario. He referred to him. He said, you know, explain to me your version of fast fast fast fast fast fast. You said whoever wins, AI wins.

[38:46](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2326s) 他说，我们都在加速，他们在加速，他们想走得快，他们想赢得与中国的竞赛，但也承认，沿途，就像你刚才说的，我们必须确保安全系统也在加速并保持同步。所以我认为那是一个重要的时刻，你有，你知道，因为我认为人们把他定位成好像他反对这个，但在我看来，你们都非常一致。你和他以及Greg和其他前沿实验室一起去了内阁会议室签署，你知道，这个协议。那个房间里有没有关于意见分歧的议论，还是因为每个人都，你显然让它签了，有没有人必须被锁头，还是每个人都自愿？不，每个人都非常建设性，非常富有成效。

> He said, that we're all speeding up, that they're speeding up, that they want to go fast, that they want to win the race with China, but also acknowledge that along the way, we just, to your point you just made, we have to make sure the safety systems are also speeding up and also keeping pace. So I thought that was an important moment where you had, you know, because I think people have positioned him as though he's in opposition right to this, but it seemed to me that you were all very much on the same page. You went to the cabinet room with him and Greg and the other Frontier Labs to sign, you know, this agreement. Was there any chatter, you know, in that room about differences of opinion or was Because everybody, you obviously got it signed, did anybody have to be put in a headlock or did everybody go willingly? No, everybody was very constructive, very productive.

[39:38](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2378s) 我认为，无一例外，每个人都致力于安全可靠地推进这个行业。原因在于，如果不安全、不可靠，功能就永远没有机会展现。能力也永远没有机会闪耀。所以我们知道，安全与可靠始终是优先事项。当然，要认识到这个行业最近刚刚经历了一次转型。就在六个月前。这项技术几乎还无法正常工作。所以，如果没有功能和能力，确保它安全可靠又有什么意义？根本没人会在意。显然，AI的首要任务是证明它能做有用的事。而它在六个月前确实达到了那个节点。所以现在，它真的想要被部署，并且广泛部署。毫无疑问，整个行业现在都有一个战略性的紧迫任务，一个不可思议的紧迫任务，以及极大的紧迫感，来确保我们部署的任何东西都是安全可靠的。所以我认为每个人都完全确信，完全一致，功能、能力、安全、可靠，它们都处于完全相同的位置，都是第一优先级。所以问题在于赢。

> I think that without exception, everybody is dedicated to advancing the industry safely and securely. The reason for that is because if it's not safe, it's not secure, the functionality will never get a chance to speak. And so the capability will never get a chance to shine. And so we know that the priorities are at all times safety and security. Of course, recognizing that this industry came through a transition just recently. It wasn't but six months ago. This technology barely worked. And so without functionality and capability, what's the point in making it safe and secure? Nobody even cares about it. And so obviously, the first job of AI is to demonstrate it could do something useful. And it really reached that point six months ago. And so now, it really wants to be deployed and broadly deployed. And so there's no question that the entire industry now has a strategic imperative, an incredible imperative, and great urgency to make sure that whatever we deploy is safe and secure. And so I think everybody is completely convinced, aligned, functionality, capability, safety, security, they're all at exactly the same place, priority number one. And so the question is winning.

### [40:59](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2459s) “美国获胜”意味着什么
*What “America Winning” Means*

[41:01](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2461s) 我想重新定义赢，不是技术行业赢，不是仅仅AI行业赢。美国赢意味着美国每个行业都赢。医疗保健。没错。没有什么比这更让我高兴的了。美国赢的定义是，当我们谈论礼来赢、沃尔玛赢、联邦快递赢，每一个都赢，波音赢、美国军队赢。每个公民都赢。每个公民都赢。所以我们的大学赢，我们的高中赢。这就是赢的定义。赢，如果只是技术赢，但没人使用它，那就没有赢。

> I would reframe winning not of the technology industry winning, not of just the AI industry winning. America winning means every industry in America winning. Health care. That's right. Nothing would give me more joy. The definition of America winning is when we talk about Lilly winning, Walmart winning, FedEx winning, every single one, Boeing winning, the US military winning. Every citizen winning. Every citizen winning. And so our university's winning, our high school's winning. That's the definition of winning. The winning, there is no winning if it's just a technology one, but nobody used it.

[41:42](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2502s) 最大的风险当然是，我就直说吧，我就说基线。如果我们不安全可靠地构建它，就不会有技术，就不会有AI技术行业。这对我们来说是一个生存性问题。如果我们让它成为对其他所有人的生存性行业，那不会发生，因为它首先会是我们的生存性问题。话虽如此，对我们来说赢的定义，这是我总是要稍微改变一点的部分。我想让我们转向。我喜欢这个。我想让我们转向。当我看到AI赢，美国AI赢，这些天我们从空中俯瞰，我们花了很多时间在空中，我们俯瞰每个州都赢，每个县都赢，每个社区都赢，每个公司都赢。每个人都赢了，因为我们能负担得起在医疗保健上花更多钱。我们能负担得起在教育上花更多钱。我们能负担得起在州际、你知道，商业、高速公路等等，所有国家基础设施上花更多钱。我听到你说的是，赢的定义是，当我们回顾25、30、40年后，经济规模需要扩大10倍。每个人的生活都需要更好，并且对所有人来说都是生存性和首要的，而且这必须以符合人类利益的方式来做，这就是我们有很多要谈的。

> The greatest risk is, of course, we're going to, I'll just take it, I'll just say baseline. There will be no technology, there will be no AI technology industry if we don't build it safely and securely. This is an existential issue for us. If we make it an existential industry for everybody else, well, that's not going to happen because it would be an existential issue for us first. Having said that, the definition of winning for us, and this is the part I always have to change a little bit. I want us to pivot. I like that. I want us to pivot. When I see AI winning, America AI winning, I'm looking down from these days, we spent a lot of time in the air, and we're looking down every state's winning, every county's winning, every community is winning, every company's winning. Everybody won because we could afford to spend more on healthcare. We could afford to spend more on education. We could afford to spend more on interstate, you know, commerce, highways, et cetera, all of the infrastructure of the country. What I hear you saying is the definition of winning is that when we look back 25, 30, 40 years from now, the economy needs to be 10 times as large. Everybody's lives need to be better and existential and primary to all and that is it has to be done in a way that's aligned with human interest, which this is the We have a lot to get to.

### [43:03](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2583s) 三条腿的凳子：美国的超级智能框架
*The Three-Legged Stool: America’s Superintelligence Framework*

[43:03](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2583s) 大多数人不了解我们国家的安全框架是什么。我说过我们拥有世界上最强大的超级智能安全框架。我们显然是地球上唯一一个今年早些时候用Mythos实际制裁了一个模型并将其下架的国家。但我把它描述为一个三条腿的凳子。我们有Casey和商务部对每个新前沿模型的前沿模型审查。所以他们做所有的工作。如果有新模型出来，他们审查它。他们测试它，看他们是否认为它会在市场上引起问题。现在我们有你帮助架构的《白宫超级智能协议》，内部和外部审计，董事会监督，以及随之而来的所有责任。第三，我们有SI Force，现在由Jay Clayton领导，当然还有Emil、Scott和Andrew，它正在做关键的行业协调、网络安全等等。所以凳子的这三条腿。

> Most people don't understand what our security framework is in the country. I've said we have the most robust super intelligent security framework in the world. We're clearly the only country on the planet that's actually sanctioned a model and pulled it off the shelf with mythos earlier this year. But I describe it as a three leg stool. We have the frontier model review of every new frontier model by Casey and the Department of Commerce. So they do all their work. If new models coming out, they review it. They test it for whether or not they think it's going to cause a problem in the marketplace. Now we have the White House Accord on Super Intelligence that you help to architect, internal and external audits, board oversight, and all of the liability that comes with that. And then third, we have the SI Force, now led by Jay Clayton, of course, Emil and Scott and Andrew, and that's doing critical industry coordination, cyber, et cetera. So these three legs on the stool.

[44:00](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2640s) 外面有人，我最近听到盖茨说，我们做得不够，我们需要制定新法律等等。这个三条腿的凳子在我看来，在很短的时间内已经成形，似乎已经很多了。作为超级智能领域最重要的公司之一的领导者，显然，你对我们要引领的未来感到舒服吗？这是向前迈出的好一步，是我们现在需要达到的位置。是的，通往未来的最好方式就是迅速采取行动。

> There are people out there, I heard Gates recently saying, we're not doing enough, we need new laws on the books and everything else. That three-legged stool seems like a lot to me in a very short period of time that's come together. As a leader of one of the most important companies, obviously, in super intelligence, do you feel comfortable for that future that we're trying to usher in? That is a good step forward where we need to be right now. Yeah, the best way to get to the future is take action quickly.

[44:31](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2671s) 如果我们争论什么是完美的还是什么是第一个，我们最终会争论完美，我们永远达不到。所以我喜欢这个三条腿的凳子是可操作的。它们是切实的。它们有真正的牙齿。正在发生。现在正在发生。你可以，你知道，我们现在正在做。我们现在正在采取行动。我们现在采取行动的原因是，这对我们的公司有好处，对我们的行业有好处，对我们的国家有好处，对社会有好处，全方位都有好处。如果我们不做这些事情，你知道，结果就不那么理想。我们就无法追求我们都梦想的那个乐观的未来。所以如果你回去看看另一个行业，我不记得有另一个行业目前处于当前状态，并且现在已经制定了所有这些跨公司、跨行业、政府到行业的协调，以确保我们以负责任和乐观的方式推进这件事。是的，我只是认为让人们理解这一点非常重要，因为当我听到参议员或他在电视上谈论这件事时，双方都是，说实话。

> And if we argue about what's perfect versus what's first, we'll end up arguing about perfection and we'll never get there. And so I love the fact that these three-legged stool are actionable. They're tangible. They have real teeth. Happening now. It's happening right now. You could, you know, we're doing it now. We're taking action now. And the reason we're taking action now is because it's good for our company, it's good for our industry, it's good for our country, it's good for society, it's good all the way around. And if we don't do these things, you know, the outcome is less desirable. We don't get to pursue that optimistic future that we all dreamt about. And so if you go back and look at another industry, I don't remember another industry that is currently at its current state and have now formulated all of these cross company, cross industrial, government to industry coordination to make sure that we advance this in a responsible and optimistic way. Yeah, I just think it's so important for people to understand because when I hear senators or is talking about it on television, on both sides, to be perfectly honest.

[45:50](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2750s) 嗯，他们首先必须了解我们在这里做什么。他们甚至不知道我们在做什么。是的，是的。对吧？比如我们正在做大量非凡的工作。顺便说一句，这些不是自愿的事情。不会有一个前沿模型被发布而Casey不审查、商务部不审查。他们展示了，他们在出口管制下有权力将它们下架。我会做两件事。第一，来花些时间了解实际需要做什么。问题是什么？如果你想要解决方案，你至少必须理解问题是什么。所以第一，来学习。我们有充足的时间，并不是世界上所有的时间都用于这类事情。就来和我们谈谈，来和我们的行业谈谈，学习。第二件事是了解我们建立的解决方案，并试着看看是否有想法可以使其更好。当然，我们所有人都完全开放。

> Well, they have to learn about what we're doing here first. They don't even know what we're doing. Yeah, yeah. Right? Like we are doing an extraordinary amount. And by the way, these aren't voluntary things. There's not going to be a frontier model that gets released that Casey is not reviewing and the Commerce Department is not reviewing. And they demonstrated, they have the authority under export controls to pull them off the I'll do two things. One, come and spend some time and learn what actually has to be done. What's the problem? If you want a solution, you've got to at least understand what the problem is. So one, come and learn. We've got plenty of time and not all the time in the world for this kind of stuff. And just come talk to us, come talk to our industry and learn. The second thing is learn about the solution that we put in place and try to see if there are ideas to make that better. of course, we're all completely open-minded.

[46:40](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2800s) 但要建设性。我认为这是超级智能协议第四点的一部分，这不是静态的。这是动态的。这个小组将定期聚会。我们将根据技术发展需要更新框架。但我想现在转向算力。当然，长期运行的智能体以数千或数百万的团队工作来解决复杂任务，这一切都消耗大量更多的算力。我们的朋友Dylan在Semianalysis说，我们明年将上线43吉瓦的新AI算力。为了提供背景，我们会在网上分享一张图表。这比当今世界上所有AI算力还要多。所以我们明年将上线那么多。后年，再增加40吉瓦算力，你是那个知道每一块GPU分配在哪里的人，所以当你坐在这里时，这听起来对吗？你认为我们会带来那么多算力吗？

> But be constructive. And I thought that was part of the fourth point on the Accord on superintelligence was this is not static. This is dynamic. This group is going to get together routinely. We're going to update the framework as required when the technology evolves. But I want to move forward now to compute. Of course, long-running agents that work in teams of thousands or millions to solve complicated tasks, it all consumes massively more compute. Dylan, our friend at Semianalysis, has said we're going to bring on board 43 gigawatts of new AI compute next year. Just to contextualize it, we'll share a chart with people online. That's more than all AI compute that exists in the world today. So we're going to bring that on board next year. The year after that, another 40 gigawatts of compute, you are the one, you know where every single GPU is allocated, so as you sit here, does that sound right to you? Do you think we're going to bring on that much compute?

### [47:07](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2827s) 2027年AI瓶颈：算力建设
*2027 AI Bottlenecks: The Compute Build Out*

[47:43](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2863s) 你认为我们听到的一些关于许可问题或供应链瓶颈或当地抗议的事情，你认为这个数字会——我是说，根据你所看到的，会有实质性不同吗？我认为需求肯定存在，而且更多。需求肯定存在，而且更多。当然，挑战在于有很多障碍。当你从数字世界转向物理世界时，有很多障碍。有土地许可。有你需要上线的发电。当然，你需要建筑工人、装配工、水管工以及各种——仅仅建造这些就创造了100万个新工作岗位。创造了100万个新工作岗位，他们的技能不是可互换的。他们有很多技能，特别是那些建造这些500亿美元数据中心的人，你会想要一些技能高超的人。

> And do you think that some of the things we're hearing out there about permitting issues or bottlenecks in the supply chain or local protests, do you think the number is going to I mean materially different than that based upon what you're seeing. I think the demand is definitely there and more. The demand is definitely there and more. The challenge, of course, is that there are a lot of obstacles. When you move from digital to the physical, there are a lot of obstacles. There are land permitting. There's power generation that you got to bring online. Of course, you need construction workers and outfitters and plumbers and all kinds of- Created a million new jobs just building these. Created a million new jobs and their skills are not fungible. They have a lot of skills and particularly the ones that are building these $50 billion data centers, you're going to want some people with great skills.

[48:45](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2925s) 在所有这些设施的建设中，每一个环节都有一个关键部分，需要一些非常特定的技能。所以我认为，然后当然还有正在发生的各种暂停令，以及我们在全国范围内看到的种种问题。所以我认为，我的猜测是，土地、电力、外壳，这个建设过程将成为实现目标的最重要障碍之一。所以，正如你所知，我们公司，现在具体说我们，我们正在做几件事来缓解这个问题。第一，因为我们是唯一一家服务所有CSP的公司，我们服务所有NCP，我们通过Dell、HP、Lenovo等公司拥有很棒的企业业务。

> There's a critical part of every single one of the build out of these things where you need some very specific skills. And so I think that, and then there's of course whatever moratoriums that are happening, whatever issues that we're seeing around the country. And so I think this is, my guess is that land power shell, the build out will be one of the most important obstacles to reaching that. And so our company, as you know, now speaking specifically for us, there's several things that we're doing to mitigate that. And one, because we're the only company that serves every CSP, we serve every NCP, we have great enterprise businesses through Dell and HP and Lenovo and others.

[49:43](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=2983s) 而且因为我们服务每一个行业，我们覆盖AI的整个生命周期。当然，从数据处理，到训练，到后训练，到推理，到智能体处理，以及我们加速的所有工具使用。我们在市场上的覆盖面是最大的。世界上每一家代工厂都在为我工作。每一个市场渠道都与我们的架构相连。所以我们在某个特定的州、某个特定的社区、某个特定的困难时期。目前我们的需求远远大于供给，我们只需相应地调整供应链。我们还有另一点，我不太喜欢这个答案，但正如你所知，NVIDIA也是这个星球的基础设施。所以无论是在欧洲还是……

> And because we serve every single industry And we're across the entire lifecycle of AI. Everything from, of course, data processing, to training, to post-training, to inference, to agentic processing, and all of the tool use that we accelerate. Our exposure to our go-to-market is the largest of all. Every contract manufacturer in the world is working on my behalf. Every go-to market is connected to our architecture. And so we have one particular state, one particular community, one particular hard time. Our demand is so much greater than our supply at the moment, we'll just adjust our supply chain accordingly. The other thing that we have, and I don't love this answer as much, but as you know, NVIDIA is also the infrastructure for the planet. And so whether it's in Europe or...

[50:42](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3042s) 我们现在在非洲、中亚、东南亚、拉丁美洲、中东等地都有早期的发展。所以我不喜欢美国跟不上这一点。但具体到我们的业务，我们拥有如此全球化的覆盖。全世界都有建设需求。如果我们不在这里建设，我想我们都必须认识到，世界上每个国家都希望在那里建设。

> And we've got early developments now in Africa, in Central Asia, in Southeast Asia, in Latin America, in, you know, in the Middle East, and so on and so forth. And so I don't love that the United States can't keep up. But for our business specifically, we have such a global reach. And there's build out demand all around the world. If we don't build it out here, I think we just all have to recognize that every country in the world would love to have R built out there.

[51:17](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3077s) 我认为具体到我们的业务，可能是最有韧性的。原因就是我提到的所有那些。我们覆盖AI的整个谱系。我们覆盖计算的整个谱系。我们存在于每一朵云、每一个地区。所以我们公司的韧性要大得多，但我确实认为土地、电力、外壳明年将是一个具有挑战性的部分。我猜我从市场上听到的是这样一种想法：也许有50吉瓦的已宣布芯片正在推进或讨论中，而也许有35到40吉瓦的土地、电力、外壳可能上线。所以，多少算力能装进那些土地、电力和外壳，以及以什么时间线，是一个悬而未决的问题。对，但我想如果我们从整个……Clark刚才打断你一下，当然，如你所知，Nvidia的架构是最高效的，对吧？它每瓦产生的token最多。嗯哼。所以如果全世界的瓦特电力能源减少，就会把他们的选择转向最高效的版本。原因在于他们仍然想赚取收入。

> I think that our business specifically is probably the most resilient. The reason for that is all the things that I mentioned. We cover the entire spectrum of AI. We covered the entire spectrum of computing. We're in every single cloud, every single region. So the resilience of our company is much greater, but I do think that land power shell is going to be a challenging part next year. And I guess what I hear from the street is there's this idea that maybe there's 50 gigawatts of announced chips that are underway or talked about, and maybe 35 to 40 gigawatts of land power shell that might come online. So, what compute fits into that land power and shell and on what timeline is a question up in the air? Yeah, but I guess like if we think of this through the entire and just Clark just interrupt you there and and You know, of course as you know Nvidia's architecture is the most productive right? It generates the most tokens per watt. Mm-hmm. And so if the world's if the world's amount of watt power energy reduces, it shifts their choices to the most productive version. And the reason for that is that they still want to make their revenues.

### [51:47](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3107s) 50 GW的芯片，35 GW的电力
*50 GW of Chips, 35 GW of Power*

[52:43](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3163s) 所以我所有的CEO同行们，他们对明年的收入都有很高的期望，人们需要交付它，他们的股东指望他们交付，他们所有的合作伙伴都需要他们交付。所以当电力减少时，你投入其中的算力质量必须是最卓越的。我们是当今世界上你能投入的最高效的计算资源。我们是最高可租赁回报的，我们是可替代的，我们是耐用的。所以由于资源短缺，世界各地的力量正在把兴趣转向我们。我们的市场份额本来就已经很高了。但令所有人惊讶的是，我们不仅是一个市场份额很高的公司，我们的市场份额……所以你会看到我们从……而且因为我们每吉瓦提供的比以前多得多，我们曾经是一家GPU公司，然后我们成为最大、增长最快的网络公司。我们现在将成为最大的CPU公司之一，肯定是增长最快的那个。

> And so all of my CEO colleagues, they all have high expectations for revenues next year and people have, they need to deliver it, their shareholders counting on the deliver, all their partners need them to deliver it. And so when the amount of power reduces, the quality of the compute you put into it has to be the most extraordinary. We're the most productive computing resource you could put in the world today. We're the most rentable, we're fungible, and we're durable. And so the forces around the world because of the shortage of resources is shifting interest towards us. Our market share was already very high. But surprisingly to everybody, not only are we a large market share site, our market share And so you're going to see us go from, and because we offer so much more per gigawatt than we used to, we used to be a GPU company, then we became the largest and fastest growing networking company. We're now going to be one of the largest CPU companies companies, surely the fastest growing one.

[53:47](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3227s) 所以因为我们提供越来越多，我们做存储计算，我们做知识基础设施计算，所以数量……现在我们还做安全。所以我们在一个吉瓦中提供的产品，从我想大约每吉瓦180亿美元左右，到每吉瓦200多亿美元，到现在每吉瓦400多亿美元。我的感觉是这还会继续上升。所以我们不仅因为人们转向我们的计算平台而增长份额，我们还因为向世界的资本支出提供更多而增长份额。对世界来说，这是市场的美元份额。对。如果你看，就回到仅仅六年前，我们占世界资本支出的百分比还是个位数。

> And so because we're offering more and more, and we're in computing for storage, we're computing for the knowledge infrastructure, and so the amount of, and now we're in security Open shell and century. And so the offering that we offer into one gigawatt went from, I guess, about 18 billion or so per gigawatt to 20-some-odd billion per gigawatt to now 40-plus per gigawatt. And my sense is that that's going to keep on going up. And so not only are we growing share because people are shifting towards our computing platform, we're also growing share because we offer more to the world's CapEx. And to the world, it's a dollar share of the market. Yeah. If you look at, just go back in just six years ago, our percentage of the world's CapEx was in single digits.

[54:44](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3284s) 然后我认为现在在全球CSP中，我们大概占40%。但我们也在把市场扩展到CSP之外。在CSP之外，无论是本地部署还是通过云，我们一半的业务，我们的资本支出占比大概是，多少，80%？所以一半业务是40%，一半业务是80%，而且两者都在每年翻倍增长。所以你可以看到，我们占世界资本支出的百分比实际上还在增长。是的。嗯，Nvidia的买家之一是我们两年前谈过的人，Elon，当他开始建造Colossus的时候。两年前看到他在孟菲斯架起几十万块Hopper，我们大为震撼。我记得你说过没人能像他那样做。现在他又在筹集400亿美元，要买也许100多万块Vera Rubin。他已经从……走到了……

> Then I think right now with the world CSPs, we're probably running about 40%. But we're also expanding the market beyond the CSPs. And outside of the CSPs, whether it's on-prem or via clouds, half of our business, our percentage of CapEx is probably, what, 80%? And so between half of our business being 40%, half of our business being 80%, and both of us growing doubling every year. And so you could see that our percentage of the world's CapEx is, in fact, even growing. Yes. Well, one of those buyers of Nvidia is somebody we talked about two years ago, Elon, when he started building Colossus. And we were blown away a couple of years ago to see him stand up a couple hundred thousand hoppers in Memphis. And I remember you saying nobody can do it like him. Now he's raising another 40 billion, going to buy maybe million plus Vero Rubens. He's moved from the case to the end.

### [55:26](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3326s) Elon、Colossus与SpaceX
*Elon, Colossus & SpaceX*

[55:48](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3348s) 他说他想在2027年底前拥有10吉瓦。那是花在Vera Rubin上的数千亿美元。这是前所未有的规模。你昨天又和他在一起，当时他也被授予了一枚奖章。Elon今天在建造算力方面比那时强那么多吗？现在对你来说是否已经很明显？SpaceX将成为世界上最大的AI超大规模厂商是否已成定局？

> He says he wants to have 10 gigawatts by the end of 2027. That's hundreds of billions spent on Vera Rubin. It's an unprecedented rampant scale. You were with him again yesterday when he was also presented with a medal. Is Elon that much better at building compute today as he was then? Is it now obvious to you? Is it a foregone conclusion that SpaceX will be the largest AI hyperscaler in the world?

[56:18](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3378s) 嗯，我们过去在我们的图表的两部分中，Elon过去被归在企业业务下。我们就是这样报告数字的。现在他是我们CSP的一部分。当然。对对。当然。所以——他们规模巨大。所以我很高兴，我喜欢和Elon合作。他在用我们的技术做他自己的AI。训练模型之类的，当然还有部署模型。Grokbot超级酷，对Grokbot的需求，Grokbot的消耗真的在飙升。然后当然他把它租出去。所以这三件事的结合，他在这方面相当独特。我猜Google也这样做。那将是又一个例子。所以我认为他还有一个优势，就是他在计算和实体建设之间……Starship上周发射，进行了首次轨道交会，并发射了那些V3卫星，真的激发了人们的想象力，让Twitter一度瘫痪。

> Well, we used to, in the two parts of our chart, Elon used to be under enterprise. That's the way we reported the numbers. Now he's part of our CSPs. Of course. Yeah, yeah. Of course. And so- They're massive. So I'm delighted that I love working with Elon. He's using our technology for his own AI. in training models and such, deploying models, of course. Grokbot is super cool and the demand for Grokbot, the consumption for Grokbot is really skyrocketing. And then of course he rents it. And so the combination of these three things, he's fairly unique in doing that. I guess Google does that. That would be yet another example. And so I think he also has the advantage that he'd section between computing and the physical buildouts. Starship launched last week, had its first orbital rendezvous and was launching those V3 Sats that literally captured the imagination and broke Twitter for a moment.

### [57:20](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3440s) 太空中的数据中心
*Data Centers in Space*

[57:51](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3471s) 人们就是着迷地看着这些发射进入太空。我迫不及待想成为有效载荷。我是说，我个人。你是否越来越有信心，到我们称之为2028、2029年，我们将通过Starship在太空拥有相当数量的算力？我是说，Elon已经放出了一些大数字，但你认为我们能在那个时间框架内在太空实现实质性的推理吗？只用NVIDIA算力。让我告诉你为什么。这不仅仅是技术的问题。

> Just people mesmerized by watching these launch into space. I can't wait to be a payload. I mean, I'm personal. Do you have increasing confidence that we will have a material amount of compute in space via Starship by, let's call it, 2028, 2029? I mean, Elon's put some big numbers out there, but do you think that we can get to material inference in space in that timeframe? Only NVIDIA compute. Let me tell you why. It's not just about the technology.

[58:23](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3503s) NVIDIA的单位能耗性能是世界上最好的。我们称之为每瓦性能。最重要的是这一点，NVIDIA算力的生产率如此之好，意味着投资资本回报率如此之好，以至于你可以把它放到太空，而发射成本只占其中一小部分。发射成本并非无关紧要。所以你不会把低成本计算放进去。你会把高价值计算放在那里。所以我们是你能想象的最高价值的计算系统，计算史上最高的计算系统。所以我们现在实际上是发射到太空的完美选择，不管发射成本如何，它仍然会非常出色。那么到29年我们能在太空获得500兆瓦吗？嗯，我会让他来告诉你，但我认为有可能。你知道，Elon的……什么发射？

> NVIDIA's performance per unit energy is the world's best. We call it performance watt. The most important thing is this, the productivity of NVIDIA's compute is so good, meaning the return on investment capital was so good that you could put it out in space and the launch cost is a small percentage of it. The launch cost is not inconsequential. So you're not going to put low cost computing in there. You're going to put high value computing right there. And so we're the highest value computing system you could imagine, the highest computing system in the history of computing. And so we're the perfect fit really to launch into space right now irrespective of the launch cost It'll still be excellent. So do we get 500 megawatts in space by 29? Well, I'll let him tell you but but I think possible I You know, what's Elon's what launching?

[59:20](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3560s) 现在一年100次发射。我不确定具体是什么。我一天要跑很多地方，对，没错。所以，你可以把Vera Rubin放进去，它就能证明这笔投入是合理的。需求就在那里。它的投资回报率太好了。所以生产力非常出色。它是可替代的，意思是你在太空放一颗卫星。如果你没有客户来租用它，那就是社会性的，你当然不想要那样。所以，NVIDIA的算力将完全可替代，会有大量的租用者。我们与世界上每一家闭源模型公司合作。我们运行世界上每一个开源模型。我们有数千家初创公司与我们一起合作。还有各个行业的企业。所以算力是完全可替代的，而且是持久的。

> 100 launches a year right now. I don't know what exactly it is. I'm gonna go to a lot of all a day Yeah, right. And so, you could put Vera Rubens in there and it'll justify it. The demand is there. His return on investment is so good on it. And so the productivity is excellent. It's fungible, meaning you put a satellite on space. If you don't have a customer to go rent it, it's societal and you certainly don't want that. And so, NVIDIA Compute is going to be fully fungible and there'll be lots and lots of renters. We work with every single closed model company in the world. We run every single open model in the world. We have thousands of startups working with us. And enterprises around in every single industry. And so the computing is completely fungible and is durable.

[1:00:04](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3604s) 你会坐在上面，我们会跟上。原因是我们持续升级软件。我们是世界上唯一一家硬件公司，你八年前付钱给我们，我们至今仍在更新你的软件、新算法。对，你知道，所以只要卫星还在上面，我们就会更新软件。所以它会非常持久，所以我认为这三个特性的组合，是的，再加上技术，使它成为SpaceX的完美载荷。建设所有这些资本开支要花费数万亿美元。你们利用自己的资产负债表帮助像neo clouds这样的公司建设起来，为了让金融人士能够获得回报，这很大程度上取决于……他们需要看到回报，对吧？你谈到一些公司看到一年回本，对吧？第一年500亿美元，也许还有SpaceX的一些现货定价。但如果我看总的增量承购，今年所有实验室的收入，所以如果我把两个前沿实验室加起来，哦对，吃点零食吧，再加上一些推理实验室，大约是2000亿美元。

> You'll sit up there and we'll keep up. And the reason for that is because we keep coup to upgrading the software. We're the only hardware company in the world that you pay us eight years ago and we're still updating your software, new algorithms. Right, you know and so what will update the software for as long as the satellites up there And so so it's gonna be really durable and so I I think the combination of these three traits Yes, along with the technology makes it the perfect payload for for SpaceX It costs trillions of dollars to build all this capex You guys have leveraged your own balance sheet to help folks like the neo clouds build that out in order for finance people to get It depends a lot. They need to see payback, right? You talked about some companies seeing one year payback, right? $50 billion in the first year, some spot pricing out of SpaceX perhaps. But if I look at the total incremental offtake, the revenue of all the labs this year, so if I add up the two Frontier Labs, oh yeah, get yourself a snack, and we add up some of the inference labs, it's about $200 billion. dollars.

### [1:00:33](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3633s) AI收入能证明资本开支合理吗？
*Does AI Revenue Justify the Capex?*

[1:01:13](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3673s) 我们需要看到明年翻倍到4000亿美元。我们需要看到后年再翻倍到8000亿美元，才能接近数万亿美元资本开支的回报范围。今年我们有了Claude Code带来的巨大解锁，导致Anthropic的收入增长了10倍。你对实验室能否继续每年将收入增长2到3倍有多大信心，以便我们获得建设数万亿美元资本开支所需的回报？

> We need to see that double next year to $400 billion. We need to see it double the year after that to $800 billion in order to get in the zip code of the payback on trillions of dollars of CapEx. We had this massive unlock this year with Claude code that led to a 10x increase in Anthropics revenues. What's your confidence that the labs can continue to 2 to 3x their revenues every year in order that we get the payback that's going to be required to build out trillions of dollars at CapEx.

[1:01:50](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3710s) 你在投资这些公司，你在这些公司有大量投资，你认为这些公司会继续保持那种增长水平吗？哦，是的。人们误解了正在发生的事情。首先，特别是，NVIDIA占当今市场上AI基础设施收入的约80%。所以我对计算宇宙和NVIDIA计算的描述在这里很重要。

> You're investing in these companies, you have big investments in these companies, do you see that level of growth continuing in these companies? Oh, yeah. People misunderstand what's going on. First of all, particularly, Nvidia represents call it 80% of the revenue share of AI infrastructure in the market. today. And so my description of the universe of computing and video computing matters here.

[1:02:18](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3738s) 他们基本上把它称为三个细分市场，三个大细分市场。可能略多于三个，但让我就分成三大细分市场。我们计算的第一个细分市场就是基线计算，就是用于数据处理、分子动力学、物理模拟、聚变模拟、裂变模拟，你能想到的，量子化学、量子物理，等等一切。模拟和数据处理的世界，更不用说计算机图形。所有这些处理都在我们的GPU上完成，如果你说根本没有AI，只有经典机器学习、XGBoost、Visa、MasterCard的欺诈检测、我们的推荐系统来决定给你推荐什么新闻，所有这些加起来，经典计算，不是AI，不是深度学习，不是智能体系统。我要说这是我们业务的三分之一。

> They really basically call it three segments, three large segments. It's probably slightly more than that, but let me just kind of categorize in three large segments. The first segment of our computing is just baseline computing, just meaning use for data processing, molecular dynamics, physics simulation, fusion simulation, fission simulation, you name it, quantum chemistry, quantum physics, and so everything. The world of simulation and data processing, not to mention computer graphics. All of that processing done on our GPUs, if you just said there was no AI at all, just classical machine learning, XGBoost, that Visa, MasterCard for fraud detection, our recommender systems to figure out what news to recommend to you, all of that stuff combined, classical computing, not AI, not deep learning, not agentic systems. I'm going to call it a third of our business.

[1:03:14](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3794s) 而那块业务在增长，大约每年30%、40%。对。好的。即使没有任何AI实验室，NVIDIA也不是零，对吧？对。我们是一家相当大的公司。这就是为什么我把它排在第一位。第二部分是AI实验室，人们看到两家，但我们与每一家都合作。MetaMuse做得非常棒，对吧？而且它大量使用NVIDIA。你知道，去年或前年它们还是零，只有推荐系统，现在它们用MetaMuse大获成功。当然，Grok三年前还不存在。我们以前不运行Gemini，现在我们运行Gemini。从今年开始，我们开始与Anthropic合作。所以如果你只看我们的算力，人们看到几个实验室，但我看到更多，我看到五六个实验室。好的，所以这些实验室都做得很好，对，如果你告诉我那代表……不完全是三分之一，比如说我们算力的四分之一，那太好了。好，就说那是四分之一。所以当人们看NVIDIA时，我们现在运行世界的基础设施，而我说其中25%，也许30%是前沿实验室。

> And that business is growing, call it 30%, 40% per year. Yeah. Okay. In the absence of any AI labs, NVIDIA is not zero, right? Right. We're quite a large company. That's why I have number one. The second part is the AI labs, people see two, but we work with every one of them. And MetaMuse is doing fantastic, right? And it's vastly NVIDIA. And, you know, there were zero last year, or a year before that, there were just recommender systems and now they're, you know, killing it with MetaMuse. And of course, Grok didn't exist three years ago. We didn't run Gemini before, we run Gemini now. And starting this year, we started working with Anthropic. And so if you just kind of look at our compute people see a couple of labs, but I see more than that I see five six labs. Okay, and so so these labs are all doing great Correct and and if you if you told me that that represents Not quite one third of our business at let's say a quarter of our compute Fantastic. Okay, so let's say that's a quarter. So When people are looking at NVIDIA and we now run the infrastructure of the world, and I just said 25% of it, maybe 30% of it is the Frontier Labs.

[1:04:41](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3881s) 大约三分之一就是一般处理。另外三分之一，这是真正很棒的部分。因为我们正处于一个新计算平台转变的开端，大量新初创公司正在涌现。这么多伟大的公司。其中一些当然在做消费级的东西。有一家做得非常好，Instinct。Instinct，我正要提到它。Noah做得很好。还有各种物理AI、生物AI、材料科学AI。

> Call it a third is just processing in general. Another third, this is the part that's really great. Because we're at the beginning of a new computing platform shift, just a vast number of new startups are coming out. So many great companies. And some of them are, of course, doing consumer stuff. One of them really doing great. Instinct. Instinct, I was just going to mention that. Noah's doing great. To all kinds of physical AI, biological AI, material sciences AI.

[1:05:16](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3916s) Jeff Bezos，这家新公司Prometheus，他们做得非常棒。所以跨行业谱系已经创立的公司数量，数千家。它们做什么？它们租用NVIDIA计算机。所以它们是租用者。所以我们业务的三分之一算是下一代云计算。有点像今天的AWS，你知道？所以它不只是未来的AWS。这总共约占我们业务的三分之一。

> Jeff Bezos, this new company, Prometheus, they're doing fantastic. So the number of companies across the spectrum of industries that have been started, thousands of them. And what do they do? They rent NVIDIA computers. So they're renters. And so what is a third of our business is kind of cloud computing for the next generation. It's kind of like AWS for today, you know? And so it's just not AWS for the future. And that represents about a third of our business altogether.

[1:05:49](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3949s) 我的感觉是，我们正处于这种新计算架构建设的早期阶段。只认识几家公司并认为它们在为基础设施提供资金，这是一个非常大的错误。然后一个观察，开源模型，你知道我一直是开源模型的伟大倡导者。原因在于开源模型开辟了新机会。如果你看所有这些初创公司，80%使用开源模型。它们不是用它来与前沿实验室竞争。它们是用它在其他地方解锁新机会，你知道，可能是物理AI、可能是制造业、可能是医疗保健，它们在完全做不同的事情。对，它们采用这些开源模型，其中许多采用Nemotron，然后这些模型被定制成超级智能。所以如果你愿意，我有点认为通用的闭源实验室将实现通用人工智能，而且会变得超级好。

> My sense is that we're in the early innings of the build out of this new computing architecture. And it's a very, it's a real big mistake to recognize just a couple of companies and think that they're funding the infrastructure. And then one observation, open models, you know that I've been a great advocate of open models. And the reason for that is because open models opens up new opportunities. And if you look at all these startup companies, 80% of them use open models. And they're not using it to compete with the Frontier Labs. They're using it to unlock a new opportunity somewhere else, you know, it could be in physical AI could be manufacturing could be healthcare They're doing something completely completely different. Yeah, and and they're taking these open models and many of them taking Nemo Tron and The these models are then customized into super intelligence Yeah, and so if you will I kind of see that the the the general the general closed labs will be artificial general intelligence and it gets super good.

### [1:06:04](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=3964s) 开放模型与闭源模型
*Open vs. Closed Models*

[1:06:51](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4011s) 然后你有所有这些垂直公司，专注于特定领域，它们有自己的专有数据，教它做非常具体的事情，它们在其领域实现超级智能。所以你不觉得，你知道，所以再说一次，如果我们回到你的算力数学，我认为很多人用它来预测实验室的未来收入，例如，你说，嘿，如果你建立一个吉瓦的这种算力，你另一端可以得到400、500亿美元。但如果我们用之前谈到的数字，明年上线40吉瓦算力，如果你另一端取400亿美元，那么那个工厂另一端将产生1.6万亿美元的收入。但当然，有些东西，不是那样运作的。归根结底，你必须看需求曲线和扩散速度，来确定工厂另一端将产生多少收入。我听到很多的是人们很紧张，对吧？他们在这一年里看到，也许实验室收入略有放缓。他们紧张的是，你知道，某个时候会出现所谓的泡沫破裂，对吧？我们某一时刻上线的供应超过需求。AI更像是阶跃函数，而不是一条直线上涨的抛物线。我想问你的问题是，你是否担心，你知道，某个时刻所有人对为建设提供融资感到紧张，对吧？

> And then you have all these vertical companies that are focused on a particular domain and they have their proprietary data, they teach it to do something very specific and they achieve super intelligence for their domain. So you don't see, you know, so again, if we go to your compute math, which I think a lot of people are using to forecast the future revenues of the labs, for example, you know, you said, hey, if you stand up a gigawatt of this compute, You can get $ R 40, $ 50 billion out the other end. But if we took the number we talked about earlier, 40 gigawatts of compute coming online next year, and if you took $ R 40 billion out the other end, then that would be $ 1.6 trillion of revenues coming out the other end of that factory. But of course, there's something, it doesn't work that way. At the end of the day, you got to look at the demand curve and the rate of diffusion to determine the revenue that's going to come out the other end of the factory. And what I hear a lot is that people are nervous, right? They've seen over the course of the year, maybe the labs revenue is slowing down a little bit. And they're nervous that, you know, at some point you're going to have the proverbial bubble popping, right? Where we have more supply coming online than we have demanded a moment in time. That AI is more like the step functions rather than just this parabolic line straight up. I guess the question to you is, are you worried about, you know, a moment in time where everybody gets nervous about financing the build out, right?

[1:08:18](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4098s) 然后我们就无法完成建设。嗯，你知道，所有这些不同的事情，你只需要分解它。回到第一性原理。人们试图外推的部分，1000亿美元或1250亿美元的ARR，或者它们所在的位置，它们试图外推，那只是整体基础设施建设的很小一部分，不是很小的一部分，只是一部分。当然。所以问题是，人们会继续在线购物吗？对。

> And then we don't get the build out. Well, you know, all these different things, you just got to break it down. Break it back down to first principles. And the part that people are trying to extrapolate, the $100 billion of AR or $125 billion where they are, they're trying to extrapolate, It's a small part, not a small part, it's just a part of the overall infrastructure build out. For sure. And so the question is, are people going to continue doing shopping online? Right.

[1:08:50](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4130s) 所以那三分之一的建设，我们就叫它三分之一吧，你说过那是你业务的三分之一。这一点板上钉钉。对。这甚至没什么可争论的。所以三分之一的建设将会发生。我说过，几年前吧，是几年前吗，随着我们走出这个十年，基础设施的建设大概会达到每年3到4万亿美元。显然，它更接近4而不是更接近3。到本十年末，每年。如果你说其中1.5万亿只是算力。只是算力。那么，为什么它会这么大？嗯，在我开始之前，它是5000亿。

> So that third of the build out, let's call that a third, you said it's a third of your business. Take that to the bank. Yeah. That's not even a debate. So a third of the build out is going to occur. I said a few couple, was it a couple years ago that as we exit this decade, infrastructure structure build out will probably be somewhere between 3 to 4 trillion a year. Obviously, it's closer to 4 than it is closer to 3. Per year by the end of the decade. And if you said that 1.5 trillion of that was just compute. Just compute. Now, why is it going to be so big? Well, it was half a trillion before I started.

[1:09:31](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4171s) 如果我告诉你，在大约六、七、八年的时间里，它变成了1.5万亿，因为世界只是在消耗更多算力，这并不不合逻辑。除此之外，记住，摩尔定律已经结束了。曾经有一段时间，如果你想得到两倍的晶体管，你只需等一年。如果你想得到两倍的晶体管，再等一年，价格相同。但现在如果你想要两倍的晶体管，无论你等多久都没用，你必须付两倍的钱，因为摩尔定律已经结束了。所以无论世界想要多少计算能力，每一年，如果你想要两倍，你就要付两倍的钱。所以在那5000亿美元之上——它本来就在自然增长，因为世界只是能用上越来越多的数字化——你还得实际上给它加一个X因子，仅仅是为了反映摩尔定律已经结束这一事实。所以，没有AI，半导体行业本来就在成为世界最大行业的路上。我很久以前就提到过这一点，只是人们忘了，那是我说过的话之一。对，对，对。我两年前刚说过，对。

> And if I told you that over the course of some six, seven, eight years, it became one and a half trillion because the world's just consuming more compute, that's not illogical. On top of this, remember, Moore's law is over. There was a time if you want to get twice as many transistors, you just wait one year. And if you want twice as many transistors, get another wait one year, same price. But if you want twice as many transistors now, it doesn't matter how long you wait, You have to pay twice as much, because Moore's law is over. And so whatever computing capacity that the world wants, every single year, if you want twice as much, you're going to pay twice as much. And so on top of that half a trillion dollars that was kind of growing just naturally, because the world just could use more and more digital, you're going to have to effectively give it an X factor just to account for the fact that Moore's law is over. And so without AI, the semiconductor industry was on its path to become the largest industry in the world. And I've mentioned that long ago, just people forgot, which is one of those things that I said. Yes, yes, yes. I just said that two years ago, yeah.

[1:10:43](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4243s) 所以那只是一部分。对。所以我们唯一要做的，然后别忘了，因为我们正处于一个新的计算平台转变之中，将会出现大量初创公司。创新是经济的一部分。对，新公司，而且我们说过它在加速。对，你知道，所以如果我们花——如果你回到摩尔定律的通用计算，每吉瓦50亿美元。对，今天是500亿。对，对，Vera Rubin就是这样。所以问题是，Vera Rubin是否能为世界带来10倍的价值？所以关键结论是，我认为没有必要把任何一家实验室外推到数万亿美元。对，没必要那样做。对，我认为那是一个——我是说，这是一个重要的点，因为我认为人们——我们昨天刚看到，甚至在市场上，有一些议论，OpenAI的收入是700亿美元还是500亿美元？我想你和我——它们会是非常成功的公司。没错。对。但你看到很多基础设施公司回落，因为担心，哦天哪，OpenAI没有钱，将来也不会有钱来支付基础设施。我认为你要强调的一点是，当我们看这40吉瓦——明年要建设的大量基础设施。至少三分之一将只是用于通用计算替换，超大规模云厂商的基础东西。三分之一将属于NeoCloud、主权云、企业，所有其他使用它的各方，所有初创公司。还有三分之一将不是给两家实验室，而是给许多实验室，对吧？包括Grok和Thinking Machines以及其他所有。

> And so that's just one part. Right. And so the only thing that we, and then don't forget, because we're in a new computing platform shift, there are going to be a whole lot of startups. Innovation is part of the economy. Right new companies and we said it's accelerating Yeah, you know, so if we spent if you go back to Moore's law general purpose compute five billion dollars per gigawatt Yeah today, it's 50. Yeah, right for a Vera Rubin So the question is is Vera Rubin deliver 10x more value in the world So the punchline is I don't think it's necessary to extrapolate any of the labs into multiple trones of dollars Yeah, unnecessary to do that. Yeah, I think that's a I mean, it's an important one because I think people We just saw yesterday, even in the market, some chatter about, was OpenAI's revenue $70 billion or $50 billion? I think you and I- They're going to be insanely successful companies. Correct. Yeah. But you saw the pullback in a lot of the infrastructure companies on the fear that, oh my gosh, OpenAI doesn't have the money, won't have the money in order to pay for the infrastructure. I think the point that you're hammering home is that when we look at the 40 gigawatts Lots of infrastructure built out next year. At least a third of it is just going to write general purpose compute replacement, the basic stuff out of the hyperscalers. That a third of it's going to kind of NeoCloud, Sovereigns, Enterprises, everybody else who's using this, all the startups. And a third of it's going to not two labs, but to many labs, right? Including Grok and Thinking Machines and all the other.

### [1:11:35](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4295s) 两家实验室承受了太多压力
*Too Much Pressure on Two Labs*

[1:12:14](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4334s) 对，我们给两家实验室施加了太多压力。对。我很高兴与他们合作并支持他们。它们将成为历史上建造过的最具影响力的两家科技公司。它们会取得巨大的成功。但它们不必接近一万亿美元才算成功。我认为我们在这方面施加了太多压力。然后行业的其余部分，我们仍然必须建设几万亿美元的计算基础设施。只是为了继续前进，我们稍微换个话题。这个有多好，你们？来吧。顺便问一下，这些叫什么？这些是西梅干。西梅干，老式西梅干。台湾这家特别的店是我最喜欢的之一。有点像甘草味。有甘草，对。一点甘草、盐、糖、肉桂。对。有点酸劲。对，你尝到了。你知道吗？对。好。

> Yeah, we're putting way too much pressure on two labs. Yeah. I'm delighted to partner with them and support them. They're going to be two of the most consequential technology companies ever built in history. They're going to be magnificently successful. But they don't have to be anywhere near a trillion dollars to be successful. I think we're putting way too much pressure on that. And then the rest of the industry, we still do have to build out a few trillion dollars of computing infrastructure. Just to keep going, let's shift a little bit. How good is this, you guys? Come on. By the way, what are these called? These are dried plums. Dried plums, the old dried plums. This one particular store in Taiwan is one of my favorites. Kind of a licorice-y. There is licorice, yeah. A little licorice, salt, sugar, cinnamon. Yeah. Has a little twang to it. Yeah, you got it. You know? Yeah. Okay.

[1:13:06](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4386s) 你说到明年，对NVIDIA芯片的需求远超100%。你给出了70%的收入增长指引，你给了一个提前一年的展望，因为你说供应受限。当时我只有这些。没错。我有18个月来运作。你说你非常有信心。我每天都会来上班。你说非常有信心，有充足的时间与供应链合作做得更好，你期望做得更好，然后不让客户失望。那么事情是不是进展得更好？订单簿是否继续增长？因为另一部分是，对，你在某个时刻拍了快照，当然，人们还在继续下单。所以你认为供应链是否有足够的余量让你挤出更多？随着更多这些能力被解锁，订单簿是否继续增长？

> You said as to next year, demand for NVIDIA's chips well over 100%. You guided to 70% revenue growth, you gave a year forward look because you said supply is constrained. That's all I had at the time. That's right. I got 18 months to work. You said you were very confident. I'll come to work every day. You said very confident and plenty of time to work with the supply chain to do even better, you expected to do even better and then not disappoint customers. So are things going even better? And has the order book continue to grow? Because the other part of that is, right, you took a snapshot at a moment time, and of course, people are continuing to place orders. So do you think there's enough slack in the supply chain you can squeeze more out? And is the order book continuing to grow as more of these capabilities are unlocked?

### [1:13:26](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4406s) AI供需动态
*AI Supply and Demand Dynamics*

[1:14:01](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4441s) 目前，我告诉你我为什么告诉所有人。我们面临的挑战是我们有巨大的订单簿。每一家CSP、每一家NCP、每一家OEM都已经向我们下了订单。初创公司、大公司、小公司，他们都已经向我们下了订单。原因是我们与他们一起做这件事，取决于客户的规模，最大的客户至少提前一年，两年的预测。所以我们已经有了一个非常大的订单簿。因为我在进行分配，而我分配给所有客户的远少于他们想要的。他们想要的。如果我不公开说出我们有什么，每个人都会猜疑，他们是不是被少给了，他们是不是被，你知道，所以这是我本质上处理每一次让我失望的人的对话的方式。

> For the moment, I'll tell you why I told everybody. The challenge that we have is we have a giant order books. Every single CSP, every single NCP, every single OEM had already placed orders on us. Startups, big companies, small companies, they already placed orders on us. And the reason for that is because we work with them on it, depending on the size of the customer, at least a year for the largest ones, two years forecast. And so we already have a very large order book. And because I was in the process of allocating, and what I was allocating to all of our customers is a lot less than they wanted. They wanted. If I didn't publicly say what we had, everybody would be second guessing, are they short-changed, are they being, you know, and so it was my way of essentially handling every single conversation where I'm disappointing people.

[1:15:05](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4505s) 所以几乎每个客户，如果你能交付更多，他们都会从你这里买更多。远远如此。然后从那时起，当然，我们几乎每周和每天都有新订单。所以此外，然而，我们正在获得更多供应。限制我们的一些供应是土地、电力、厂房，下游。所以几周前，我们在澳大利亚宣布了2吉瓦。所以那是一件非常大的事。2吉瓦对我们来说就像800亿美元。

> So pretty much every customer would buy more from you if you could deliver more. By far. And then since then, of course, we have new orders almost weekly and daily. And so in addition, however, we're getting more supply. And some of the supply that is limiting us is land power shout downstream. And so a couple of weeks ago, we announced two gigawatts in Australia. And so that's a very big deal. 2 gigawatts to us is like $80 billion.

[1:15:34](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4534s) 然后我希望当我去欧洲参加GTC时，我们能在那里宣布一些新的合作伙伴关系，并解锁更多的土地、电力、厂房。所以这个土地、电力、厂房显然是障碍之一。而这是NVIDIA拥有巨大优势的领域。因为基于我们构建的东西，如果有一家NeoCloud，比如CoreWeave或EBS，他们想扩展业务，NVIDIA拥有整个全栈完整的基础设施。

> And then I'm hoping when I go to Europe with GTC, we'll be able to announce some new partnerships there and unlock more land power shell. So this land power shell is clearly one of the obstacles. And this is an area where NVIDIA has a huge advantage. Because based on what we build, if there's a Neo Cloud and this Neo Cloud like Core Weave or EBS, and they would like to scale out their business, NVIDIA has the entire full stack complete infrastructure.

[1:16:10](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4570s) 所以我们可以支持他们扩展。所以我们可以在别人无法建设的地方建设，与别人无法合作的伙伴和客户合作。所以我认为这给了我们更大的韧性，但尽管如此，随着我们继续在上游和下游增加供应，关键领域之一是基板。基板变得更大，芯片也更多。所以这在两方面对他们不利。每一颗出货的芯片都使用更多的材料。

> And so we could support them as their scale out. And so we could build in places and we can build with partners and customers that nobody else can. And so I think this gives us a lot more resilience, but nonetheless, as we continue to increase supply upstream and downstream, one of the critical areas is substrate. Substrates are getting larger, and there are more chips. And so it's working against them in two ways. Every single chip that's being shipped uses more of the material.

[1:17:09](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4629s) 所以那就是供应约束。这些只是一些例子。硅光子学是供应约束。所以一切都是供应受限。所以我们会继续努力，需要几个月时间来做。你知道，每过一周，我们都会增加，但尽管如此，我们不断增加更多供应，订单簿也不断上升。所以几年内很难弥合这个差距。看看这张图。

> the So that's supply constraint. And those are just some examples. Silicon photonics is supply constraint. So everything is supply R And so we're gonna keep working on it a few months to do so. You know, every single week that goes by, we increase, but nonetheless, we keep increasing more supply and order books keeps going up. And so it's gonna be hard to close that gap for a few years. Take a look at this chart.

### [1:17:37](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4657s) 华尔街共识回顾
*Review of Wall Street Consensus*

[1:17:39](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4659s) 你知道，这就像是一场庭外取证。不，没错。好。不好意思。通常流程是这样的：黄先生，请，请现在，请现在看这里。庭外取证。好，那么这条灰线，这是2025年初的共识。这是2025年初对NVIDIA的共识。蓝线是Altimeter当时很早的预测，所以我们比共识高得多。绿线是现在基于你们指引的新共识所在的位置。让我震惊的一件事是，那25个拿钱专门预测这家公司会怎么走的人。他们所有想法的共识，对吧，偏差了百分之几百，对吧，就在20……不，是去年这个时候。去年这个时候，甚至不是25。对。那是去年这个时候的共识。好，一年前。上周。好，所以一年之内，对这家公司的共识预测上移了100%到200%。嗯，太惊人了。在我20多年的从业经历里，你很少看到共识偏差这么大。然而股价只涨了20%。好，你真的把这家公司的营收预测和自由现金流预测翻了一倍。你现在按13到14倍市盈率交易，取决于你看谁的数据。那么你认为市场在就Nvidia说什么，或者说市场对预期的信心在说什么，对吧？

> You know, this is like a deposition. No, that's right. Okay. Excuse me. It usually goes like this, Mr. Huang, please, please now, please now look at it. The deposition. Okay, so the gray line, this was the consensus at the start of 2025. This was the consensus for NVIDIA at the start of 2025. The blue line was altimeters forecast way back then, so we were a lot higher than the consensus. The green line is now where the new consensus is based upon your guidance. One of the things that blows my mind about this is the 25 people who get paid to just forecast what this company is going to do. Their consensus of all their ideas, right, was off by a couple hundred percent, right, just at the beginning of 20 No, it's time last year. At this time last year not even 25. Yes. That was the consensus at this time last year. Okay, one year ago. Last week. Okay, so in a year The consensus forecast has shifted up by a hundred to two hundred percent for this company. Mm-hmm extraordinary. Like in my 20 plus years of doing this, you don't often see consensus being off by that much. And yet the stock's only up 20%. Okay, you literally doubled what the revenue forecast was and the free cash flow forecast was for the company. You're now trading at 13 to 14 times earnings, depending upon who you're looking at. So what do you think the market is saying about Nvidia or about the confidence in the expectations of the market, right?

[1:19:12](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4752s) 才会造成这种情况。因为通常如果你看到它上移那么多，股价会上移得多得多。首先，一家像我们这么大的公司还能提升份额，这本身就说明了我们优势的韧性、规模和体量。这个优势是在扩大还是在缩小？你的护城河是在扩大还是在缩小？在扩大。在扩大。你的竞争护城河在规模上还在扩大。对，我举个例子。我们现在就做个思想实验。如果你现在必须递给某人一张1000亿美元的支票，去下下一代产品的采购订单，如果你想排上Vera Rubin Ultra的队，现在差不多是下单的时候了。

> To cause that to be the case. Because usually if you see it shift up by that much, the stock would have shifted up by a lot more. First of all, for a company to be as large as us gaining share tells you something about the resilience and the size and scale of our advantage. Is that advantage increasing or decreasing? Is your moat increasing or decreasing? Increasing. Increasing. Your competitive moat is increasing at scale. Yeah, let me give you an example. Let's just do a thought experiment right now. If you had to hand somebody $100 billion check to place a purchase order right now for a next generation, if you want to get in line for Verre Ruben Ultra, it's close to placing order time.

### [1:19:40](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4780s) Vera Rubin Ultra订单
*Vera Rubin Ultra Orders*

[1:19:59](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4799s) 见都没见过。对吧？但他们有信心。所以假设你得把1000亿美元押在某人身上。第一，你怎么知道它会好？但因为NVIDIA造它，你就知道它会比Vera Rubin更好。而Vera Rubin很棒。Vera Rubin之所以、所有人之所以对Vera Rubin有信心，是因为Blackwell很棒。而在计算世界里，这些计算机跑的工作负载太复杂了。你不能只靠测一个基准。对，对。你知道，那是500亿美元。对。而要交付这样的性能，需要汇聚的技术数量多得离谱，你知道吗？所以问题是，见都没见过，你能把500亿美元的订单押在谁身上？对。我们真的是唯一一个。当年下采购订单，一张10亿美元的订单，那已经够难了。但如果你下的是1亿美元的订单，而系统就摆在你面前，你直接测芯片就行了。

> Side unseen. Yeah. Right? But they have the confidence. And so let's say you got to place $100 billion on somebody. One, how do you know it's going to be any good? But because NVIDIA builds it, you know it's going to be better than Vera Rubin. And Vera Rubin's awesome. And the reason why Vera Rubin, the reason why everybody had confidence in Vera Rubin is because Black Hole is awesome. And the world of computing, the workloads that these computers do are so complex. You can't just have like measure one benchmark. Right, right. You know, that's $50 billion. Yes. And the number of technologies that have to come together to deliver the performance is insane, you know? And so the question is, sight unseen, who can you place a $50 billion order on? Yeah. We're literally the only one. Now, back when it was placed a P.O., a billion-dollar P.O., that was challenging enough. But if you had placed $100 million P.O. and the system is sitting right in front of you, you could just benchmark the chips.

[1:21:04](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4864s) 你怎么测这些系统？光是把它建起来就是个奇迹。你知道，所以如果有人想做评估，我很乐意支持，但你得给我一个巨大的数据中心，我们得花几年时间去改造它，你知道吗？所以这些玩意你现在根本没法测。有意思。所以这跟台积电的情况一模一样。我们把100%的业务给台积电的原因，是因为他们是世界上唯一一家我能提前五年规划的公司。而且在见都没见过的情况下，知道那个晶体管会很好，因为它比那个好，比那个好。

> How do you benchmark these systems? Just building it is a miracle. You know, so if somebody wants to do an evaluation, I'm happy to support it, but you're going to need to give me a giant data center and we're going to need a couple of years to retrofit it, you know? And so you can't even benchmark these things anymore. Interesting. And so it's exactly what happened with TSMC. The reason why we give 100% of our business to TSMC is because they're the only company in the world that I can plan five years out. And inside Unseen, know that that transistor is going to be really good because it's better than that one, better than that one.

[1:21:40](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4900s) 而且他们会有产能给我，他们知道怎么跟我的业务团队合作，他们有敏捷性来应对我所有的起伏。而且，你知道，比如应对我从灰线到蓝线到绿线这个事实。没错。好。他们必须足够大，才能帮我从灰线走到蓝线再到绿线。世界上没有另一家公司能帮我做到这一点。没错。而且要拥有所有这些不同的工艺技术，从硅光到逻辑到存储，推、拉、推、拉、封装。我们做的事情数量如此动态。那么你怎么可能用一家只有他们十分之一规模的公司做到这些？不可能。所以我认为核心观点是：因为我们在为这个星球建设基础设施，而NVIDIA已经成为一家基础设施公司。而且因为这些基础设施花费巨大，它们在动，技术在飞速变化。对。

> And that they will have capacity for me and that they know how to work with my business team, that they have the agility to handle all of my ups and downs. And, you know, for example, to deal with the fact that I went from gray to blue to green. Exactly. Okay. They have to be big enough to help me go from gray to blue to green. There's no other company in the world that can help me do that. That's right. And to have all these different process technologies from silicon photonics to logic to memory to push, pull, push, pull, packaging. The number of things that we do is so dynamic. And so how do you possibly do that with a company that is one-tenth their size? It's impossible. And so I think the big idea is this, because we're building infrastructure for the planet and NVIDIA has become an infrastructure company. And because these infrastructures cost so much money and they're moving and technology is moving so fast. Yeah.

[1:22:37](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=4957s) 我们几乎是信任的结合体，事实上我们跑一切。所以一开始是一些GPT模型，但看看被发明的flash attention机制，看看我们把模型并行化、跨NVLink 72分片的方式，以及这一切如何改变，MoE改变了。现在人们在谈论transformer扩散模型是混合的、SSM，它们是transformer、混合体，所有这些不同的技术都是，对吧？这一切的演进，我们知道它全都跑在CUDA上。它是唯一……我们不知道你是需要做更多预训练、后训练，还是只需要智能体推理。嗯，结果是我们反正全都跑。所以你可以建一个500亿美元的数据中心，并且知道工作的可替代性，如果发生了什么事，不管什么未知原因，你想，如果你愿意，把一半产能转租给别人，NVIDIA是完全可替代的。所以我们在工作负载上可替代，在模型上可替代，在客户之间可替代。

> We're practically the combination of the trust, the fact that we run everything. So it started out with some GPT models, but look at the flash attention mechanisms that were invented, the way that we paralyzed the models and sharded across NVLink 72 and how all of that changed, MOE changed. And now people are talking about transformer diffusion models that are hybrid, SSMs that They're transformers, hybrids, and all these different technologies are all, right? The evolution of all that, we know that it all runs on CUDA. It is the only... We don't know whether you need to do more pre-training, post-training, or just agentic inference. Well, it turns out we run it all anyways. And so you could build a $50 billion data center and know that the fungibility of the work, and if something were to happen, for whatever reason unknown, you wanted to, if you will, sublease half of your capacity to somebody, NVIDIA is completely fungible. So we're fungible in workload, we're fungible in models, we're fungible across customers.

[1:23:42](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5022s) 而且我们可以在世界上任何地方部署。你在欧洲开公司，你在美国开公司，但猜怎么着？你想要你的客户，你想服务欧洲的客户，你想服务日本的客户。嗯，结果是我们已经在那儿了。NVIDIA的基础设施已经在那儿了。所以，OpenAI想去欧洲，没问题……我们眨眼间就能把他们带过去。如果他们眨眼间想去印度，如果他们想，对吧？如果Anthropic想眨眼间去日本、韩国，我们已经在那儿了。所有这些架构都已经在那儿了。所以我认为，如果你把NVIDIA放进一个洞里，然后你回来说，好，我想做个思想实验。我想把1000亿美元押在某人身上。

> And we're deployable everywhere in the world. You start a company in Europe, you start a company in United States, but guess what? You would like your customer, you want to serve customers in Europe, you want to serve customers in Japan. Well, it turns out we're already there. NVIDIA's infrastructure is already there. So, OpenAI wants to go to Europe, nothing... We'll take them there in the heartbeat. If they want to go to India in a heartbeat, if they want, isn't it right? If anthropic wants to go to Japan, Korea in a heartbeat, we're already there. All of these architectures are already there. And so I think if you take NVIDIA in a hole and then you come back and you say, okay, I want to do the thought experiment. I want to place $100 billion on somebody.

### [1:23:49](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5029s) 全球布局：欧洲、日本及其他地区
*Global Reach: Europe, Japan & Beyond*

[1:24:22](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5062s) 你怎么做？一年前，我们坐在这里，共识预测是你在2027年会持平或萎缩。现在，共识预测是你会增长74%。所以，我只想让大家消化一下，因为很多人跟着这些共识预测走，而它错得离谱。他们在预测。如果他们回去听你上次的播客。对。你告诉过他们。我解释过。听着，女士们先生们，我确实解释过。所以我看2029年。当时没有ROI的问题。然后是定制与通用之争。NVIDIA通用之所以如此，是因为我能拥有所有这些可替代性。可替代性的好处是低风险。

> How do you do it? A year ago, we were sitting here and the consensus forecast was that you were going be flat or shrink in 2027. Now, the consensus forecast is you're going to grow 74%. So, I just want to let people let that sit in because a lot of people follow these consensus forecasts, and it was really wrong. They were forecasting. If they go back and listen to your podcast from last time. Right. You told them. I explained it. Listen, ladies and gentlemen, I did explain And so I look out at 2029. There was wasn't an ROI question. And then there was this custom versus general purpose. The reason why NVIDIA's general purpose is so I can have all this fungibility. The benefit of fungibility is low risk.

[1:25:15](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5115s) 而低风险在你押大钱的时候才真正发挥作用。当你建设500亿、1000亿美元的基础设施时，你需要知道，无论发生什么，那套基础设施在很长一段时间里都会很好。所以专业化是可替代性的敌人。专业化是耐久性的敌人。只要我能证明我们的性能非凡，而我们每天都在证明这一点。一旦我们能证明这一点，那么可替代性和耐久性就转化为可租赁性，进而转化为生产力。以及可融资性。因此，可融资性，进而带来最低的TCO。对，对，没错。对。嗯，现在你们在回购史上最大规模的股票。按我的计算，你们大概能在未来两年内回购公司10%以上的股份。

> And low risk really comes in to play when you're placing big dollars. When you're building out 50, $100 billion worth of infrastructure, you need to know that that infrastructure is going to be really good for a long time, no matter what. So specialization is the enemy of fungibility. Specialization is the enemy of durability. And so long as I can prove that our performance is extraordinary, and we prove that every single day. The moment we can prove that, then fungibility and durability translates to rentability, which translates to productivity. And finance-ability. And therefore, finance-ability, which leads to the lowest TCO. Right, right, exactly. Yeah. Well, now you're buying back the largest amount in history. And on my math, you could probably buy upwards of 10% of the company back within the next two years.

[1:26:14](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5174s) 在高盛，你把Nvidia称为第一家成长价值型公司。听起来很像苹果那套追求更多利润的说辞。哦，是吗？好。是吗？对。或者更多。对。他们能返还资本。它奏效了。对。当然。它奏效了。它从当时的30美元涨到300美元。对。300美元。在很多方面，我们正在经历类似的转型。比如，我忘了提的一件事是，Brad，我想你正要问我那个。

> At Goldman, you called Nvidia, the first growth value company. Sounds a lot like Apple's pitch to more profit. Oh, is that right? Okay. Is that right? Yeah. Or more. Yeah. That they could return the capital. It worked. Yeah. For sure. It worked. It went from at the time $30 to $300. Yeah. 300 bucks. In a lot of ways, we're going through a similar transition. For example, the thing I forgot to mentioned is, and Brad, I think you were about to ask me about that.

[1:26:46](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5206s) 我们的大部分投资和融资合作策略，是在尝试建立两种商业模式，两种新的商业模式。其中一种我们已经研究了一段时间，并将在不久的将来加速推进，那就是软件许可。你知道，我们公司拥有的软件体系非常庞大。而且那个数据中心如此之大，想象一下其中运行的所有软件。我们提供整个技术栈。而且，我们的整个技术栈完全是工业级的。对于云服务提供商（CSP），他们想运行自己的技术栈。因此，我们以许可方式提供，将我们的软件开源给他们，然后他们拿去修改并集成到自己的系统中，再运营他们的服务。但对于世界上许多公司来说，他们没有IT部门，没有软件部门来做这件事，所以他们需要我来维护。因为过去几年AI主要聚焦于训练，但现在他们正转向部署，他们需要像运营工厂一样运营这些。他们需要知道由我负责，我承担SLA来支持和维护，保持其安全可靠，只要我们还活着。面向企业业务的软件许可将会非常、非常庞大。所以，你知道，如果我们的软件许可，粗略简单算一下，每GPU每年几千美元，而我们每年出货数百万GPU，你知道，每年2000万GPU，还在增长。

> Most of our investment and financing partnership strategies is trying to set up two business models, two new business models. One business model we've been working on for some time and will be accelerated in the near future is software licensing. You know, the body of software that we have in our companies is extraordinary. And that data center is so large, imagine all the software that runs in it. And we provide the entire stack. Well, and our entire stack is completely industrial strength. For the CSPs, they want to run their own stack. And so we license it where we give them our software open source, and then they take it and they modify it and they integrate it into theirs, and then they operate their services. But for a lot of the world's companies, they don't have the IT department, they don't the software department do it, so they need me to maintain it. Because AI in the last several years was largely focused on training, but now they're moving into deployment, they need to operate these as factories. They need to know that I'm on the hook, I'm on the SLA to support it and maintain it, keep it secure, keep it safe for as long as we shall live. Software license for our enterprise business will be very, very large. And so, you know, if our software licenses, you know, just roughly say for simple math, a couple thousand dollars per GPU per year, and we have millions of GPUs, we ship, you know, 20 million GPUs a year, I'm growing.

[1:28:19](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5299s) 所以其中一半进入全球企业市场，加起来。对。好。所以我相当确定，这将是我们业务中软件许可的很大一部分。然后，当然，有些云依赖 NVIDIA 提供我们的全栈。他们依赖我们的承购、我们的生态。他们也依赖我们，也许作为投资伙伴，帮助他们确保融资和资本。他们已经搞定了土地、电力、外壳，他们是区域冠军。他们有各种东西在为他们所用。但这三个领域，他们需要我们的支持。所以在那种关系里，我们很可能会转向收入分成。所以我们会做投资，我们会获得前期销售的好处，但我们也会获得未来收入分成的好处。所以我们不是云服务提供商，但我们会为他们提供承购，提供软件支持，为期四、五、十年。所以我们应该从中获得一些经济回报。所以收入分成会是其中很大一部分。就像，绝对。我刚刚描述了公司两个独立于硬件销售的新增长向量，而且它更多与我们平台的装机量相关。如果你没有装机量，软件许可的好处是什么？

> And so half of it goes into world's enterprise, it adds up. Yeah. Okay. And so I'm fairly certain this is going to be a very big part of our business software license. And then, of course, there are some clouds that rely on NVIDIA for our full stack. They rely on us for our offtake, our ecosystem. They also rely on us for maybe as an investment partner and helping them secure their financing and capital. They've secured Land Power Shell, they're a regional champion. They got all kinds of stuff working in their advantage. But these three areas, they need our support on. And so in those kind of relationships, it's very likely that we'll move to refshare. And so we'll make an investment, we'll get the benefit of the sale upfront, but we'll also have the benefit of refshare going forward. And so we're not a cloud service provider, but we're going to be providing offtake and providing software support for them for four, five, 10 years. And so we ought to get some economic return on that. And so refshare is going to be a good part of that. Like, absolutely. I just described two new growth vectors independent of the hardware sale of our company and it's much more related to the install base of our platform. What's the benefit of software licensing if you have no install base?

[1:29:36](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5376s) 如果没有承购，怎么会有收入分成？所以装机量要素、我们软件栈的丰富性、我们承购的丰富性以及我们在全球的生态系统，将推动这一点。其中一个方式，我们谈到以13倍交易，当Tim和Luca去见Warren Buffett时，Buffett说过他不喜欢科技公司的原因之一是他们有这种命中，他们产生大量自由现金流，而不是返还给股东，他们拿去花在别的东西上。而且通常，那个别的东西不如最初产生现金流的那个东西好。你们现在将超过50%的自由现金流用于回购股票，有效地返还给股东，同时你们还有指数级增长，正如我们已经讨论过的，我们说50，对吧？我认为上个季度我们做了60多。正确。是的，我们会增加。你知道，我们投资于真正高质量的货币，你们知道，过去几年，NVIDIA从谁那里买，Jensen从谁那里买，你就得投资那家公司。现在街上的新说法是，Jensen投资谁，你就得投资那家公司。这是真的。原因在于，像你一样，我们身处生态系统中，我们处于信息流中。

> And how could you have refshare if there's no offtake? And so the install base element, the richness of our software stack, the richness of our offtake and the ecosystem that we have around the world, that's going to fuel that. One of the ways, so we're talking about this trading at 13 times, when Tim and Luca went to see Warren Buffett, and one of the reasons that Buffett has said he doesn't like technology companies is they have this hit, they produce an enormous amount of free cash flow, and rather than ever giving any of it back to the shareholders, they go and spend it on something else. And oftentimes, the something else isn't as good as the thing that caused them to generate the cash flow in the first place. You guys are now spending over 50% of your free cash flow buying back your stock effectively returning it You know to shareholders while you have this exponential growth at the same time as we've already discussed we said 50, right? I think this last quarter we did 60 some on correct. Yeah, we're we're gonna increase You know, we invest in really high quality currencies you guys know and and and you know the last couple of years, whoever NVIDIA buys from, whoever Jensen buys from, you gotta invest in that company. And now the new word on the street is whoever Jensen invests in, you gotta invest in that company. And it's true. And the reason for that is because, you know, like you, we're in the ecosystem, we're in the flow of information.

[1:30:59](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5459s) 我们知道谁在做。我们一起做了很多这样的投资。是的，我们认识很多聪明人，当他们想做伟大的事情时，我们想支持这些人。所以我们算是内部人士。我们有很多信息。因此我们的投资是信息充分的。更不用说，NVIDIA也在生态系统中，所以我们可以帮助他们成功。所以投资成功有很多不同的原因。但作为我们总现金生成的比例，它将迅速下降，部分原因是OpenAI和Anthropic不会再次出现。坦率地说，那个机会不会再来。其次，我们的现金流生成增长相当快。

> We know who's doing it. We do a lot of those investments together. Yeah, and we know a lot of smart people, and we wanna back these people when they wanna do something great. And so we're kind of on the inside. We have a lot of information. And so our investments are well-informed. And not to mention, NVIDIA is also in the ecosystem, and so we can help them succeed. And so there's a lot of different reasons why the investments turn out great. But as a percentage of our total cash generation, it's going to decline very rapidly, partly because open AI and anthropic isn't going to come around again. That opportunity is not going to come around again, frankly. And then secondarily, our cash flow generation is growing quite fast.

[1:31:44](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5504s) 就此而言，在我看来，如果你回购那么多，而且收入的持久性，正如你今天在这里描述的，比人们预期的更长。我认为这是人们仍然过于依赖的一点。这只会持续这么长时间。Ezra前几天在播客上问过你，泡沫破裂还有多久？我想你说了大约两到三年，你有很好的可见性，需求在那里，之后谁知道呢。我听到有人说，不，可能只有两年就会破裂。但当我看到这一点时，这从根本上是对所有全球算力的永久性重新架构。

> And to that point, it seems to me that if you're buying back that much, and the durability of the revenues, I think, as you've described here today, is longer than people expect. And I think that's the one place people are still hanging too much on. This is only going to last as long. Ezra asked you this on the pod the other day, how long until the bubble burst? And I think you said something about two to three years you have good visibility that the demand is there and who knows kind of after that. And I heard some people say, no, it may only be two years before this burst. But when I look at this, this is fundamentally a re-architecture of all global compute forever.

[1:32:24](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5544s) 所以除非出现全球萧条，否则每一年，人们都将消耗大量的NVIDIA算力。是的，所以这样想，我们运行在，CSP运行大约半万亿美元，整个行业，全世界每年运行大约1万亿美元的IT建设。每年建得更多，大约每五年，一些会折旧或关闭下线等等，一些持续更久。所以旧世界是检索式的计算方式，新世界是新的计算方式。你知道，我预测未来十年，三、四，比方说四，比方说五，我的感觉是我们将继续从那里增长。如果你告诉我，年产能建设、基础设施建设、更新，新旧结合，更新，大约10万亿美元，我一点也不惊讶。

> So unless there's a global depression, then every single year, people are going to have to consume an enormous amount of NVIDIA compute. Yeah, and so the way to think about that is we were running in the, the CSPs were running about half a trillion, the whole industry, the whole world is running, call it a trillion dollars of IT build out each year. And every year, more gets built and every five years or so, some gets depreciated or turned off and taken offline and so on and so forth, and some of it lasts longer. And so that old world was in the retrieval way of doing computation, in the new way of doing computation. You know, I've predicted that going out of the decades, so three, four, let's say four, let's say five, and my sense is that we're going to continue to grow from there. If you told me that the annual capacity building, infrastructure building, refreshing, combination of new and old, refreshes, runs around $10 trillion, I'm not even a little bit surprised.

[1:33:35](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5615s) 再次，对听众来说，你认为到本十年末每年4万亿美元的支出，因为每次你消费token，每次你使用智能，你都必须运行机器。工厂必须运行。我想你已经描述了这一点。那大约是50%。如果你假设全球GDP中有50万亿美元与知识工作相关，如果我们在五年内获得20%的生产力提升，就像你在NVIDIA内部看到的，你在NVIDIA内部看到了更大的生产力提升。但如果你看到那种生产力提升，那么如果到2029、2030年我们获得10万亿美元的增量经济活动，那似乎非常合理。你只是说，我们将其中一半用于资本支出，用于生产智能。所以也许这是我们稍微下坡的好地方。你非常慷慨。

> Just again, for people listening, you think $4 trillion of annual spend by the end of the decade because every time you consume tokens, every time you use intelligence, you have to run the machine. The factory has to run. And I think you've described that. That's about 50%. If you assume that $50 trillion of global GDP is related to knowledge work, if we get a 20% productivity gain over five years on that, like you've seen inside Nvidia, you've seen much bigger productivity gains inside Nvidia. But if you see that sort of productivity gain, then if we get $10 trillion of incremental economic activity by 2029, 2030, that seems very reasonable. And you're just saying, we'll spend half of that on CapEx, on the intelligence to produce it. And so maybe that's a good place for us to take a bit of an off ramp. You've been incredibly generous.

[1:34:32](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5672s) 我可以一直说下去，但我想回到我们开始的地方。当我们回顾——播客和吃播。我们可以继续。我们可以继续。但不久之前，人类状况还被描述为肮脏、野蛮和短暂。对吧？那是上周吗？创新确实推动了所有人类进步，主要是在过去200年。

> I could go on forever, but I kind of want to wrap where we started. When we look back- Podcasts and mukbang. We could keep going. We could keep going. But it wasn't so long ago that the human condition was described as nasty, brutish, and short. Right? Was that last week? Innovation has really driven all human progress, mostly in the last 200 years.

[1:35:05](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5705s) 更多时间与家人在一起，更长的寿命，更少时间做人类真的不想做的苦差事，更安全。我们是第一代认为63岁还年轻的人。当人们还说，你知道，人们叫我年轻人。更安全、更好的生活。尽管你和我听到外面有所有这些担忧，顺便说一句，同样的担忧在工业革命开始时也被表达过。

> More time with family, longer lives, less time doing drudgery that humans really didn't want to do safer. We're the first generation where 63 years old consider young. When people still got, you know, people call me young man. Safer, better lives. All of this despite all of these concerns that you and I hear out there, and by the way, those same concerns were expressed at the beginning of the Industrial Revolution.

[1:35:37](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5737s) 从现在起五年后，我们是否已经做出了正确的决定、世界是否会变得比现在好得多——也就是我们所描述的那个样子，因为有了这些投资和我们所做的一切——你对这一点的信心有多大？相比12个月前，你今天坐在这里是不是更有信心了？我有99.99%的信心——只是为了致敬一位好朋友，才不用0和100。我有99%……说伪证罪有点过了，99%的信心，相信我们中有太多人如此深切地在意，以至于结果会变好。相信我们会倾尽全力、倾尽智慧、倾尽资源，确保我们负责任且乐观地走到那一步。所以，有那么多人在做正确的事。当然，很多时候，有些人会对这些事情高谈阔论，他们的判断受若干因素影响，却没有意识到，在他们的信息圈之外，有很多人、很多行业、了不起的研究人员和工程师，正在独自做着惊人的事情，与他们无关，只是因为他们没有说出来——他们没说，并不意味着很多聪明人没有在做聪明的事。

> What's your level of confidence that in five years from now, we will have made the right decisions and the world will be that much better place, right, that we describe because of the investment and all that we're doing. Are you more confident as we sit here today than you were 12 months ago? I'm 99.99% just in service in honor of a good friend not using zero and 100. I'm 99. perjury, 99% confident that too many of us care do deeply that the outcome turns out well. That we will dedicate our might, our intellect, our resources to ensure we get there responsibly and optimistically. And so there's just so many people doing the right things. Now, oftentimes, there are people who pontificate about these things, and they're informed by several things, and they don't realize that there are a lot of people, industries, incredible researchers and engineers, outside of their sphere of information that are doing amazing things all by themselves, independent of them, just because they didn't say, they They didn't say it doesn't mean a lot of smart people are doing smart things.

### [1:35:57](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5757s) 未来十年：99.99%的信心
*The Decade Ahead: 99.99% Confident*

[1:37:01](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5821s) 我在公司内部一直这么说：即使没有我的领导、没有战略，公司本身也在做着聪明的事。聪明人会做聪明的事。当你把一条线组织起来时，他们会做出更有影响力的事，但了不起的人本身就在独自做着聪明的事。所以，当有人对我说“你在为我们担心”时，担心是好事。我们应该表达我们的担忧。我们应该表达我们的关切。但我们不应该预言世界末日。对。因为他们低估了人类的处境。他们低估了人类的能力。而且他们不知为何认为——也许是狂妄，也许是傲慢——除非他们来做，否则事情就做不成。对。对。除非他们来做，否则事情就做不成。对。事实是，这家公司正在做着惊人的事情，不管我说不说。对。所有的赞誉和指责都落到我头上，但在这两种情况下，我都坦然接受，尤其是赞誉。

> And I say this all the time inside the company, even without my leadership, without strategies, the company is doing smart things all by itself. Smart people do smart things. They do more impactful things when you organize a line, but amazing people are doing smart things all by themselves. And so when somebody tells me that you're worrying about us, it's good that we worry. And we should express our worries. We should express our concerns. But we shouldn't predict the end of the world. Yeah. Because they're underestimating the human condition. They're underestimating the human capacity. And they think for some reason with maybe it's just hubris, maybe it's arrogance, that unless they do it, it won't get done. Right. Right. Unless they do it, it won't get done. Right. The fact that it matters is this company is doing amazing things, whether I say it or not. Right. And I get all the accolades and the blame, but in both cases, I'm fair, especially the accolades.

[1:38:05](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5885s) 原因就在于这些……当你身边围绕着了不起的人，并且你创造出一种条件，让每个人都想要同样的结果——好的结果、好的版本。我认为大多数人都是这样。这就是为什么我说我们需要对自己问责、承担责任。但我们应该表达一个乐观的未来。当然。而且我99.99%确定，有世界上所有了不起的人围绕在身边，我们会一起走到那里。对，同意。而且我认为你是国宝。我们很幸运有你帮助引领这场冲锋。国家感激你把大家聚在一起、让我们达成共识、建立框架。还有很多工作要做。但作为两个孩子的父亲——我知道你有孩子，你将来还会有孙辈——我们最在乎的莫过于真正为他们建设一个更好的世界。我们大可以停下来。但如果我们今天停下，我们就会失去。

> And the reason for that is just these... When you surround yourself with amazing people and you create the conditions by which everybody wants the same outcome, the good outcome, the good version of it. And I think most people do. That's the reason why I say we need to hold ourselves accountable and responsible. But we should express a future that's optimistic. For sure. And I am 99.99% certain that surrounded by all the amazing people in the world, we're going to get there together. Yeah, agreed. And I think you're a national treasure. We're lucky to have you helping to lead the charge. The nation's indebted to you for getting everybody together, getting us all on the same page, building the framework. There's going to be more work to do. But as a father of two, and I know you have kids, you're going to have grandkids, and we don't care about anything more than actually building a better world for them. We could easily stop. But if we stop today, we lose.

[1:39:05](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5945s) 我们就会失去那个更好世界的机会。那我们就到这里。谢谢你，Jensen。谢谢你们，各位。Brad，谢谢你。我从未见过比你更伟大的爱国者，一个如此关心这个行业、但更关心这个国家的人，还有你的天赋、你的热情，以及你在协调、把很多人聚拢起来去做像你所做的那样伟大的事情时那种不知疲倦的精力。

> We lose the opportunity at that better world. So we'll wrap it there. Thank you, Jensen. Thank you, guys. Brad, thank you for it. I've never met a greater patriot, somebody who cares so much about the industry, but even more so the country and your gift and your enthusiasm and your tireless energy in coordinating, pulling together a lot of people in order to do something great like what you did.

[1:39:48](https://www.youtube.com/watch?v=eS7qQ3WJSrw&t=5988s) 提醒大家，这只是我们的观点，不是投资建议。

> As a reminder to everybody, just our opinions, not investment advice.

---

*来源：[研几 · AI](https://ai-daily-cn.pages.dev/watch/bg2-jensen-huang/) · 原节目 [YouTube](https://www.youtube.com/watch?v=eS7qQ3WJSrw)。文字稿：faster-whisper small.en（本机 CPU），时间轴即 YouTube 原片时间（Bg2 Pod）；未区分说话人；英文原声：DeepSeek 逐段翻译（105 段），编辑抽查；核心总结由 DeepSeek 依据译文生成*