Nobody Needs Atomic Bombs. Everybody Needs AI.

I think that OpenAI is likely going to be the next multi- trillion dollar hypers scale company.

我认为 OpenAI 很可能会成为下一家数万亿美元级的超大规模公司。

Okay, [Music] [Applause]

好。[音乐][掌声]

Jensen.

Jensen。

Great to be back, of course, with my partner Clark Tang.

很高兴又回来了,当然,还是和我的搭档 Clark Tang 一起。

You know, I can't believe it's

我真不敢相信已经

Welcome to Invidia.

欢迎来到 NVIDIA。

Uh, oh, and nice glasses.

哦,还有,眼镜不错。

Um, those actually look really good on you.

这副眼镜你戴着真好看。

The problem is now everybody's going to want you to wear them all the time.

问题是,现在所有人都会希望你一直戴着它。

They're going to say, "Where are the red glasses?"

他们会说:「那副红眼镜呢?」

I can vouch for that.

这一点我可以作证。

So, it's been over a year since we did the last pod.

我们上次录播客到现在,已经一年多了。

Yeah.

对。

Over 40% of your revenue today is inference, but inference is about ready because of chain of reasoning.

今天你们 40% 以上的营收来自推理,而因为思维链,推理马上就要——

Yeah.

对。

Right.

没错。

It's about

它马上就要

It's about to go up by a billion times,

它马上就要涨 10 亿倍,

right?

对吧?

By by a million x by a billion.

涨 100 万倍——涨 10 亿倍。

That's right.

没错。

That's the part that most people have, you know, haven't completely internalized.

这一点大多数人还没有完全消化。

This is that industry we were talking about.

这就是我们当时讲的那个产业。

This is the industrial revolution.

这就是工业革命。

Honestly, it's it's felt like you and I have had a continuation of the pod every day since then.

老实说,从那以后,感觉这期播客我们俩天天都在往下接。

You know, in AI time, it's been about a hundred years.

按 AI 的时间尺度算,这中间已经过去了大概一百年。

I was re-watching the pod recently and the many things that we talked about that stood out.

我最近重看了那期播客,当时聊到的很多东西现在格外醒目。

The most the one that that was probably most profound for me was you pounding the table

对我来说最有分量的一点,大概就是你拍着桌子讲的那段——

that you know remember at the time there was kind of a slump in terms of pre-training

还记得吗,当时预训练那边有点低潮,

and people were like, "Oh my god,

大家都在说:「我的天,

the end of pre-training,

预训练到头了,

right?

对吧?

The end of pre-training.

预训练到头了。

We're not going.

我们走不下去了。

We're overbuilding.

我们建过头了。

This is about a year and a half ago.

这大概是一年半以前的事。

And you said inference isn't going to 100x, a thousandx, it's going to 1 billionx.

而你说,推理不是要涨 100 倍、1000 倍,而是要涨 10 亿倍。

Mhm.

嗯。

Which brings us to where we are today.

这就说到我们今天的局面了。

You know, you announced this huge deal.

你们宣布了这笔巨额交易。

We ought to start there.

我们应该从这里聊起。

I underestimated.

我低估了。

Let me just go on record.

我先把话说在这儿。

I estimated we we now have three scaling laws, right?

我当时估计——我们现在有三条 scaling law,对吧?

We have pre-training scaling law.

我们有预训练的 scaling law。

We have post-training scaling law.

我们有后训练的 scaling law。

Post-training is basically like uh AI practicing.

后训练基本上就是 AI 在练习。

Yes,

对,

practicing a skill until it gets it right.

练一项技能,一直练到做对为止。

And so it tries a whole bunch of different ways and and uh in order to do that, yeah, you've got to do inference.

它会试一大堆不同的路子,而要做到这件事,你就得做推理。

So now training and inference are now integrated in reinforcement learning.

所以现在训练和推理在强化学习里合到一起了。

Really complicated.

非常复杂。

And so that's called post training.

这就叫后训练。

And then the third

然后第三条

is inference.

是推理。

The old way of doing inference was one shot,

过去做推理的方式是一次成型(one-shot),

right?

对吧?

But the new way of doing inference, which we appreciate, is thinking.

但我们现在体会到的新推理方式是思考。

So think before you answer.

回答之前先想一想。

Yeah.

对。

And so now you have three scaling laws.

所以现在你有三条 scaling law。

The the longer you think, the better the quality answer you get.

你想得越久,拿到的答案质量越高。

While you're thinking, you do research, you go check on some ground truth.

思考的过程中,你会去做研究,去核对一些 ground truth。

And you you learn some things, you think some more, you go learn some more, and then you generate an answer.

你学到一些东西,再多想一会儿,再去多学一点,然后生成答案。

Don't just generate right off the bat.

不要一上来就直接生成。

And so thinking, post-training, pre-training, we now have three scaling laws, not one.

所以思考、后训练、预训练——我们现在有三条 scaling law,不是一条。

You knew that last year, but is your level of confidence this year in the inferences going to 1 billionx and where that will take the levels of intelligence is it higher?

这一点你去年就知道了。但今年,你对推理会涨 10 亿倍、对它能把智能水平推到什么高度,信心是不是更足了?

Are you more confident this year than you were a year ago?

今年你比一年前更有信心吗?

I'm more confident this year and the reason for that is because look at the agent systems now

我今年更有信心,原因是,你看看现在的智能体系统——

and AI is no longer a language model and AI is a system of language models and they're all running concurrently maybe using tools.

AI 已经不再是一个语言模型,而是一个由语言模型组成的系统,它们全都在并发运行,可能还在调用工具。

Some of us using tools, some of us doing research and yeah, there's a whole bunch of stuff and it's all multimodality and look at all the video that's being generated.

其中一些在用工具,一些在做研究,还有一大堆事情同时发生,而且全都是多模态的,你看看现在生成出来的那些视频。

I mean, it's just crazy stuff.

真的太疯狂了。

Yeah.

是啊。

It really brings us to, you know, kind of the seminal moment this week that everybody's talking about the massive deal.

这就正好说到本周那个所有人都在谈的标志性时刻——那笔巨额交易。

You announced a couple days ago with OpenAI Stargate where that you're going to be a preferred partner, invest hundred billion dollars in the company over a period of time. they're going to build 10 gigs and if they used Nvidia for those 10 gigs that could be upwards of 400 billion in revenue to Nvidia.

你们几天前宣布了跟 OpenAI Stargate 的合作:你们会成为优先合作伙伴,在一段时间里陆续向这家公司投资 $100 billion;他们要建 10 gigs,而如果这 10 gigs 用的是 NVIDIA,那可能给 NVIDIA 带来高达 $400 billion 的营收。

So help us understand you just tell us a little bit about that partnership what it means to you right and why that investment makes so much sense for Nvidia.

所以帮我们理解一下,讲讲这段合作、它对你意味着什么,以及这笔投资为什么对 NVIDIA 特别合理。

So first of all that I'll answer that last question first and then I'll come back and present my way through.

首先,我先回答最后那个问题,然后再回过头来一步步讲。

I think that OpenAI is likely going to be the next multi- trillion dollar hypers scale company.

我认为 OpenAI 很可能成为下一家数万亿美元级的超大规模公司。

Okay.

好。

I think you and I

我想你和我

Why do you call it a hypers scale company?

你为什么把它叫超大规模公司?

Hypers scale like uh like Meta is a hypers scale.

超大规模,就像 Meta 就是一家超大规模公司。

Uh Google's a hypers scale.

Google 也是超大规模公司。

They're going to have consumer and enterprise services and and uh they are very likely going to be the world's next multi-t trillion dollar hypers scale company.

他们会有面向消费者和面向企业的服务,而且很可能成为全世界下一家数万亿美元级的超大规模公司。

Yes.

是的。

And I think you would agree with that.

而且我想你会同意这一点。

I agree.

我同意。

If that's the case, the opportunity to invest before they get there, this is some of the smartest investments we can possibly imagine.

如果真是这样,能在他们走到那一步之前投进去,这就是我们能想到的最聪明的投资之一。

And you got to invest in things, you know,

而且你得投你自己懂的东西,

right?

没错。

And it turns out we happen to know this space.

而事实证明,我们恰好懂这个领域。

And so the opportunity to invest in that,

所以投进去的这个机会,

the return on that money is going to be fantastic.

这笔钱的回报会非常可观。

So we love the opportunity to invest.

所以我们很喜欢这个投资机会。

We don't have to invest, right?

我们并不是非投不可,对吧?

And it's not required for us to invest, but they're giving us the opportunity to invest.

投资对我们不是必须的,但他们把这个机会给了我们。

Fantastic thing.

这是大好事。

Now let me start from the beginning.

现在我从头讲起。

So we're partnering with with OpenAI in several projects.

我们和 OpenAI 在好几个项目上合作。

First the first project is the buildout of Microsoft Azure.

第一个项目是 Microsoft Azure 的建设。

We're going to continue to do that

我们会继续做下去

and that that partnership is going fantastically.

而这个合作进展得非常好。

We're going we have several years of buildout to do hundreds of billions of dollars of work just to do there.

我们还有好几年的建设要做,光是那边就有数千亿美元的活要干。

Right.

没错。

The second is the OCI buildout

第二个是 OCI 的建设

and uh I think there's some five, six, seven gigawatts that are about to be built out.

我想大概有 5、6、7 吉瓦马上就要建起来。

And so we're working with OCI and OpenAI and and SoftBank to build that out,

所以我们在和 OCI、OpenAI 还有 SoftBank 一起把它建起来,

right?

没错。

Those projects are contracted.

那些项目都已经签了合同。

Um we're working on it.

我们正在推进。

Lots of work to do.

有很多活要干。

And then the third the third is Core Weave, right?

然后第三个是 CoreWeave,对吧?

And so uh all of Core Wee 4, I'm talking about OpenAI still.

所以 CoreWeave 那一整块,我说的还是 OpenAI。

Yes.

是的。

Okay.

好。

Everything in the context of OpenAI.

所有这些都是在 OpenAI 这个范围里。

And so so the question is what is this new partnership?

所以问题是,这个新合作到底是什么?

This this new partnership is about helping OpenAI working partnering with OpenAI to build their own selfbuild AI infrastructure for the first time,

这个新合作,是帮 OpenAI、和 OpenAI 一起,第一次建起他们自己的 AI 基础设施,

right?

没错。

And so this is us working directly with OpenAI at the chip level, at the software level, at the systems level, at the AI factory level um to help them become a fully operated hypers scale company that I mean this is going to go on for some time.

所以这是我们直接和 OpenAI 一起干,在芯片层面、软件层面、系统层面、AI 工厂层面,帮他们变成一家完全自主运营的超大规模公司,这件事会持续相当长一段时间。

And it's going to supplement it's going to supplement the amount of you know they're going through two exponentials as you know

而这会补上其中一部分——你也知道,他们正在经历两条指数曲线

right

没错。

the first exponential is the number of customers is growing exponentially and the reason for that is the AI is getting better the use case is getting better just about every every application is connected to open now

第一条指数曲线是客户数量在指数级增长,原因是 AI 越来越好、用例越来越好,几乎每一个应用现在都接到了 OpenAI 上

and so they're going through the the usage exponential the second exponential is the computational exponential of every use.

所以他们在经历使用量的指数曲线,第二条指数曲线是每一次使用背后计算量的指数曲线。

Yes.

是的。

Right.

没错。

Yes.

是的。

Instead of just a oneshot inference is now thinking before it answers.

不再是只做一次成型(one-shot)的推理,现在是回答之前先思考。

Yeah.

对。

And so these two exponential is compounding their compute requirements.

所以这两条指数曲线叠加,把他们的算力需求成倍推高。

And so we we got to build out these all these different projects.

所以我们得把所有这些不同的项目都建起来。

And so this last one is an additive on top of everything that they've already announced, all the things that we're already working on with them.

所以最后这一个,是加在他们已经宣布的所有东西、我们已经在跟他们做的所有事情之上。

It's additive on top of that

它是叠加在那之上的

and uh it's going to support the you know this incredible exponential growth.

而且它会支撑这种惊人的指数级增长。

One of the things you said there that's really interesting to me is kind of you know they're going to be high probability multi-t trillion dollar company in your mind.

你刚才说的其中一点我觉得特别有意思,就是在你心里,他们有很高概率成为一家数万亿美元的公司。

I think it's a great investment.

我觉得这是一笔很好的投资。

At the same time, you know, they're selfbuilding.

同时,他们在自建。

You're helping them self-build their data centers.

你在帮他们自建数据中心。

So, here to four, they've been outsourcing to Microsoft to build the data center.

所以到目前为止,他们一直把数据中心的建设外包给 Microsoft。

Now, they want to build full stack factories themselves.

现在他们想自己建全栈工厂。

They want to do they want to they want to basically have a relationship with us the way that that Elon and X has relation.

他们想要的,基本上就是和我们建立一种像 Elon 和 X 那样的关系。

Correct.

对。

I mean, Elon and Exel built.

我是说,Elon 和 xAI 是自己建的。

Exactly.

正是。

But I think that's

但我觉得这

this is a very big deal

这是一件非常大的事

when you think when you think about

当你想到,当你想到

the advantage that Colossus had.

Colossus 带来的那种优势。

Yeah,

对,

they're building full stack.

他们在建全栈。

That is a hyperscaler because if they don't use the capacity, they could sell it to somebody else.

那就是超大规模厂商,因为如果这些产能他们自己用不掉,可以卖给别人。

In the same way Stargate, they're building monstrous capacity.

同样地,Stargate 他们在建的产能大得离谱。

They think they'll need to use most of it, but it puts them in a position to sell it to somebody else as well.

他们觉得自己会用掉其中大部分,但这也让他们有条件把产能卖给别人。

It sounds very much like AWS or GCP or Azure.

这听起来非常像 AWS、GCP 或者 Azure。

That's what you're saying.

你说的就是这个意思。

Yeah, I I think they'll likely use it themselves and um just think the case of X, they'll likely use it themselves.

对,我觉得他们很可能自己用掉——想想 X 的情况,他们很可能自己就用掉了。

Um but they would like to have the the same direct relationship with us, direct working relationship and direct purchasing relationship.

但他们希望和我们有同样的直接关系:直接的工作关系,直接的采购关系。

Um uh Meta just as with Zuck and Meta has with us uh it's exactly a direct

就像 Zuck 和 Meta 跟我们的关系那样,那完全是直接的

um our relationship with uh between us and Sunund and Google direct our partnership with Satia and Azure direct.

我们和 Sundar Pichai、和 Google 的关系是直接的,我们和 Satya、和 Azure 的合作是直接的。

Isn't that right?

不是吗?

And so they've gotten to a large enough scale that they believe it's time for them to start building these direct relationships.

所以他们已经做到足够大的规模,觉得是时候开始建立这些直接关系了。

So, I'm delighted to support that and and all of and and Satia knows it and Larry knows it and everybody everybody's aware of what's going on and everybody's very supportive of it.

所以我很乐意支持,Satya 知道,Larry 也知道,大家都清楚在发生什么,而且大家都非常支持。

So, one of the things I find mysterious, right?

有件事我一直觉得很费解,对吧?

You know, you just mentioned Oracle 300 billion Colossus what they're building.

你刚提到 Oracle 的 300 billion、Colossus,他们在建的那些东西。

We know what the sovereigns are building.

我们知道各国主权方在建什么。

We know what the hyperscalers are building.

我们知道超大规模厂商在建什么。

You know, Sam's talking in terms of trillions.

Sam 开口谈的就是 trillions 这个量级。

But of the 25 sellside analysts on Wall Street who cover your stock, if I look at the consensus estimate, it basically has your growth flatlining starting in 2027.

但华尔街覆盖你们股票的那 25 位卖方分析师,我去看卖方共识预期,基本上是说你们的增长从 2027 年开始就走平了。

8% growth 2027 through 2030.

2027 到 2030 年,8% 增长。

Okay, that is the 25 people in their only job.

好,而这 25 个人就干这一件事。

They get paid to forecast the growth rate for Nvidia.

他们拿钱就是为了预测 NVIDIA 的增长率。

So clearly

所以很明显——

we're comfortable with that by the way.

顺便说一句,我们对这个没意见。

Right.

没错。

Look, we're comfortable with that.

听着,我们对这个没意见。

Okay, we have no trouble beating the numbers on a regular basis,

我们要定期超掉这些数字,一点都不难,

right?

对吧?

No, I understand that.

不,这我明白。

But, but there is this interesting disconnect.

但这里有一个很有意思的脱节。

No,

不,

right.

没错。

I hear it every day on CNBC and Bloomberg.

我每天在 CNBC 和 Bloomberg 上都听到这个说法。

And I think it goes to, you know, some of these questions around, you know, uh, shortages leading to a glut that they don't believe.

我觉得这跟围绕「短缺会不会演变成产能过剩」的那些疑问有关,他们就是不信。

They say, "Okay, we'll give you credit for 26, but 27, you know, maybe we'll have too much and you're not going to need that.

他们说:「好,26 年我们认你,但 27 年嘛,可能就供过于求了,你也不需要那么多了。」

But it is interesting to me and I think it's important to point out that the your consensus forecast is that this won't happen right and we also put together forecast uh you know for the company taking into account all of these numbers

但我觉得这很有意思,而且我认为有必要指出:卖方共识预期就是认为这不会发生,对吧,而我们自己也给公司做了一份预测,把所有这些数字都算了进去。

and what it shows me is still even though we're two and a half years into the age of AI a massive divergence of belief

它给我看到的是:哪怕我们已经进入 AI 时代两年半了,信念上依然存在巨大分歧。

between what we hear Sam Alman saying you saying Sundar saying Satcha is saying and what Wall Street still believes and you know again you're comfortable with that.

一边是我们听到 Sam Altman 说的、你说的、Sundar Pichai 说的、Satya Nadella 说的,另一边是华尔街到现在还相信的,而你刚才说了,你对这个没意见。

I also don't think it's inconsistent.

我也不觉得这两者矛盾。

Okay.

好。

So explain that a little bit.

那你稍微解释一下。

So first of all uh for the builders we're supposed to be building for opportunity

首先,我们这些搞建设的人,本来就该为机会而建。

right?

对吧?

We're we're builders.

我们是建设者。

Let me give you three points to think through and and and these three points uh it'll help you um hopefully uh be more comfortable with Nvidia in this future.

我给你三点,你顺着想一想,这三点应该能让你对未来的 NVIDIA 更放心一些。

So the first point and this is the laws of physics point.

第一点,关于物理定律。

This is the most important point that general general purpose computing is over and the future is accelerated computing and AI computing.

这是最重要的一点:通用计算结束了,未来是加速计算和 AI 计算。

That's the first point.

这是第一点。

And so the way to think about that is there's how much how many trillions of dollars of computing infrastructures in the world that has to be refreshed.

所以该这么想:世界上有多少万亿美元的计算基础设施等着换掉。

Right.

没错。

Right.

没错。

And when it gets refreshed it's going to be accelerated comput.

而换的时候,换上来的会是加速计算。

That's right.

对。

And so the first thing you have to realize is that general purpose computing and nobody disputes that.

所以你首先要意识到,通用计算——这一点没人有争议。

Everybody goes, "Yeah, we completely agree with that.

所有人都说:「是啊,我们完全同意。

General purpose computing is over.

通用计算结束了。

Moore's law is dead.

摩尔定律死了。「

People say these things.

人们都在这么说。

And so what does that mean?

那这意味着什么?

So general purpose computing is going to go to accelerated computing.

通用计算会走向加速计算。

Our partnership with Intel is recognizing that general purpose computing needs to be fused with accelerated computing to create opportunities for them.

我们和 Intel 的合作,就是认识到通用计算需要跟加速计算融合起来,才能给他们创造机会。

Is that right?

不是这样吗?

And so one,

所以第一点,

general purpose computing is shifting to accelerated computing and AI.

通用计算正在转向加速计算和 AI。

Two, the first use case of AI is actually already everywhere,

第二,AI 的第一个用例其实已经无处不在了,

right?

对吧?

It's in search recommener engines,

它就在搜索的推荐引擎里,

isn't that right?

不是这样吗?

In shopping.

购物里也是。

The basic hypers scale computing infrastructure used to be CPUs doing recommenders, right?

超大规模计算最基础的那层基础设施,以前是 CPU 在跑推荐,对吧?

Is now going to GPUs

现在正在换成 GPU,

doing AI,

跑的是 AI,

right?

对吧?

So you just take classical computing, it's going to accelerated computing AI.

所以你看,经典计算正在走向加速计算和 AI。

You take hypers scale computing is going from CPUs to accelerated computing and AI and then now that's the second point just feeding the metas the Google's the bite dances the Amazons

超大规模计算正在从 CPU 走向加速计算和 AI,这就是第二点——光是喂 Meta 这些、Google 这些、字节跳动这些、Amazon 这些公司。

and take their classical traditional way of doing hyperscaling and moving into AI

把他们做超大规模那套经典的传统方式搬到 AI 上,

that's hundreds of billions of dollars

那就是几千亿美元。

and and because that may be four billion people on the planet today if you take Tik Tok meta into account

而且因为今天地球上可能有 40 亿人,如果把 TikTok、Meta 算进去,

that's Google into account who are already demanding workloads that are driven by accelerated comput.

把 Google 也算进去,这些人已经在产生由加速计算驱动的负载需求了。

That's exactly right.

完全正确。

And so there a simp without even thinking about AI creating new opportunities.

所以这里就已经有一块——甚至还没算上 AI 去创造新机会。

It's about AI shifting how you used to do something to the way new way of doing something.

这只是 AI 把你过去做一件事的方式,换成新的做法。

Okay.

好。

And then now let's talk about the future.

那现在我们来谈未来。

I just so far I've only spoken kind of largely about

到目前为止,我讲的大体上只是——

just mundane stuff.

只是些平常的东西。

Just mundane stuff.

只是些平常的东西。

The old way is now wrong.

老办法现在是错的。

You're going to go, you're no longer going to use uh uh fuel light lanterns.

你不会再用煤油灯了。

You're going to go to electricity.

你会用电。

That's all.

就是这样。

Right.

没错。

Okay.

好。

And you no longer, you know, prop planes.

你也不会再用螺旋桨飞机了。

You're going to go to jets.

你会用喷气机。

That's all.

就是这样。

And so, you know, so far, you know, that's all I've talked about.

所以到目前为止,我讲的就只有这些。

And then

然后——

now that the incredible thing is when you go to AI, when you go to accelerated computing, then what happens?

现在,不可思议的地方在于:当你走向 AI、走向加速计算之后,会发生什么?

What are the new applications that emerge as a result?

结果会涌现出哪些新应用?

And that's all the AI stuff that we're talking about.

那就是我们在谈的所有 AI 的东西。

And that's the that opportunity.

那就是那个机会。

What is it?

它是什么?

How do what does that look like?

它长什么样?

Well, the simple way of thinking about that is where motors replace labor and physical activity.

简单想一想:马达取代的是劳动力和体力活动。

We now have AI.

现在我们有了 AI。

These AI supercomputers, these AI factories that I talk about, they're going to generate tokens to augment human intelligence, right?

这些 AI 超级计算机、我一直在讲的这些 AI 工厂,会生成 token 来增强人的智能,对吧?

And human intelligence represents what 55 65% of the world's GDP.

而人的智能占世界 GDP 的多少?55% 到 65%?

Let's call it $50 trillion.

就算它 $50 trillion。

And that $50 trillion is going to get augmented by something.

而这 $50 trillion,会有东西来增强它。

And so let's you just let's come back to a single person.

我们回到单独一个人身上。

Suppose I were to hire a $100,000 employee.

假设我雇一个 $100,000 的员工。

And I augmented that $100,000 employee with a $10,000 AI.

然后我拿一个 $10,000 的 AI 去增强这个 $100,000 的员工。

Yes.

是。

And that $10,000 AI as a result made that $100,000 employee twice more productive, three times more productive.

而这个 $10,000 的 AI 让这个 $100,000 的员工生产力翻一倍、变成三倍。

Would I do it?

我会做吗?

Heartbeat.

毫不犹豫。

I I'm doing it across every single person in our company right now.

我现在就在公司里对每一个人这么做。

Right.

没错。

Every single co-agents.

每个人都配了协同 agent。

That's right.

对。

Every That's right.

每一个——对。

Every single software engineer, every single chip designer in our company already has AIS working with them.

我们公司里每一个软件工程师、每一个芯片设计师,已经都有 AI 在跟他们一起干活。

100% coverage.

100% 覆盖。

As a result,

结果就是,

the number of chips we're building is better.

我们造的芯片数量更好了。

The number is growing.

数量在增长。

The pace at which we're doing it is right.

我们做这件事的节奏也对了。

And so we're we're growing faster as a company.

所以我们这家公司增长得更快。

As a result, we're hiring more people.

结果是我们雇了更多人。

Our productivity is greater.

我们的生产力更高。

Our top line's greater.

我们的营收更高。

Our profitability is greater.

我们的盈利能力更强。

What's not to love about that?

这有什么不好的?

Now apply the Nvidia story to the world's GDP.

现在把 NVIDIA 这个故事套到全世界的 GDP 上。

Yeah.

对。

And so what's likely to happen is that that $50 trillion is augmented by let's pick a number

所以很可能发生的是:那 $50 trillion 会被增强——随便挑个数字,

10 trillion that $10 trillion

10 trillion,这 $10 trillion

needs to run on a machine. M

需要跑在机器上。

now the reason that AI is different than it in the past

现在,AI 跟过去不一样的原因在于,

in a way software was written a priori

过去软件某种意义上是事先写好的,

and then it runs on a CPU and it doesn't it runs it a a person would operate it

然后跑在 CPU 上,而且得有个人去操作它。

in the future of course AI is generating tokens

而在未来,AI 当然是在生成 token,

but a machine has to generate the tokens and it's thinking

但必须由一台机器来生成这些 token,而它在思考。

so that software is running all the time whereas in the past the software was written once now the software is in fact writing all the time it's thinking

所以那套软件是一直在运行的,而过去软件只写一次,现在软件事实上一直在写、它在思考。

in order for the AI to think it needs a factory.

而为了让 AI 思考,它需要一座工厂。

And so let's say that that 10 trillion of token generated

所以假设那 10 trillion 的 token 生成——

50% gross margins and 5 trillion of it needs a factory needs an AI infrastructure.

按 50% 毛利率算,其中 5 trillion 需要工厂、需要 AI 基础设施。

So if you told me that on an annual basis the capex of the world was about $5 trillion

所以如果你告诉我,全世界每年的 capex 大概是 $5 trillion,

I would say the math seems to make sense.

我会说这个数算下来是说得通的。

Yeah.

对。

And that's kind of the future, right?

这大概就是未来,对吧?

Yeah. the going from Excel general purpose computing to accelerated computing replacing all the hypers scales with AI and then now augmenting human intelligence for the world's GDP

对。从 Excel、通用计算走向加速计算,把所有超大规模计算都换成 AI,然后现在去增强世界 GDP 里的人类智能。

and today that market is about our estimate is about 400 billion annually.

而今天这个市场,我们估算大概是每年 400 billion。

Yeah.

对。

So the the TAM you know is is a four to 5x increase over where it is today.

所以这个 TAM 是今天的 4 到 5 倍。

Yeah.

对。

Eddie last night, Eddie Woo at Alibaba said between now and the end of the year and excuse me, now and the end of the decade, they're going to increase their data center power by 10x.

昨晚 Eddie,阿里巴巴的 Eddie Wu 说,从现在到今年底——抱歉,是从现在到这个十年结束,他们要把数据中心的电力提升 10 倍。

Right.

对。

Right.

没错。

You just said how much?

你刚才说是多少?

4x.

4 倍。

There you go.

这就对了。

Yeah.

对。

They're going to increase power by 10x.

他们要把电力提升 10 倍。

And we we correlate to power.

而我们是跟电力挂钩的。

Nvidia's revenue is almost correlated to power.

NVIDIA 的营收几乎跟电力挂钩。

Isn't that right?

不是这样吗?

Yeah,

对,

that's right.

没错。

Yeah.

对。

Because

因为——

one other thing, what he what else did he say?

还有一件事,他还说了什么?

Yeah.

对。

He said token generation is doubling every few months.

他说 token 生成每隔几个月就翻一倍。

Yeah.

对。

What's that saying?

这说明什么?

The the perf per watt has to keep on going exponentially.

每瓦性能必须持续按指数往上走。

That's why Nvidia's like cranking it out with perf per watt.

所以 NVIDIA 才在每瓦性能上死命往前推。

And revenue per watt is, you know, watt is basically revenues in this future.

而每瓦营收——在这个未来里,瓦基本上就等于营收。

Embedded in this assumption, I find it very fascinating historical context, right?

这个假设里隐含着一段我觉得非常迷人的历史背景,对吧?

For 2,000 years, basically GDP did not grow.

有 2,000 年的时间,GDP 基本上不增长。

Okay?

对吧?

And then we get the industrial revolution, GDP accelerates.

然后工业革命来了,GDP 加速。

We get the digital revolution, GDP accelerates.

数字革命来了,GDP 又加速。

And it basically what you're saying, and Scott Besson has said it, he said, I think we're going to have 4% GDP growth next year.

而你说的基本上就是——Scott Bessent 也说过,他说我认为明年我们会有 4% 的 GDP 增长。

Basically, what you're saying is the world's GDP growth is going to accelerate because now we are giving the world billions of co-workers that will do work for us.

你说的基本上是:世界 GDP 增速会加速,因为我们现在给这个世界几十亿个替我们干活的同事。

And if GDP is an amount of output for a fixed amount of labor and capital, right, it has to accelerate.

而如果 GDP 就是固定劳动力和资本投入下的产出量,对吧,那它必然会加速。

It has to,

必然会,

right?

对吧?

It has to look at what's going on with AI as a result of the technology of AI.

必然。看看 AI 正在发生的事,这都是 AI 这项技术带来的结果。

And that technology of AI, let's just call it the large language models and all the AI agents.

而 AI 这项技术,我们就叫它大语言模型和所有那些 AI agent。

It's now creating a new industry of AI agents.

它现在正在创造一个新产业:AI agent 产业。

There's no question about that.

这一点毫无疑问。

Okay.

好。

So, so that's OpenAI is the fastest growing revenue company in history,

所以,OpenAI 是史上营收增长最快的公司,

right? and they're growing exponentially, right?

对吧?而且他们是指数级增长,对吧?

And so, so AI itself is a fast growing industry because of AI needs a factory behind it, right?

所以 AI 本身就是一个高速增长的产业,而因为 AI 背后需要一座工厂,对吧?

An infrastructure behind it.

背后需要基础设施。

There's this industry is growing.

这个产业在增长。

My industry is growing.

我的产业在增长。

And because my industry is growing, the industry underneath it is growing.

而因为我的产业在增长,它下面那一层产业也在增长。

Energy is growing.

能源在增长。

Power shell.

电力。

This is the

这就是——

This is like renaissance for the energy industry, isn't that right?

这简直是能源行业的一场文艺复兴,不是这样吗?

Nuclear energy, you know, gas turbines.

核能、燃气轮机。

I mean, look at all of those companies in the in the infrastructure ecosystem underneath us.

你看我们下面那整个基础设施生态里的所有公司。

They're doing incredibly well.

他们的表现好得不得了。

Everybody's growing.

每个人都在增长。

These numbers have everybody talking about a glutter bubble, right?

这些数字让所有人都在谈产能过剩、谈泡沫,对吧?

Zuckerberg said last week on a podcast, you know, he said, "Listen, I think it's quite possible at some point that we will have an air pocket and Meta may in fact overspend by $10 billion or whatever, but he said it doesn't matter.

Zuckerberg 上周在一个播客里说:「听着,我觉得很有可能在某个时点我们会碰上需求真空,Meta 甚至可能多花 $10 billion 之类的」,但他说这不重要。

It's so existential to the future of his business that it's a risk that they have to take.

这对他生意的未来太生死攸关了,所以这是他们必须承担的风险。

But when you think about that, it sounds a little bit like prisoners dilemma.

但你这么一想,这听起来有点像囚徒困境。

Right.

对吧。

And walk us again through

你能不能再给我们讲一遍——

these are very happy prisoners.

那这些囚徒可开心得很。

Walk us again through Right.

再帮我们捋一遍 —— 好。

Today our estimate is that we're going to have a 100 billion of AI revenue in 2026 excluding meta and excluding you know the GPUs running recommener engines.

我们现在的估算是,2026 年会有 100 billion 的 AI 营收,不算 Meta,也不算跑推荐引擎的那些 GPU。

Okay.

好。

So there's

所以还有 ——

or search or

或者搜索,或者 ——

correct.

对。

So there there's other stuff but let's call it 100 billion.

所以确实还有别的部分,但就算它 100 billion 吧。

What is that industry anyways?

那个产业本身到底有多大?

What is the industry already in hypers scale?

已经算在超大规模里的那部分产业有多大?

What is the hypers scales you know between trillions?

超大规模那块儿有多少?好几万亿?

Yeah exactly by the way that industry is going to AI

对,没错 —— 顺便说一句,那个产业正在整个转向 AI。

before anybody starts at zero.

在那之前,没有人是从零开始的。

You got to start there.

你得从那儿起算。

But I think the skeptics would say we need to go from a 100red billion of AI revenue in 26 to at least a trillion of AI revenue in 2030.

但我想怀疑的人会说,我们得从 2026 年的 100 billion AI 营收,涨到 2030 年至少 1 trillion 的 AI 营收。

Okay.

好。

You just were talking a minute ago about five trillion when you look at kind of global GDP.

你一分钟前刚说到 five trillion,那是从全球 GDP 的角度看。

If you do did a bottoms up, can you see your way to a trillion dollars of AIdriven revenues from a hundred billion over the course of the next 5 years?

如果做自下而上的测算,你看得到未来 5 年里从 100 billion 走到 1 trillion 美元 AI 驱动营收的路径吗?

Are we growing that fast?

我们真有这么快的增长吗?

Yes.

有。

And I would also say we're already there.

而且我还要说,我们其实已经到了。

Okay.

好。

So explain that.

那你解释一下。

Because the hyperscalers, they went from CPUs to AI.

因为超大规模厂商已经从 CPU 转到 AI 了。

Okay.

好。

Their entire revenue base is all now AIdriven.

他们整个营收基本盘,现在全是 AI 驱动的。

Correct.

对。

Uh you can't do Tik Tok without AI.

没有 AI,你做不了 TikTok。

Correct.

对。

You can't do YouTube short without AI.

没有 AI,你做不了 YouTube Shorts。

You can't you know you can't do any of this stuff without AI.

这些事,离了 AI 一件也做不成。

Uh the the amazing things that that Meta is doing for for uh um uh you know customized content, personalized content.

Meta 在定制内容、个性化内容上做出来的那些了不起的东西。

You can't do that without AI.

没有 AI,这些都做不到。

It's all of that stuff used to be humans, you know, doing uh content a priori creating four choices that are then

这些以前全靠人 —— 事先做内容,先造出四个选项,然后再

selected by a recommener engine.

由推荐引擎挑出来。

Correct.

对。

And now it's infinite number of choices generated by an AI, right?

而现在是 AI 生成无穷多个选项,对吧?

But those things are already like we had the transition from CPUs to GPUs largely for those recommener engines and now they're going

但那些已经…… 我们已经经历过一轮从 CPU 到 GPU 的迁移,主要就是为了那些推荐引擎,而现在它们又要

and that's fairly new I would

而这本身也很新,我得说

in the last three or four years.

就是过去三四年的事。

Zuck would tell you I was at Sigraph and Zuck would tell you you know they were late getting to GPU for sure for sure.

Zuck 会告诉你 —— 我在 SIGGRAPH 上,Zuck 会告诉你,他们上 GPU 确实晚了,确实晚了。

GPUs for for meta is what couple years and a half

Meta 用 GPU 才多久,两年半?

it's pretty new search with GPUs

挺新的,用 GPU 做搜索。

for sure

确实。

brand spanking new

崭新得不得了。

for sure for sure

确实,确实。

brand spanking new

崭新得不得了。

search for GPUs on GPUs

搜索 —— 在 GPU 上跑搜索。

so your argument would be the probability that we're going to have a trillion dollars of AI revenues by 2030 is near certain because we're almost already there.

所以你的论点是,到 2030 年有 1 trillion 美元 AI 营收的概率几乎是确定的,因为我们差不多已经到了。

Okay,

好,

let's just talk about incremental from where we are.

那就只谈从现在往上的增量。

Now we can talk about incremental from where we are today, right?

现在我们可以谈从今天往上的增量了,对吧?

As you do your bottoms up or your tops down, I just heard your tops down about percentage of global GDP.

你做自下而上、或者自上而下的测算 —— 自上而下的我刚听到了,是按全球 GDP 的百分比算。

Yeah.

嗯。

What is the percentage probability that you think we'll have a glut will run into a glut in the next three or four or five years?

你觉得未来三年、四年、五年里撞上产能过剩的概率是多少?

Right.

对。

It's a distribution of we don't know the future.

这是一个分布 —— 未来我们并不知道。

It's a distribution of power

这是一种力量的分布 ——

until until we fully convert all general purpose computing to accelerated computing and AI.

在我们把所有通用计算彻底转成加速计算和 AI 之前 ——

Until we do that,

在做到那一步之前,

yes,

是的,

I think the chances are extremely low.

我认为出现产能过剩的概率极低。

Okay.

好。

Okay.

好。

And that will take a few years.

而那要花几年。

That'll take a few years.

那要花几年。

Yeah.

嗯。

Let me ask one more and then

我再问一个,然后

until all recommener engines are AI based. until all content generation is AI based because content generation consumer oriented content generation is very largely recommener systems and so on so forth um and all of that's going to be AI generated until until all of the stuff what classically was hypers scale now transitions to AI you know everything from shopping to e-commerce to you know all that stuff until everything goes over

要等到所有推荐引擎都跑在 AI 上,要等到所有内容生成都基于 AI —— 因为内容生成、面向消费者的内容生成,很大程度上就是推荐系统那一套,而这些全都会变成 AI 生成;要等到过去叫做超大规模的那一整套全部转向 AI,从购物到电商,所有那些;要等到一切都转过去。

because but all this new build right when we're talking about trillions.

但所有这些新建设 —— 我们谈的是几万亿。

We're investing ahead of where we are.

我们是超前于现状在投资。

Um, you know, is that like at will?

这算是可以随时收放的吗?

Are you obliged to invest the money even if you see a slowdown or a kind of a glut coming or is this one of these things that you're just waving the flag to the ecosystem to say get out and build and at some point in time if we see some of this slow down, we can always pull back on the level of investment.

就算你看到增速放缓、或者看到产能过剩要来了,你也非投不可吗?还是说更像这样:你只是在向整个生态摇旗,让大家出去建,而到了某个时候,如果我们看到有些地方放慢了,随时可以把投资力度收回来?

Actually, it's the other way because we're at the end of the supply chain, right?

其实反过来 —— 因为我们在供应链的末端,对吧?

And so, we respond to demand.

所以我们是响应需求。

Okay?

好吧?

And right now all the VCs will tell you and you guys know

现在所有 VC 都会告诉你,你们俩也知道

the demand the short there's a shortage of compute in the world not because there's a shortage of GPUs in the world.

需求 —— 短缺 —— 世界上算力短缺,并不是因为世界上 GPU 短缺。

Okay, if they give me an order I'll build it.

好,他们给我订单,我就造。

Mhm.

嗯。

Right.

没错。

We've over the last couple years we've really plumbed the supply chain.

过去这两年,我们真把整条供应链打通了。

So all of the supply chain behind me from wafer starts to co-ass HBM memories you know all of that technology.

所以我身后这一整条供应链,从晶圆投片到 CoWoS、HBM 内存,所有这些技术。

We've really geared up.

我们真的把产能拉起来了。

Yeah.

嗯。

If we need to double we'll double.

需要翻倍,我们就翻倍。

Yes.

是的。

Okay.

好。

So the supply chain is ready.

所以供应链已经准备好了。

Now we're just waiting for demand signals and when when uh the the CSPs and the hyperscalers and our customers uh do their annual plan and they give us you know their forecast um we respond to that and we build to that.

现在就等需求信号 —— CSP、超大规模厂商和我们的客户做年度计划,把预测给我们,我们照着响应、照着生产。

Now what's what's going on of course is that every one of their forecasts that they provide us turns out to have been wrong

而现在的情况当然是,他们给我们的每一份预测,最后都证明是错的

right

没错

because they under forecasted

因为他们预测得太低了

and so now we're always in a scramble mode.

所以现在我们一直在手忙脚乱地赶。

Mhm.

嗯。

And so we've been in the scramble mode now for you know a couple of years

这种手忙脚乱的状态,我们已经持续了好几年

and it's whatever forecast we've been given has been always

而不管给我们的预测是什么,总是

significant increase from last year but not not enough.

比去年大幅增长,但还是不够。

Satcha last year seemed to be pulling back a little bit.

Satya Nadella 去年好像收了一点手。

You know seemed to be you know some people called him the adult in the room tamping down kind of some of these these expectations.

有人把他叫作「房间里的成年人」,说他在给这些预期降温。

A few weeks ago he said hey I've also built two gigs this year and we're going to accelerate in the future.

几周前他说,嘿,我今年也建了 2 吉瓦,而且未来还要加速。

Do you see some of the traditional hyperscalers that may have been moving a little slower than let's call it a coreweave or or or Elon X or maybe a little slower than Stargate?

你有没有看到,那些传统超大规模厂商 —— 原本可能比 CoreWeave、比 Elon 的 xAI 走得慢一点,或者比 Stargate 慢一点的那几家?

Do you see them all?

你看到他们全都 ——

It it sounds like to me they're all leaning in more now and they're all also

我听着像是,他们现在全都在加码,而且全都还

because of the second exponential.

因为第二条指数曲线。

Okay.

好。

We've already had one exponential we were experiencing which was the adoption rate of AI, the engagement of AI was growing exponentially.

我们已经经历过一条指数曲线,就是 AI 的采用速度 —— AI 的使用在指数级增长。

Yes.

是的。

The second exponential that kicked in was reasoning.

而第二条开始发力的指数曲线,是思考。

Yeah.

嗯。

That was the conversation we had one year ago.

那正是我们一年前聊过的。

One year ago.

一年前。

Yeah.

嗯。

We said, "Hey, listen.

我们当时说,嘿,听着。

The moment you take AI from one shot,

一旦你把 AI 从一次成型(one-shot)带出来,

memorizing an answer and Right.

记住一个答案 —— 对。

Memorizing and generalizing, that's basically pre-training."

记住,然后泛化,这基本上就是预训练。

Yeah.

嗯。

So memorizing an answer, you know what's 8* 8?

所以说,记住一个答案 —— 8 乘 8 等于多少?

Just memorize it.

背下来就行。

Okay.

好。

And so memorizing an answer and generalizing, that was one shot AI.

所以记住答案加上泛化,那就是一次成型(one-shot)的 AI。

Now a year ago, reasoning came about

而一年前,思考出现了

for sure

确实

research came about, tool use came about, and now you're a thinking AI

研究出现了,工具调用出现了,现在你手上是一个会思考的 AI

1 billion X.

1 billion 倍。

It's going to use a lot more compute.

它会吃掉多得多的算力。

Certain hypers scale customers to your point had internal workloads that they had to migrate anyways from um from general purpose computing to accelerated computing.

按你说的,某些超大规模客户本来就有内部工作负载,反正都得从通用计算迁到加速计算。

So they built through the cycle.

所以他们是穿越周期在建。

I think maybe some hyperscalers had different workloads so they weren't quite sure how quickly they could digest it but everyone

我猜有些超大规模厂商的负载不一样,所以他们没那么确定自己能多快消化掉,但每个人

has now concluded that they dramatically underbuilt.

现在都得出了结论:自己建得远远不够。

One of the applications that my favorite is just good oldfashioned data processing

有一类应用是我最喜欢的,就是老老实实的数据处理

structured data and unstructured data.

结构化数据和非结构化数据。

Just good oldfashioned data processing.

就是老老实实的数据处理。

And very soon we're going to announce a very big initiative of accelerated data processing. M

很快我们就会宣布一项非常大的加速数据处理计划。

data processing represents the vast majority of the world's CPUs today.

今天全世界的 CPU,绝大部分都在做数据处理。

It still completely runs on CPUs.

到现在还是完全跑在 CPU 上。

You know, if you go to data bricks, it's mostly CPUs.

你去看 Databricks,大部分是 CPU。

You go to snowflakes, mostly CPUs.

你去看 Snowflake,大部分是 CPU。

Uh SQL processing at Oracle, mostly CPUs.

Oracle 的 SQL 处理,大部分是 CPU。

Everybody's using CPUs to do SQL structured data.

所有人都在用 CPU 处理 SQL 结构化数据。

In the future, that's all going to move to AI data.

将来这些全都会转到 AI 数据上。

That is one gigantic massive market

那是一个极其巨大的市场

that we're going to move to.

是我们要进去的市场。

But you need you need a you need everything that Nvidia does requires acceleration layers and requires you know

但你需要 —— NVIDIA 做的每一件事都需要加速层,都需要

domain specific right data processing lies recipes we got to go build that but that's coming

领域专用的,没错 —— 数据处理库、配方,这些我们得去建,但它们就要来了

so one of the push backs I you know I turned on CNBC yesterday they were like oh glut bubble when I turned on Bloomberg it was about roundtpping and circular revenues okay and so for the benefit of people you know at home know these arrangements are when companies enter into a misleading transaction ction that artificially inflates revenue without any underlying economic substance.

所以有一种反驳是 —— 我昨天打开 CNBC,他们在讲产能过剩、讲泡沫;打开 Bloomberg,讲的是循环交易和循环收入。为了让在家看的人听明白:这类安排指的是,公司之间做一笔带误导性的交易,人为把营收吹起来,底下没有任何真实的经济实质。

So in other words, growth propped up by financial engineering, not by customer demand.

换句话说,增长是靠财务工程撑起来的,不是靠客户需求。

And the canonical case everybody's referencing of course is Cisco and Nortell from the last bubble 25 years ago.

而所有人引用的经典案例,当然是 25 年前上一轮泡沫里的 Cisco 和 Nortel。

So when you guys or Microsoft or Amazon are investing in companies that are also your big customers, in this case you guys investing in Open AI while Open AI is buying tens of billions of chips, just remind us and remind everybody else like what is it what are the analysts on Bloomberg and otherwise getting wrong when they're hyperventilating about circular revenues or about roundtpping?

所以当你们、或者 Microsoft、Amazon 去投资那些同时也是你们大客户的公司时 —— 这里就是你们投 OpenAI,而 OpenAI 在买几百亿美元的芯片 —— 请给我们、也给所有人再讲清楚一次:Bloomberg 上那些分析师和其他人为循环收入、为循环交易大呼小叫的时候,他们到底错在哪儿?

10 gigawatts is like $400 billion, right?

10 gigawatts 差不多就是 $400 billion,对吧?

Something like that. and and that $400 billion dollars will have to be largely funded by their offtake,

差不多这个数。而这 $400 billion,大头得靠他们的承购(offtake)来筹,

right?

对吧?

Their revenue

也就是他们的营收——

which is growing exponentially.

而这营收正在指数级增长。

It has to be funded by their capital, the money they've raised through equity and whatever debt they can raise.

还得靠他们自己的资本:股权融来的钱,再加上他们能借到的所有债。

Those are the three vehicles.

融资渠道就这三条。

And the equity that they could raise and the debt that they could raise has something to do with the confidence of the revenues that they could sustain

而他们能融到多少股权、能借到多少债,跟外界对他们能不能持续做出这些营收有多少信心有关。

for sure.

这是肯定的。

And so smart investors and smart lenders will consider all of these factors.

所以精明的投资人、精明的放贷方,会把这些因素全考虑进去。

Fundamentally, that's what they're going to do.

说到底,他们就是会这么做。

That's their company.

那是他们的公司。

It's not my business.

不是我的生意。

And of course, we have to stay very close to them to make sure that we build in support of their continued growth.

当然,我们得跟他们贴得很紧,确保我们建的东西是在支撑他们继续增长。

Okay?

明白吧?

And so, um, there's the revenue side of it and has nothing to do with the investment side of it.

所以这里面有营收这一边,它跟投资那一边毫无关系。

The investment side of it is not tied to anything.

投资这一边不跟任何东西挂钩。

It's an opportunity to invest in them.

这就是一个投资他们的机会。

And as we were mentioning earlier, this is likely going to be the next multi- trillion dollar hypers scale company.

而就像我们前面说的,这很可能会成为下一家 multi-trillion dollar 级的超大规模公司。

And who doesn't want to be an investor in that?

谁不想成为它的投资人?

You know, my only regret is that that they invited us to invest early on.

我唯一的遗憾是——他们当年很早就邀请我们投资,

I remember those conversations

我记得那几次谈话。

and we were so poor, you know, that we were so poor, we didn't invest enough, you know, and I should have given them all my money.

而我们当时太穷了,真的太穷,投得不够多,我真该把我全部的钱都给他们。

And the reality is if you guys don't do your jobs and keep up with, you know, if if Ver Rubin doesn't turn into a good chip, they can go get other chips and put them in these data centers,

而现实是,如果你们不好好干、跟不上——如果 Vera Rubin 没做成一颗好芯片,他们完全可以去拿别的芯片装进这些数据中心,

right?

没错。

There's no obligation that they have to use your chips and and like you said, you're you're looking at this as an opportunistic equity investment.

他们并没有义务非用你们的芯片,而且像你说的,你把这看成一笔机会型的股权投资。

The other thing I would say

我还想说一点——

and we've made some great investments.

我们确实投出过一些很棒的项目。

I got to put it out there, you know, we invested in XAI, we invested in Corewave.

这我得说出来:我们投了 xAI,我们投了 CoreWeave。

Incredible.

太厉害了。

Yeah.

是啊。

Yeah.

是啊。

How smart was that?

这多聪明啊?

Yeah.

对。

As I go back to this, the other fundamental thing it seems to me is, you know, you're putting it out there.

回到这个话题,我觉得另一个根本点是,你把话摆在台面上说了。

You're saying this is what we're doing.

你说,这就是我们在做的事。

And the underlying economic substance here, right?

还有这背后真实的经济实质,对吧?

It's not that you're just some somehow sending revenues back and forth between the two companies.

并不是说你只是在两家公司之间来回倒营收。

We got people sending money every month for Chat GPT, a billion and a half monthly users using the product.

有人每个月真金白银为 ChatGPT 付钱,每月 15 亿用户在用这个产品。

You just said every enterprise in the world is either going to do this or they will die.

你刚才说了,世界上每一家企业要么做这件事,要么就得死。

Every sovereign views this as existential to their national security and economic security as as nuclear power.

每个主权国家都把这看成跟核力量一样、攸关国家安全和经济安全的生死问题。

What person, company or nation says intelligence is bas basically optional for yeah for us.

哪个人、哪家公司、哪个国家会说,智能对我们基本上可有可无?

I mean it's fundamental to them.

这对他们是根本性的。

Well the automation of intelligence

就是智能的自动化——

I beat the demand question to death.

需求这个问题,我算是问到底了。

So let's jump in a little bit to system design and I'm I'm I'm going to turn to Clark here in a sec second on that.

那我们进到系统设计这块,这部分我一会儿就交给 Clark。

But in 2024, you switched to your annual release cycle,

但在 2024 年,你们切换到了年度发布节奏,

right, with Hopper.

对,从 Hopper 开始。

You then had a massive upgrade which required, you know, significant data center overhaul with Grace Blackwell.

之后你们又做了一次大升级——Grace Blackwell,数据中心得跟着大改一遍。

In 2025, and in the back half of 26, we're going to get Vera Rubin.

那是 2025 年,26 年下半年我们会拿到 Vera Rubin。

27 we'll get Ultra and 28 Fineman.

27 年是 Ultra,28 年是 Feynman。

How is the annual release cycle going?

年度发布节奏推进得怎么样?

Okay.

好。

What were the main goals of going to an annual release cycle?

改成年度发布节奏,主要目标是什么?

And did AI inside Nvidia allow you to execute the annual release cycle?

还有,NVIDIA 内部用上 AI,是不是才让你们跑得动这个年度发布节奏?

Yeah, the answer is yes.

对,答案是肯定的。

On the back on the last question, without it, um, Nvidia's velocity, our pace, our scale would be limited.

先回答最后一个问题:没有 AI,NVIDIA 的速度、节奏、规模都会受限。

And so without without AI these days, um, it's just simply not possible to build what we built.

所以今天没有 AI,我们已经造出来的这些东西,根本造不出来。

Now, um, why do we do it?

那我们为什么要这么做?

There's something that remember uh uh Eddie said it uh at his earnings call um or his conference.

记得吗,吴泳铭在财报电话会、或者他自己的大会上讲过一件事。

Uh Satia has said it, Sam has said it.

Satya Nadella 说过,Sam Altman 也说过。

The token generation rate is going up exponentially.

token 生成的速率正在指数级上升。

Yeah.

对。

And the customer use is going up exponentially.

客户的使用量也在指数级上升。

Um I think they're at 800 million weekly active users or something like that.

我记得他们已经到了 800 million 周活跃用户上下。

Yes.

是。

I mean that's less than two years from chat GPT, right?

距离 ChatGPT 出来还不到两年,对吧?

And each of those users is generating massively more tokens because they're using inference time reasoning.

而且每个用户生成的 token 也多得多,因为他们用的是推理阶段的思考。

That's right.

没错。

Exactly.

正是。

And so so the first thing is because the token generation rate is going up so incredibly two exponentials.

所以第一点是,token 生成速率涨得实在太猛,是两个指数。

Yeah.

对。

On top of each other,

叠在一起,

we have to un unless we increase the performance at incredible rates,

我们必须——除非我们以惊人的速率把性能提上去,

the cost of token generation will keep growing because Mo's law is dead, right?

token 生成的成本就会一直往上涨,因为摩尔定律已经死了,对吧?

Because transistors basically cost the same every single year now.

现在晶体管的成本每年基本没变。

And power is largely the same.

功耗也基本没变。

And between those two fundamental laws, unless we come up with new technologies to drive the cost down, even if there's a slight difference in growth, you give somebody a discount of a few percent.

卡在这两条基本定律之间,除非我们拿出新技术把成本压下来——哪怕增长上有一点差别,你给人家打个百分之几的折,

How's that going to make up for two exponentials?

那怎么补得上两个指数?

And so we have to increase our per performance annually at a pace that keeps up with that exponential.

所以我们每年都得把性能提上去,速度要跟得上那条指数曲线。

Yeah.

对。

So in the case of in the case of um uh going from uh I guess uh uh Kepler to yeah all the way to Kepler all the way to um uh Hopper was probably 100,000x

那比如说,从 Kepler 一路到 Hopper,大概是 100,000x,

that was the beginning of the AI journey for Nvidia.

那就是 NVIDIA AI 之旅的起点。

100,000x in 10 years.

10 年 100,000x。

Okay.

好。

Between Hopper and Blackwell, we increased because of MVLink 72,

从 Hopper 到 Blackwell,靠 NVLink 72,我们又提升了

right?

对吧?

30x

30x,

in one year and then we'll get another X factor again with Reuben and then we'll get another X factor with Fineman.

就在一年之内。到 Rubin 还会再拿到一个倍数,到 Feynman 又是一个倍数。

And the way we we do that is because the transistors aren't really helping us very much, right?

我们能做到,是因为晶体管其实帮不上多少忙,对吧?

Moors law is largely the density is growing up but going up but the performance is not.

摩尔定律那边,密度还在涨,性能没跟上。

And so if that's the case, one of the challenges that we have to do is we have to break the entire problem down at the system level and change every chip at the same time and all the software stack and all the systems all at the same time.

既然如此,我们要面对的挑战之一,就是在系统层面把整个问题拆开,所有芯片同时换代,整个软件栈、所有系统也同时换代。

The ultimate extreme code design.

这是终极的极致协同设计。

Nobody's ever codees at this level before,

以前没人在这个层级上做过协同设计,

right?

对吧?

We change the CPU, revolutionize the CPU, a GPU, the networking chip, the MVLink scale up, the Spectrum X scale out.

我们换掉 CPU、把 CPU 彻底革新,还有 GPU、网络芯片、NVLink 的 scale up、Spectrum-X 的 scale out。

Somebody said I heard somebody said, "Oh yeah, it's just Ethernet.

有人说——我听过有人说:「哦,那不就是以太网嘛。」

Yeah.

是啊。

Right.

没错。

Okay.

行吧。

So, Spectrum X Ethernet is not just Ethernet.

可 Spectrum-X 以太网不只是以太网。

And people are starting to discover, oh my god, the X factors is pretty incredible,

大家开始意识到,天哪,这个倍数相当惊人,

right?

对吧?

You know, Nvidia's Ethernet business, the just Ethernet business is the fastest growing Ethernet business in the world.

NVIDIA 的以太网业务,那个「不就是以太网」的业务,是全世界增长最快的以太网业务。

Yeah.

对。

And so, so scale out and of course now we have to uh build even larger systems.

所以有了 scale out;当然,现在我们还得造更大的系统。

So, we scale across

于是我们做 scale across,

um multiple AI factories connected together.

把多座 AI 工厂连起来。

And then we do this at an annual pace.

而且这件事我们一年做一轮。

And so we now have an exponential of exponentials going ourselves from technology.

所以现在我们自己在技术上也跑出了一个指数之上的指数。

And that allows our customers to drive the cost of tokens down.

这让客户能把 token 的成本压下去,

Keep making those tokens smarter and smarter with pre-training and post-training and thinking.

同时靠预训练、后训练和思考,把这些 token 变得越来越聪明。

And as a result, when the AI gets smarter, they get more used.

结果是,AI 越聪明,用的人就越多。

When they get more used, they're going to grow exponentially.

用得越多,增长就越呈指数级。

For people who may not be as familiar.

可能有人还不太熟悉——

Yeah.

嗯。

What is extreme code design?

极致协同设计到底是什么?

Extreme code design means that you have to optimize the model, algorithm, system, and chip at the same time.

极致协同设计的意思是,你得同时优化模型、算法、系统和芯片。

You have to innovate outside the box,

你必须跳出框框去创新,

right?

对吧?

Because Moore's law said you just have to keep making the CPU faster and faster.

因为摩尔定律讲的是,你只要不停把 CPU 做得更快就行。

Everything got faster.

一切都跟着变快。

You were innovating within a box.

那时你是在框框里创新。

Just make that chip faster.

把那颗芯片做快就行。

Yeah.

对。

Well, if that chip doesn't go any faster, then what are you going to do?

可这颗芯片再快不上去了,你怎么办?

Innovate outside the box.

跳出框框去创新。

And so Nvidia really changed things because we did two things.

NVIDIA 真正改变了局面,因为我们做了两件事。

We invented CUDA, invented GPUs,

我们发明了 CUDA,发明了 GPU,

and we invented the idea of code design at a very large scale.

还发明了在极大规模上做协同设计这件事。

That's why there's all these industries we're in.

这就是我们为什么进了这么多行业。

We're creating all these libraries and code design.

我们做出这么多库,做了这么多协同设计。

Number one, full stack extreme is even beyond software and GPUs.

第一,全栈的极致早已不止软件和 GPU。

It's now at the data center level switches and networking and you know all of that all of that software in the switches and the networking and the nicks the scale up the scale out

现在做到了数据中心层级:交换机、网络,交换机和网络里跑的所有软件,还有网卡、scale up、scale out,

optimizing across all of that as a result of that blackwell to hopper is 30x

把这一切放在一起优化,结果就是 Blackwell 相对 Hopper 提升 30x。

no moors law could possibly achieve that

摩尔定律绝无可能做到。

right

对吧?

and so that's extreme

所以这才叫极致。

and that comes from the extreme code design

而它来自极致协同设计。

that's because Nvidia has that's why we got into networking and switching and scale up and scale out and scale across and building CPUs and building GPUs and building nicks.

这也是 NVIDIA 为什么进网络、进交换,做 scale up、scale out、scale across,自己做 CPU、做 GPU、做网卡。

You know that that's the reason why Nvidia is so rich in software and people we we check in more open-source software in the world than just about anybody except one other company.

这也是为什么 NVIDIA 在软件和人上这么厚——我们提交的开源软件比世界上几乎任何一家都多,只有一家比我们多。

I think it's AI2 or something like that.

我记得是 AI2 之类的。

And so so we have such enormous richness of software and that's just in AI.

所以我们的软件积累极其丰厚,而这还只算 AI。

Don't forget computer graphics and digital biology and autonomous vehicles and you know the amount of software we produce as a company is incredible that allows us to do deep and extreme code designs.

别忘了计算机图形、数字生物学、自动驾驶——我们这家公司产出的软件量惊人,正是这一点让我们能把协同设计做到又深又极致。

I heard from one of your competitors you know yes he's doing this because it helps drive down the cost of token generation but at the same time your annual release cycle makes it almost impossible for your competitors to keep up. um the supply chain gets locked up more because you're giving three-year visibility to your supply chain.

我从你们的一位竞争对手那儿听到这么个说法:是的,他这么干是因为能压低 token 生成的成本,但同时,你们的年度发布节奏让竞争对手几乎追不上。供应链也被锁得更死,因为你们给了供应链三年的可见性。

So now the supply chain has confidence as to what they can build to.

现在供应链心里有底,知道该按什么量去建。

So do you think about this?

那你是这么想的吗?

Wait wait wait before you before you ask the question.

等等等等,你先别急着问——

Think about this.

想想这个。

In order for us to do several hundred billion dollars a year

我们一年要做到几千亿美元

of AI infrastructure buildout.

的 AI 基础设施建设,

Yes.

是。

Think about how much capacity we had to go start a year ago.

你想想一年前我们就得启动多少产能。

Yes.

是。

We're talking about building hundreds of billions of dollars of

我们说的是要建起几千亿美元的

wafer starts and DRAM buys and are you you guys talking

晶圆投片和 DRAM 采购——你们说的是这个吧——

Yeah.

对。

This is now at a scale that hardly any company can keep up with.

这个规模,现在几乎没有哪家公司跟得上。

So would you say your competitive moat is greater today than it was three years ago?

那你会说,今天你们的护城河比三年前更深了吗?

Yeah.

对。

You know, first of all, there's just more competition than ever before, but it's harder than ever before.

首先,竞争比以往任何时候都激烈,但难度也比以往任何时候都大。

Mhm.

嗯。

And the reason why I say that is because um wafer cost is getting wafer costs are getting higher

我这么说,是因为晶圆成本在往上走,

which means that unless you do code design at an extreme scale you're just not going to be able to deliver the X factor growth

这意味着,除非你把协同设计做到极致规模,否则根本交不出那种倍数级的增长。

number one number and so you you know unless you unless you're working on six seven eight chips a year

这是第一点。所以说,除非你一年在做六颗、七颗、八颗芯片——

right

对吧?

that's amazing thing it's not about building an ASIC it's about building an AI factory

了不起的地方就在这儿:关键不是造一颗 ASIC,而是造一座 AI 工厂。

system

一整套系统。

and this system that has a lot of chips in and they're all co-designed

这套系统里有很多颗芯片,全都是协同设计出来的。

and together they deliver that you know that 10x factor that we get

它们合在一起,才交付出我们那个 10x,

uh almost regularly.

而且几乎成了常态。

Okay.

好。

So number one the code design is extreme.

所以第一,协同设计是极致的。

The second thing is that the scale is extreme.

第二,规模是极致的。

When your customers deploy a gigawatt that's a you know 400,000 500,000 GPUs right

客户部署 1 个 gigawatt,那就是 400,000、500,000 颗 GPU,对吧?

getting 500,000 GPUs to work together is a miracle.

让 500,000 颗 GPU 协同工作,是个奇迹。

I mean it's just a miracle.

真的就是奇迹。

And so they your customers are taking enormous risk on you to go buy all of this.

所以客户要买下这一切,是在你身上承担了巨大的风险。

You got to ask yourself what customer would place a $50 billion PO

你得问问自己,哪个客户会下一张 $50 billion 的订单(PO),

on an architecture,

押在一个架构上,

right?

对吧?

On an unproven architecture, a new one,

押在一个还没验证过的架构上,一个全新的架构,

right?

对吧?

A new architecture.

一个新架构。

Yeah.

对。

You just take out a whole new chip.

你才刚把一颗全新的芯片流片出来。

You're as excited as you are about it, you know, and everybody's excited for you

你自己兴奋得不行,大家也都替你高兴,

and and you just show the first silicon,

而你才刚拿出首批硅片,

right?

对吧?

Who's going to give you $50 billion PO, right?

谁会给你 $50 billion 的订单(PO)?对吧?

And why would you start $50 billion worth of wafers for a chip that just taped out?

又为什么要为一颗刚流片的芯片启动 $50 billion 的晶圆投片?

But for Nvidia, we could do that because our architecture is so proven.

但 NVIDIA 可以,因为我们的架构经过了充分验证。

So number the the scale of our customer is so incredible.

所以第一,我们客户的规模惊人。

Now the scale of our supply chain is incredible,

现在供应链的规模也惊人,

right?

对吧?

Who's going to start all of that stuff, pre-build all of that stuff for a company unless they know that Nvidia can deliver through?

谁会为一家公司把这些全都启动、全都预先备好?除非他们知道 NVIDIA 一路交付得出来。

Isn't that right? and they believe that we can we can deliver through to all of the customers around the world.

对不对?他们相信我们能一路交付给全世界所有客户。

They're willing to start several hundred billion dollars at a time.

他们才愿意一次就启动几千亿美元的量。

This is just the scale is incredible.

这个规模真的惊人。

To that point, you know, one of the biggest key debates and controversies in the world is this question of GPUs versus A6,

说到这一点,当今世界上最关键、争议最大的辩论之一,就是 GPU 与 ASIC 之争,

Google's TPUs, Amazon's Tranium, and it seems like everyone from ARM to OpenAI to Enthropic are, you know, rumored to be building one.

谷歌的 TPU、亚马逊的 Trainium,而且看起来从 ARM 到 OpenAI 到 Anthropic,人人都有传闻要自己做一颗。

Mhm.

嗯。

Last year you said, you know, we're building systems, not chips, and you're driving performance through every single part of that stack.

去年你说过,我们做的是系统,不是芯片,你们是在整个技术栈的每一个环节上推性能。

You also said that many of these projects may never get to production scale.

你还说过,这些项目里很多可能永远做不到量产规模。

But given like the

但考虑到

most of them,

大多数都不会。

given the seeming success of Google's TPUs,

考虑到谷歌的 TPU 看起来相当成功,

yeah,

对,

you know, how are you thinking about this evolving landscape today?

你今天怎么看这个正在演变的格局?

Yeah.

对。

And

而且

first of all,

首先,

Mhm. the the advantage that that Google had is foresight.

谷歌的优势是有先见之明。

Remember, they started TPU1 before everything started.

记住,他们在这一切开始之前就上了 TPU1。

You know, this is no different than a startup.

这跟创业公司没什么两样。

You're supposed to build a startup.

你应该去做一家创业公司。

You're supposed to create a startup before the market grows.

你应该在市场长大之前就把这家公司做出来。

Mhm.

嗯。

You're not supposed to come up as a startup when the market's a trillion dollars large.

你不该等市场已经有 $1 trillion 那么大了,才作为一家创业公司出场。

Mhm.

嗯。

You know the this fallacy and and all VCs know this this fallacy that a large market if you could just take a few percent market share you could be a giant company.

有一个谬误,所有 VC 都懂这个谬误 —— 说市场足够大,你只要拿下几个百分点的份额,就能变成一家巨头公司。

That's actually fundamentally wrong.

这其实根本上就是错的。

Yeah.

对。

You're supposed to take 100% of a tiny company, a tiny industry, which is what Nvidia did did, right?

你应该拿下一家小公司、一个小行业的 100%,NVIDIA 当年就是这么干的,对吧?

Which is what TPUs did.

TPU 也是这么干的。

There were only the two of us.

当时就只有我们两家。

But you better hope that that industry gets really big.

但你最好祈祷那个行业真的会变得很大。

You're creating an industry.

你是在创造一个行业。

That's right.

没错。

Right. and and I mean the Nvidia story you know which

对。我是说 NVIDIA 的故事,就是

and so that's the challenge for the for people who are building A6 now it looks like a juicy market

所以这就是现在做 ASIC 的人面临的挑战 —— 眼下看着是个肥美的市场,

but remember this juicy market has evolved from a chip called a GPU to I just described an AI factory and you guys just saw I just announced a chip called CPX for context processing and diffusion video generation a very specialized workload but an important workload inside a data center I just prelude it to maybe AI data processing processors because guess what?

但记住,这个肥美的市场已经从一颗叫 GPU 的芯片,演变成我刚描述的那种 AI 工厂;你们也刚看到,我发布了一颗叫 CPX 的芯片,做上下文处理和 diffusion 视频生成,是一类非常专门的负载,但在数据中心里是很重要的负载;我刚才还预告了一下,也许还会有做 AI 数据处理的处理器 —— 因为你猜怎么着?

You need long-term memory.

你需要长期记忆。

You need short-term memory.

你需要短期记忆。

The KV cache processing is really intense.

KV cache 的处理量非常大。

AI memory is a big deal.

AI 的记忆是件大事。

You know, you kind of like your AI to have good memory and just dealing with all the KV caching around the system.

你总希望自己的 AI 记性好一点,而光是在整个系统里处理 KV cache,

Really complicated stuff.

就已经相当复杂了。

Maybe it wants to have a specialized processor.

也许它需要一颗专门的处理器。

Maybe there's other things, right?

也许还有别的东西,对吧?

So you you see that Nvidia's our our our viewpoint is now not GPU.

所以你看,NVIDIA 我们今天的视角不是 GPU。

Our viewpoint is looking at the entire AI infrastructure and what does it take for these incredible companies to get all of their workload through it which is diverse and changing.

我们的视角是看整个 AI 基础设施 —— 这些了不起的公司要把自己所有的负载跑通,需要什么;而这些负载既多样又一直在变。

Look at the transformer.

看看 transformer。

The transformer architecture is changing incredibly.

transformer 架构正在以惊人的速度变化。

If not for the fact that CUDA is easy to operate on and iterate on, how do they try all of their vast number of experiments to decide which one of the transformer versions, what kind of attention algorithm to use?

要不是 CUDA 好上手、好迭代,他们怎么跑得完那么多实验,去决定用哪个版本的 transformer、用哪种注意力算法?

How do you disagregate?

你怎么做解耦?

CUDA helps you do all that because it's so programmable.

这些都是 CUDA 帮你做到的,因为它可编程性太强。

And so the way to think about our our our business now is you look at when when all of these ASIC companies or ASIC projects start 3, four, five years ago, I got to tell you that industry was super adorable and simple.

所以现在看我们这门生意,你要回头看:这些 ASIC 公司、ASIC 项目是三年、四年、五年前启动的,我得说,那时候这个行业又可爱又简单。

There was a GPU involved,

里面就涉及一颗 GPU,

right?

对吧?

But now it's giant and complex

但现在它又大又复杂,

and in another two years it's going to be completely massive.

再过两年会彻底变成庞然大物。

The scale is going to be so large.

规模会大到不行。

And so I think that the battle the the battle of getting into

所以我认为,要挤进

a very large market as a nent player is just hard you know as you guys know

一个非常大的市场,还是个刚起步的玩家,这件事就是难 —— 你们也清楚。

even for the customers who perhaps are successful with AS6.

就算是那些用 ASIC 或许真做成了的客户 ——

Yeah.

对。

Isn't there an optimal balance in their compute fleet like it's you know I think investors are very much binary creatures.

在他们的算力机队里,难道不存在一个最优配比吗?我觉得投资人是非常二元的生物。

They just want a yes or no black and white answer.

他们只想要一个是或不是、非黑即白的答案。

But even even if you get the ASIC to work, isn't there an optimal balance because you think I'm buying the Nvidia platform, CPX is going to come out for prefill for, you know, for video generation,

但就算你把 ASIC 做成了,难道不存在一个最优配比吗?因为你会想,我买的是 NVIDIA 平台,CPX 会出来做 prefill、做视频生成,

maybe a decode, you know, you know, a platform video.

也许还有做 decode 的,一个平台,视频。

Exactly.

正是。

Yeah.

对。

So there will be like many different

所以会有很多不同的

um chips or parts to add to the Nvidia ecosystem,

芯片或者部件加进 NVIDIA 生态,

accelerated compute fleet,

加进这支加速计算机队,

right? as new workloads are, you know, are are born.

对吧?随着新负载不断出现。

That's right.

没错。

And, you know, people trying to tape out new chips today are not really anticipating what's happening a year from now.

而今天那些想流片做新芯片的人,其实并没在预判一年之后会发生什么。

They're just trying to get a chip to work.

他们只是想先让一颗芯片跑起来。

That's right.

没错。

Set another way, Google's a big GPU customer.

换个说法,谷歌是 GPU 的大客户。

Google's a big GPU customer.

谷歌是 GPU 的大客户。

Um, if you look at and Google is a very special case.

你去看,谷歌是个非常特殊的例子。

I mean, we just have to, you know, show respect where respect is really deserved.

该给的尊重就得给。

I mean,

我是说,

uh, TPU is on TPU7.

TPU 已经做到 TPU7 了。

Yes.

是的。

Right.

对。

And so um and and uh and it's a challenge for them as well, right?

所以这对他们来说也是个挑战,对吧?

And so so the work that they do is incredibly hard.

他们做的工作难度极高。

So so I think the first thing to to let let me do a um you know remember there there are three categories of chips.

所以我想先说第一件事 —— 记住,芯片分三类。

There's the category chips that are architectural.

有一类芯片是架构级的。

X86 CPUs, ARM CPUs, Nvidia GPUs.

X86 CPU、ARM CPU、NVIDIA GPU。

Architectural And it has an ecosystem above and and um uh the architecture uh allows uh has rich IP and rich ecosystem very complicated technology.

架构级的,上面顶着一整个生态;架构本身有丰富的 IP 和丰富的生态,技术非常复杂。

It's built by the owners like us.

它由像我们这样的架构拥有者自己来造。

Okay.

好。

There's AS6.

然后是 ASIC。

I worked for the original company LSI Logic who invented the idea of A6.

我在最早那家公司 LSI Logic 干过,ASIC 这个想法就是他们发明的。

As you know LSI Logic is not here anymore,

你也知道,LSI Logic 已经不在了,

right?

对吧?

And the reason for that is because AS6 is really fantastic.

原因就在于,ASIC 真的很棒 ——

When the the market size is not very large,

在市场规模还不大的时候。

um it's easy to have somebody uh be a contractor to help you put the packaging of all that stuff together and do the manufacturing on your behalf and they charge you 50 60 points of margin.

那时候很容易找一家承包商,帮你把封装那些活儿都拼起来,替你做制造,他们向你收 50、60 个点的毛利。

But when the market gets large for an ASIC, there's a new way of doing things called coot,

但等 ASIC 的市场做大了,就出现了一种新做法,叫 COT ——

customerowned tooling.

客户自有流片。

And who would who would do something like that?

谁会这么干?

Um Apple's as Apple Apple's smartphone chip the volume is so large they would never go pay somebody else 50 60% gross margin to be an ASIC.

苹果 —— 苹果的智能手机芯片量太大了,他们绝不会为了做一颗 ASIC 去付给别人 50%、60% 的毛利。

They do customer own tooling.

他们做的就是客户自有流片。

And so um where will TPUs go when they when it becomes a large business?

那么等 TPU 变成一门大生意,它会走向哪里?

Customer own tooling.

客户自有流片。

There's no question about it.

这一点毫无疑问。

And so a but there's a place for A6.

但 ASIC 仍然有它的位置。

Uh video transcoders will never be too large.

视频转码器永远大不到哪儿去。

Um, smart nicks will never be too large.

智能网卡也永远大不到哪儿去。

And so when when there's 10, 12, 15 as projects going on at an ASIC company, I'm not surprised by that,

所以一家 ASIC 公司同时有 10 个、12 个、15 个 ASIC 项目在跑,我一点都不意外,

you know, because there there probably five smart nicks and four transcoders and, you know, are they all AI chips?

因为里面大概有五个智能网卡、四个转码器 —— 它们都是 AI 芯片吗?

Of course not.

当然不是。

You know, and and if somebody were to build an embedded embedding processor for a specific recommener system and that was an ASIC, of course you could do that.

如果有人为某个特定的推荐系统做一颗 embedding 处理器,而那是一颗 ASIC,当然可以那么干。

But would you do that as the fundamental compute engine for AI that's changing all the time?

但 AI 一直在变,你会拿它当 AI 的根本计算引擎吗?

You've got low latency workload.

你有低延迟的负载。

You got high throughput lo workload.

你有高吞吐的负载。

You have token generation for chat.

你有聊天的 token 生成。

You have thinking workload.

你有思考的负载。

You have AI video generation workload.

你有 AI 视频生成的负载。

Is there a you know now you're talking about

有没有 —— 现在你说的是

the workhorse backbone of your accelerated?

你那套加速计算的主力骨干?

That's what Nvidia is all about.

这正是 NVIDIA 的全部意义所在。

Again, dumb this down.

我再把这事说简单点。

It's like playing chess and checkers, right?

这就好比一边在下国际象棋,一边在下跳棋,对吧?

The fact of the matter is the folks who are starting AS6 today whether it's Tranium or whether it's some of these other accelerators etc. they're building a chip that's a component of a much larger machine. you've built a very sophisticated system, platform, factory, whatever you want to call it, and now you're opening up a little bit, right?

事实就是,今天开始做 ASIC 的这些人,不管是 Trainium 还是别的一些加速器,他们做的是一颗芯片,是一台大得多的机器里的一个部件。你已经造出了一套非常精密的系统、平台、工厂,随你怎么叫,而现在你开始稍微开放一点了,对吧?

So, you mentioned CPXGPU, right?

所以你提到了 CPX GPU,对吧?

That is, it seems to me that it in some ways you're disagregating the workloads to the best slice of the hardware for that particular domain.

在我看来,你某种程度上是在把负载解耦,分给最适合那个特定领域的那一块硬件。

Well, we did we announced this thing called Dynamo,

我们发布了一个叫 Dynamo 的东西,

right? disagregated orch AI workload orchestration and we open sourced it because the future AI factory is disagregated

对吧?解耦式的 AI 负载编排,而且我们把它开源了,因为未来的 AI 工厂是解耦的。

right

没错。

and you launched MV Fusion

而你们还推出了 NVLink Fusion,

that even said to your competitors

那等于是在跟你的竞争对手说 ——

including Intel which you just invested in

包括你刚投资的 Intel ——

that's right

没错。

you know the way in which you participate in this factory that we're building because nobody else is crazy enough to try to build the entire factory but you can plug into that if you have a product that's good enough, compelling enough that the end user says, "Hey, we want to use this instead of an ARM GPU or we want to use this instead of your inference accelerator, etc.

你参与我们正在建的这座工厂,方式是这样:没有别人疯到要去建整座工厂,但只要你有一个足够好、足够有说服力的产品,让终端用户说「嘿,我们想用这个,而不是用 ARM 的 GPU」,或者「我们想用这个,而不是用你的推理加速器」,你就可以插进来。

Is that correct?"

这样理解对吗?

We're delighted delighted to connect you in.

我们非常非常乐意把你接进来。

Yeah.

对。

Tell us a little bit fusion such a great idea and we're so happy to partner with Intel on that.

跟我们讲讲 Fusion —— Fusion 这主意太好了,我们非常高兴能跟 Intel 在这上面合作。

It takes takes the Intel ecosystem, you know, there most of the world's enterprise still runs on Intel.

它把 Intel 的生态 —— 世界上大部分企业级业务今天还是跑在 Intel 上 ——

It takes the Intel ecosystem, takes the Nvidia AI ecosystem, accelerated computing, and we fused it together,

它把 Intel 的生态,和 NVIDIA 的 AI 生态、加速计算,融合到了一起,

right?

对吧?

And we did that with ARM,

ARM 那边我们也做了同样的事,

right?

对吧?

And there are several others we're going to be doing it with.

而且还有好几家我们也会这么做。

And and uh that that opens up opportunities for both of us.

这为双方都打开了机会。

It's a win for both of us.

这对双方都是赢。

Great great win.

很大的赢。

I'll be a large customer of theirs and uh they're going to expose us to a a much much larger market opportunity.

我会成为他们的大客户,而他们会把我们带到一个大得多的市场机会面前。

Yeah. that's deeply related to this idea is the argument you've made that kind of um c shock some people where you say our competitors building A6 they could literally all their chips are cheaper already today but they could literally price them at zero our objective is they could price them at zero and you would still buy an Nvidia system because the total cost of operating that system power data center land etc the intelligence out is still a better bet than buying a chip even if it's given to you for free

对。跟这件事深度相关的,是你提过的一个论点 —— 那个论点让一些人有点震惊 —— 你说,我们那些做 ASIC 的竞争对手,他们的芯片今天本来就更便宜,但他们甚至可以把价格定成零;我们的目标就是,哪怕他们定价为零,你还是会买 NVIDIA 的系统,因为把运营那套系统的总成本 —— 电力、数据中心、土地等等 —— 拿去换产出的智能,仍然是更划算的一笔账,哪怕那颗芯片是白送给你的。

because the land power and shell is already $15 billion,

因为土地、电力和厂房外壳本身就已经是 $15 billion 了,

right?

对吧?

Yeah.

对。

So, we've taken a crack at kind of the math on that.

这笔账我们自己也试着算过。

But walk us through your math because I think for people who don't spend as much time here that it just doesn't compute.

但请你带我们走一遍你的算法,因为很多人没像我们这样在这上面花这么多时间,这笔账在他们看来根本算不通。

How could it possibly be that you were pricing your competitor's chips at zero given the expense of your chips and it still is a better bet?

你自己的芯片那么贵,怎么可能把竞争对手的芯片按零定价,买你的还是更划算?

There's two ways to think about it.

可以从两个角度想这件事。

Um, one way is um, uh, let's just think about it from a perspective of revenues.

一种是,我们就从营收的角度看。

Yes.

是。

Okay.

好。

So everybody's power limited and let's say uh you were able to secure two more gigawatts of power.

每个人都受电力限制,假设你又多拿到 2 gigawatts 的电。

Well, that two gawatts of power you would like to have translate to revenues.

那么这 2 gigawatts 的电,你希望它能变成营收。

Yes.

是。

So your performance or tokens per watt was twice as high as somebody else's token per watt because you did I did deep and extreme code design,

那么假设你的性能,或者说每瓦 token,是别人每瓦 token 的两倍,因为我做了深度的、极致的协同设计,

right? and my performance was much higher per unit energy,

对吧?我在单位能耗上的性能高得多,

then my customer can produce twice as much revenues

那我的客户就能产出两倍的营收,

from their data center.

从他们的数据中心里出来。

And who doesn't want twice as much revenues? and and and if somebody gave them a 15% discount,

谁不想要两倍的营收?而如果有人给他们打个 15% 的折扣 ——

you know, the difference between our gross margins, which is called the 75 points, and somebody else's gross margins, call it the 50 to 65 points, is not so much as to make up for the 30 times difference between Black Wall and Hopper.

我们的毛利率,就说 75 个点,和别人的毛利率,就说 50 到 65 个点,这中间的差距还不足以弥补 Blackwell 和 Hopper 之间 30 倍的差距。

Let's pretend Hopper Hopper is an amazing chip, an amazing system.

我们就假设 Hopper 是一颗了不起的芯片、一套了不起的系统。

Let's pretend somebody else's ASIC is Hopper.

我们就假设别人的 ASIC 相当于 Hopper。

Yeah.

对。

Black Wall's 30 times.

Blackwell 是它的 30 倍。

So you've got to give up 30x revenues in that one gigawatt.

所以在那 1 gigawatt 里,你得放弃 30 倍的营收。

Mhm.

嗯。

It's too much to give up.

要放弃的太多了。

So even if they gave it to you for free, you you you only have 2 gigawatts to work with.

所以哪怕他们白送给你,你手上也只有 2 gigawatts 可用。

Your opportunity cost is so insanely high.

你的机会成本高得离谱。

You would always choose the best perf per watt.

你永远都会选每瓦性能最好的那个。

So I heard this from one of the CFOs at one of the hyperscalers that given the performance improvement right that's coming out of your chips again precisely to that point tokens per per gig um and power being the limiting factor right that they had to upgrade uh to the new cycle so when you look ahead at Ruben at Ruben Ultra at Fineman

我从一家超大规模厂商的 CFO 那里听到过这个说法:考虑到你们芯片带来的性能提升 —— 恰恰就是这一点,每 gigawatt 产出的 token,而电力是限制因素 —— 他们不得不升级到新一代;那么你往前看 Rubin、看 Rubin Ultra、看 Feynman,

does that trajectory continue

这条轨迹会继续下去吗?

we're building what six seven chips a year now

我们现在一年要做六七颗芯片,

yeah and and each one

对,而每一颗

that's part of that system.

都是那套系统的一部分。

That's right.

没错。

And those that system software is everywhere and it takes

而那套系统软件无处不在,而且需要

it takes the integration and the optimization across all of those six seven chips to deliver on the 30x blackwell.

需要在这六七颗芯片之间做整合和优化,才能兑现 Blackwell 那 30 倍。

Now imagine I'm doing this every single year.

现在想象一下,我每一年都要来这么一遍。

Bam bam bam bam bam bam.

砰、砰、砰、砰、砰、砰。

And so if you build one ASIC in that soup of AS6 in that soup of chips and we're optimizing across that, you know, it's a hard problem to solve.

所以如果你在这一大锅 ASIC、这一大锅芯片里只做出一颗 ASIC,而我们是跨整锅在做优化 —— 这问题很难解。

This does bring me back to where we started about the competitive moat.

这确实又回到我们开头聊的护城河问题。

We've been covering this and investors for a while.

作为投资人,我们关注这件事已经有一阵子了。

We're investors throughout the ecosystem and in competitors of yours, you know, from Google to to Broadcom.

我们在整个生态里都有投资,也投了你的竞争对手,从 Google 到 Broadcom。

But when I really just first principles around this and say

但我真的用第一性原理去想这件事,我会这么问:

are you increasing or decreasing your competitive mode, you move to an annual cadence.

你的护城河是在变宽还是变窄?你转到了一年一代的节奏。

You're co-developing with a with a supply chain.

你在和整条供应链协同开发。

The scale is massively bigger than anybody anticipated which requires scale both of balance sheet and of development.

规模比所有人预期的都大得多,这既要有资产负债表的体量,也要有研发的体量。

Right? the moves you made both through acquisition and organically with things like Envy Fusion, CPX, which we just talked about.

对吧?你通过收购、也靠自研做的那些动作,比如 NVLink Fusion、我们刚聊过的 CPX。

All of those things together cause me to believe that your competitive mode is increasing visav at least in so far as building out the factory or the system.

这些加在一起,让我相信你的护城河相对同行是在变宽,至少在把整座工厂、整个系统建出来这件事上是这样。

It's at least surprising.

起码可以说,这挺出人意料的。

But but but I think it's interesting that your multiple is much lower than most of those other people.

但我觉得有意思的是,你的估值倍数比那几家里的大多数都低得多。

And I think part of that has to do with this law of large numbers.

我认为这里面一部分要归到大数定律上。

A $4.5 trillion company couldn't possibly get any bigger.

一家 $4.5 trillion 的公司,不可能再大下去了。

But I asked you this a year and a half ago as you sit here today.

但一年半前我就问过你这个问题,今天你坐在这里 ——

If the market's going to AI workloads are going to 10x or 5x, you know, we know what capex is doing, etc.

如果市场要……AI 工作负载要涨 10 倍或 5 倍,capex 往哪走我们也清楚,等等。

Is there any conceivable world in your mind where your top line in 5 years isn't two or 3x bigger than it is in 2025?

在你心里,有没有任何想象得到的世界,是你五年后的营收没比 2025 年大出两三倍?

Like what's the probability that it's actually not not much higher than it is today given those advantages?

考虑到这些优势,最后它并没有比今天高出多少的概率有多大?

I'll answer it this way.

我这么回答吧。

Our opportunity as I described it is much larger than the consensus.

按我前面描述的,我们的机会比卖方共识预期大得多。

I'll say it here.

我就在这儿说了。

I think Nvidia will likely be the first 10 trillion dollar company.

我认为 NVIDIA 很可能会成为第一家 $10 trillion 的公司。

And I would I've been here long enough, it wasn't that long ago, just a decade ago, as you well remember, that people said there could never be a trillion dollar company.

我在这行待得够久了 —— 就在不久之前,也就十年前,你也记得很清楚,大家还说不可能出现 $1 trillion 的公司。

Now we have 10, right?

现在有 10 家了,对吧?

And today the world's bigger,

而今天,世界更大了,

right?

对吧?

And today, this is this is the back to the exponentials around GDP and the growth.

而今天,这又回到 GDP 和增长那些指数曲线上了。

The world is bigger and and and people misunderstand what we do.

世界更大了,而且大家误解了我们做的事。

They they uh they remember we're a chip company, right?

他们记得我们是一家芯片公司,对吧?

And we are we build chips.

我们确实是,我们造芯片。

Boy, do we build chips and build the most amazing chips in the world.

天哪,我们太会造芯片了,造的是全世界最了不起的芯片。

But NVIDIA is really an AI infrastructure company.

但 NVIDIA 其实是一家 AI 基础设施公司。

You know, we are your AI infrastructure partner and our partnership with OpenAI is a perfect demonstration of that.

我们是你的 AI 基础设施伙伴,我们和 OpenAI 的合作就是最好的例证。

Yeah.

对。

That we are their AI infrastructure partner and we work with people in a lot of different ways.

我们就是他们的 AI 基础设施伙伴,而我们跟人合作的方式有很多种。

You know, you you you would uh we we don't require anybody to buy everything from us.

我们不要求任何人把所有东西都从我们这儿买。

Um we don't we don't require that they buy uh the full rack.

我们不要求他们买整个机架。

They could buy a chip.

他们可以只买一颗芯片。

They could buy a component.

可以只买一个部件。

They could buy our networking.

可以只买我们的网络。

They could buy our We have customers buying only our CPU.

可以只买我们的……我们有客户只买我们的 CPU。

You know, just buy our GPUs and buy somebody else's CPUs and somebody else's networking.

只买我们的 GPU,CPU 买别人的,网络也买别人的。

You know, we're kind of okay selling any way you like to buy.

你想怎么买,我们基本上都卖。

You know, my only request is just buy a little something from us.

我唯一的请求是:从我们这儿多少买一点。

You know, you said, you know, this isn't just about better models.

你说过,这不只是模型更好的问题。

We also have to build.

我们还得真的把东西建出来。

We have to we have to have worldclass builders.

我们必须有世界级的建造者。

And you said, you know, the most world-class builder maybe that we have in the country is Elon Musk.

你还说,这个国家里最世界级的建造者,可能就是 Elon Musk。

And we talked about Colossus one and what he, you know, what he was doing there, standing up a couple hundred thousand, you know, at the time H100s, H200s in a coherent cluster.

我们聊过 Colossus 1,聊过他在那里干的事 —— 在一个一致性集群里立起当时二十来万张 H100、H200。

Now he's working on Colossus 2, you know, which may be 500,000 GBs, um, millions of H100 equivalents in a coherent cluster.

现在他在做 Colossus 2,可能是 500,000 张 GB,相当于数百万张 H100,全在一个一致性集群里。

I would not be surprised if he gets to a gigawatt before anybody else does in one.

如果他比任何人都先在单个集群里做到 1 gigawatt,我不会意外。

Yeah.

对。

So say a little bit about that.

那你稍微讲讲这个。

The advantage of being, you know, the builder who, you know, isn't just building the software and the models, but understands what it takes to to build those clusters.

作为建设者的优势在哪——你不只是做软件和模型,还懂得建起这些集群要付出什么。

Well, you know, these AI supercomputers are complicated things.

这些 AI 超级计算机很复杂。

The technology is complicated.

技术复杂。

Procuring it is complicated because of financing issues.

采购也复杂,因为有融资问题。

Securing the land power and shell, powering it is complicated. building it all, bringing it all up.

搞定土地、电力和厂房外壳,再给它供上电,复杂;把这一切建起来、全部跑起来,也复杂。

I mean, these are this is unfortunately the most complex systems problem humanity has ever endeavored.

很不幸,这是人类迄今挑战过的最复杂的系统难题。

And and so Elon has has a great advantage that in his head um all of these systems are interoperating and um and the and the interdependencies um are are um you know resides in one head including the financing.

所以 Elon 有个巨大优势:在他脑子里,所有这些系统是打通在一起运转的;那些相互依赖关系——包括融资——全装在同一个脑子里。

Yes.

对。

And so

所以说——

he's a big GPT.

他自己就是个大 GPT。

He's a big supercomputer himself.

他本人就是一台大型超级计算机。

He's the Yeah. the ultimate GPU.

他就是——对,终极 GPU。

Yeah.

对。

Yeah.

对。

And so so he has a great advantage there.

所以他在这上面有巨大优势。

Yeah.

对。

And and he has a great sense of urgency.

而且他有很强的紧迫感。

He Yeah.

他——对。

He has a has a real desire to to build it and and so when when will comes together with

他真有一股想把它建出来的渴望,所以当意志遇上——

with skill.

遇上本领。

Yeah.

对。

You know, unbelievable things can happen.

就会发生不可思议的事。

Yes.

对。

Yeah.

对。

Quite unique.

非常独特。

Something you've been so involved in is I want to talk about sovereign AI.

有件你参与很深的事——我想聊聊主权 AI。

I want to talk about China and the global AI race that's going on.

我想聊中国,聊眼下这场全球 AI 竞赛。

You know, when I look back at you 30 years ago, you couldn't have imagined you were going to be hanging out in palaces with airs and the king this week and you're at the White House all the time.

回头看 30 年前的你,你根本想象不到,这一周你会在宫殿里跟埃米尔和国王坐在一起,还成天出入白宫。

The president has said that you and Nvidia are critical to uh US, you know, national security.

总统说过,你和 NVIDIA 对美国国家安全至关重要。

So when when you look at that first just contextualize for me like it's hard to believe that you would be in those places if sovereigns didn't view this at least as existential as important as maybe we did nuclear in the 1940s right we don't have a Manhattan project today at least funded by the government but it's funded by Nvidia it's funded by open AI it's funded by Meta it's funded by Google we have companies today the size of nation states and thank god God for America, right?

所以你看这件事——先帮我把背景铺一下:很难相信,要不是各国主权政府至少把它看得像我们 1940 年代看核武器那样攸关存亡、那样重要,你会出现在那些场合,对吧?今天我们没有曼哈顿计划——至少不是政府出钱的——出钱的是 NVIDIA、是 OpenAI、是 Meta、是 Google;今天有些公司的体量顶得上一个民族国家,感谢上帝有美国,对吧?

Who are funding something that it appears to me presidents and kings think are think is existential to their future economic and national security.

而它们在投的这件事,在我看来,总统和国王们都视为攸关自身未来经济安全与国家安全的存亡大事。

Would you agree with that?

你同意吗?

Nobody needs atomic bombs.

没人需要原子弹。

Everybody needs AI.

人人都需要 AI。

Well said.

说得好。

Okay.

好。

Here.

说得对!

Here.

说得对!

Yeah.

对。

Here.

说得对!

And and so that's a very very large difference.

所以这里的差别非常非常大。

Um AI AI as you know is modern software.

AI 就是现代软件。

I just that's where I started from general purpose computing to accelerated computing from human written code line at a time to AI written code that foundation can't be forgotten we've reinvented computing there's not a new species on earth we just reinvented computing and everybody needs computing it needs to be democratized

我就是从这儿讲起的:从通用计算到加速计算,从人一行一行写的代码到 AI 写出来的代码——这个基础不能忘。我们重新发明了计算,地球上并没有多出一个新物种,我们只是重新发明了计算;而所有人都需要计算,计算必须民主化。

which is the reason why everybody all of these all of the countries realize they have to get into the AI world because everybody needs to stay in computing.

这就是为什么所有这些国家都意识到,自己必须进入 AI 的世界——因为所有人都得留在计算里。

There's nobody in the world that says, "Guess what?

世界上没有人会说:「你猜怎么着?

You know, I used to use computers yesterday.

我昨天还在用计算机。

I'm pretty good with, you know, clubs and fire tomorrow, you know, and so everybody needs to move into computing.

明天我拿根棒子、点堆火就挺好。「所以所有人都得转向计算。

It's just it's just being modernized.

只是在现代化而已。

That's all.

仅此而已。

Okay.

好。

Number one, um it it is the case that that in order to participate in AI, you have to encode within AI, your your history, your culture, your values.

第一点,确实是这样:要参与 AI,你就得把自己的历史、文化、价值观编码进 AI 里面。

And and of course, AI is getting smarter and smarter so that even the core AI is able to learn these things fairly quickly.

当然,AI 越来越聪明,所以连核心的 AI 都能很快学会这些。

You don't have to start from the ground, you know, from ground zero.

你不必从零开始。

And so I I think that that every country um needs to have some sovereign capability.

所以我认为,每个国家都需要有一定的主权能力。

I recommend that they all use OpenAI, they all use Gemini, they all use, you know, these open models use Grock and I think they I recommend they all do that.

我建议他们都用 OpenAI,都用 Gemini,都用这些开放模型、用 Grok,我建议他们都这么做。

I I recommend they all use anthropic.

我建议他们都用 Anthropic。

Um but they should also dedicate resources to learn how to build AI.

但他们也应该投入资源,去学会怎么造 AI。

And the reason for that is because they need to learn how to build it not just for language models, but they need to build it for industrial models, manufacturing models, national security,

原因在于,他们要学会造的不只是语言模型,还得造工业模型、制造业模型,还有国安——

national security models.

国安模型。

There's a whole bunch of intelligence they had to go cultivate themselves.

有一大堆智能,得靠他们自己去养出来。

So they they ought to have sovereign capability.

所以他们理应有主权能力。

Every country should develop it.

每个国家都该把它发展起来。

And is that what you see?

你看到的就是这样吗?

Is that what you're hearing around the world?

你在世界各地听到的就是这样?

They all realize it.

他们全都意识到了。

They all realize

他们全都意识到——

they all realize it.

他们全都意识到了。

And they they all are going to be customers of OpenAI and Throbic and Grock and Gemini, but they all really need to also build their own infrastructure.

而且他们都会成为 OpenAI、Anthropic、Grok 和 Gemini 的客户,但他们确实也都需要建自己的基础设施。

And this is this is the big idea that that what Nvidia does is we're building infrastructure.

这才是最大的那件事:NVIDIA 做的就是建基础设施。

Just as every country needs energy infrastructure, the communications and internet infrastructure, now every single country needs AI infrastructure.

就像每个国家都需要能源基础设施、通信和互联网基础设施,现在每个国家也都需要 AI 基础设施。

So you let's start with the rest of the world.

那我们先从世界其他地方说起。

You know, our our good friend David Saxs.

我们的好朋友 David Sacks。

Um the AIS are doing a heck of a job.

AI 这块的几位负责人干得非常出色。

We are in so lucky.

我们太幸运了。

Yeah.

是啊。

To have David and Shriram in Washington DC.

能有 David Sacks 和 Sriram Krishnan 在华盛顿特区。

Um doing you know and David doing Yeah. doing AI in the AISR.

在做…… David 负责的是 AI,做 AI 事务主管。

Uh this at what a what a smart move by President Trump to put them in the White House.

特朗普总统把他们放进白宫,这一步实在太聪明了。

Um because during this pivotal time

因为在这个关键时刻,

Yes.

是的。

the technology is complicated.

技术很复杂。

Shriram is the only person in Washington DC that I think knows CUDA.

我觉得 Sriram 是华盛顿特区唯一懂 CUDA 的人。

Yeah.

是啊。

Um and and which is strange anyways.

而这件事本身就挺奇怪的。

But but I I just love the fact that during this pivotal time when technology is complicated, policy is complicated, the impact to the future of our nation is so great

但我特别喜欢这一点:在这个关键时刻,技术复杂,政策复杂,对我们国家未来的影响又这么大,

that we have somebody who is clear-minded, dedicating the time to understand the technology

我们有一个头脑清醒的人,肯花时间把技术弄懂,

and thoughtful to uh help us through that.

而且想得周全,帮我们把这一段走过去。

And it would seem to me, yeah, I'm going back to the Manhattan Project analogy.

在我看来,对,我又要回到曼哈顿计划那个类比了。

Yeah.

是啊。

Right. that you have a president who understands

没错。你有一位总统,他明白

how existential this is.

这件事关乎生死存亡。

You have governors like Greg Abbott in Texas who want to remove regulations to accelerate because they understand how important it is.

你有像德州的 Greg Abbott 这样的州长,他们想砍掉监管来加速,因为他们懂这有多重要。

You have secretaries right at energy and Doug Bergram at interior and Lutnik at commerce who also understand how important this is, how pro- energy they are.

你有能源部的部长、内政部的 Doug Burgum、商务部的 Lutnick,他们也都明白这有多重要,他们非常支持能源。

Could you imagine?

你能想象吗?

Could you imagine the alternative if we had an administration right now who is not proen energy and want energy to grow in our nation so that we could have AI?

你能想象另一种可能吗 —— 如果现在这届政府不支持能源,不希望国内能源增长、好让我们能搞 AI?

I find it

我觉得这

I just can't even think about it.

我简直都不敢去想。

I find it ironic that that that just a couple years ago we were saying China's building a 100 nuclear reactors.

我觉得挺讽刺的,就在两三年前,我们还在说中国在建 100 座核反应堆。

They're so far ahead of us.

他们领先我们太多了。

Like that's the the the primitive to AI.

那是 AI 的前置条件。

But now you have people when we go to build it, everybody says, "Oh, it's a glut, right?

可现在,等我们真要建了,所有人都说:「哦,这是产能过剩」,对吧?

Like it seems to me that this is something that the government, it is in their interest.

在我看来,这件事符合政府自身的利益。

And we have industry and government working together in a way that I haven't seen in a long time.

而且产业和政府现在的合作方式,是我很久没见过的。

You've been around a long time.

你在这行很久了。

You you're very close with President Trump at this stage.

到今天这个阶段,你跟特朗普总统走得很近。

Help us understand like what is the nature of industry government relationships?

帮我们理解一下,产业和政府之间的关系到底是什么性质?

We saw that dinner last week with all the CEOs.

上周那场所有 CEO 都在的晚宴,我们都看到了。

You know, you spent a lot of time.

你在这上面花了很多时间。

Is it unique?

这是独一无二的吗?

Have you seen anything like this in your career over the last 30 years?

过去 30 年的职业生涯里,你见过类似的情形吗?

It was it was hard to go to DC in the past as you know.

你也知道,过去要去华盛顿特区很难。

Uh getting an appointment is almost impossible,

约上一次见面几乎不可能,

right?

对吧?

Uh President Trump has a open door to leaders who wants to come in and uh help them understand the future.

特朗普总统对那些愿意进来、帮他们理解未来的领导者是敞开大门的。

Um this is an administration that believes in growth.

这是一届相信增长的政府。

Fundamentally, President Trump wants America to grow.

从根本上说,特朗普总统希望美国增长。

Yeah.

是啊。

If we can grow economically, we will be strong militarily.

如果我们经济上能增长,军事上就会强大。

If we could be if we could grow economically, we will be secure.

如果我们经济上能增长,我们就会安全。

I've never met somebody who is secure who's poor.

我从没见过哪个穷人是安全的。

Being being rich as a nation is an essential part of national security.

一个国家富有,是国家安全不可或缺的一部分。

And he knows that.

而他懂这一点。

He also wants America to win the AI the the AI race.

他也希望美国赢下这场 AI 竞赛。

This is going to be a very long-term race.

这会是一场非常长期的竞赛。

And um and he understands that this is a pivotal time.

而他明白,现在是关键时刻。

He wants the technology industry to run.

他希望科技产业跑起来。

He wants everybody in the world to be built on American technology.

他希望全世界都建在美国技术之上。

These are sensible, logical things.

这些都是合情合理、符合逻辑的事。

You know, the opposite is strange to me.

反过来的那一套,我觉得很奇怪。

If I take everything and I just reversed it, we want our country not to grow

如果我把所有这些都反过来:我们不希望自己的国家增长,

and because we don't want our country to grow, we don't need any energy because we know we need energy to grow and so let's not have any energy and in fact we don't want our technology industry to lead. uh he understands that our technology industry is our national treasure.

又因为不希望国家增长,我们就不需要能源 —— 我们知道增长需要能源,那就干脆别要能源;而且事实上,我们也不希望自己的科技产业领先。他明白,我们的科技产业是国家的珍宝。

Correct.

对。

And that technology like corn and steel and things in the past are now such fundamental trade opportunities.

而技术,就像过去的玉米和钢铁一样,现在成了这么根本的贸易机会。

It's an essential part of trade.

它是贸易里不可或缺的一环。

And why would you not want American technology to be coveted by everyone so that it could be used for trade?

你为什么会不希望人人都想要美国技术、好让它能用来做贸易呢?

Right.

没错。

So let's talk about you know

那我们来聊聊

the internet.

互联网。

Yeah.

是啊。

Google spread around the world.

Google 传遍了全世界。

Yeah, we had democratic values spread around the world by way of search and Google didn't have to go to Washington to get permission to do it.

是啊,我们靠搜索把民主价值观传遍了全世界,而 Google 不需要跑去华盛顿拿许可才能这么做。

It just happened.

它就那么自然发生了。

We diffused our technology around the world.

我们把自己的技术扩散到了全世界。

David Sachs has been crystal clear of the need to accelerate export licenses so that the American AI stack wins around the world.

David Sacks 讲得非常清楚:必须加快出口许可,让美国的 AI 技术栈在全世界赢。

Right?

对吧?

We're talking chips, we're talking models, we're talking data centers, etc.

我们说的是芯片,是模型,是数据中心,等等。

We know a year and a half ago that wasn't happening. was a concept that was called small yard tall fence or something like that.

我们知道,一年半前并不是这样。当时有个概念,叫「小院高墙」之类的。

A small yard tall fence and and the irony of it was it was described in such a way and it was recommended in policy in such a way it was a small yard tall fence around America.

小院高墙 —— 讽刺的是,照它当时那种描述方式、那种写进政策建议的方式,这道小院高墙最后是围着美国建的。

That was the strange part.

这才是奇怪的地方。

I think President Trump's got it right that we want to maximize exports.

我认为特朗普总统这一点看对了:我们要把出口最大化。

We want to maximize American influence around the world.

我们要把美国在全世界的影响力最大化。

We're supposed to maximize those things.

我们本来就该把这些都做到最大。

And do you see those licenses coming?

那你看到那些出口许可下来了吗?

Are you seeing the acceleration in Washington?

你在华盛顿感觉到那种加速了吗?

I know it's being said at the top, but are you seeing it flow down through government that's accelerating us around the world?

我知道高层是这么说的,但你有没有看到它往下贯穿到整个政府,在全世界替我们加速?

Secretary Lutnick was all over it.

Lutnick 部长一直盯得很紧。

Great.

太好了。

Yeah.

是啊。

So, now let's talk about China.

那么,现在我们来聊聊中国。

You know what most people may not realize is I think you understand China as well as any leader in the United States.

大多数人可能没意识到的是,我觉得你对中国的了解,不输美国任何一位领导者。

We've been there for 30 years.

我们在中国已经 30 年了。

Been there for 30 years.

在中国 30 年了。

What most people don't realize is is up until a couple years ago, you had dominant market share within China in terms of

大多数人没意识到,直到几年前,你们在中国的市场份额一直是压倒性的——

95% market share.

95% 的市场份额。

95% market share in the most important thing arguably.

95% 的市场份额,而且可以说是在最重要的那件事上。

And you have said that our biggest goal that we as a country could have under the guise of somehow trying to slow them down is we've unilaterally disarmed.

你说过,我们作为一个国家,打着「设法拖慢他们」的旗号,能实现的最大目标就是单方面解除了自己的武装。

We forced Nvidia out of China, which has allowed Huawei to accelerate on the back of monopoly profits within China.

我们把 NVIDIA 逼出了中国,让华为靠着中国国内的垄断利润加速起来。

And I just saw this morning, you're seeing announcements out of Huawei and Baba and others that they're going to build data centers around the world.

我今天早上刚看到,华为、阿里巴巴这些公司都在宣布要在全球各地建数据中心。

Now Huawei has a three-year plan to pass Nvidia funded by the monopoly profits in the biggest AI market in the world.

现在华为有一个三年计划,要靠全球最大 AI 市场里的垄断利润超过 NVIDIA。

So it's looking like your admonition that this is a huge mistake to hand China, you know, monopoly markets is coming true.

所以看起来,你那句「把垄断市场拱手交给中国是个巨大错误」的警告,正在成真。

The president said, you know, a a after kind of the ban on H20s, now we have a situation where you can sell, you know, chips to China, but there's a 15% export tax.

总统说过,在 H20 被禁之后,现在的局面是:你可以向中国卖芯片,但要交 15% 的出口税。

But now it appears that the Chinese perhaps offended by statements out of the United States are saying no, Nvidia is not allowed to sell here.

但现在看起来,中国那边可能被美国的一些表态惹到了,说不行,NVIDIA 不许在这儿卖。

Now where do we stand today between Nvidia and China?

那么今天 NVIDIA 和中国之间是什么状况?

And can you reiterate kind of what you think we as a country should be doing to put ourselves in a best position to win the AI race around the world?

你能不能再讲一遍,你认为我们作为一个国家该怎么做,才能让自己处在赢下全球 AI 竞赛的最佳位置?

We have a competitive relationship with China.

我们和中国是竞争关系。

We should acknowledge that China rightfully should want their companies to do well.

我们该承认,中国理所当然希望自己的公司做得好。

We I don't I don't for a second begrudge them for them.

我一秒钟都不会怪他们。

They should do well.

他们就该做得好。

They should they should give them as much support as they like.

他们想给多少支持,就给多少支持。

It's all their prerogative.

这完全是他们的权利。

And don't forget that China has some of the best entrepreneurs in the world because they came from some of the best STEM schools in the world.

而且别忘了,中国有全世界最好的一批创业者,因为他们出自全世界最好的 STEM 院校。

They're they're the most hungry in the world.

他们是全世界最有饥饿感的。

Yes.

是的。

996 as you know.

就是 996,你知道的。

This is a very

这是一个非常

producing the most AI engineers in the world.

培养出全世界最多的 AI 工程师。

96.

996。

So the audience knows 9 in the morning to 9 at night 6 days a week.

跟听众说一下,就是早上 9 点到晚上 9 点,一周 6 天。

That is their culture.

那就是他们的文化。

Yeah.

对。

Okay.

好。

We're up against a formidable, innovative, hungry, fastm moving, underregulated.

我们面对的是一个强悍、有创新力、有饥饿感、动作快、监管又松的——

Yeah.

对。

Okay.

好。

People don't realize this.

人们没意识到这一点。

They are very lightly regulated,

他们受的监管非常宽松,

right?

对吧?

Less regulated, ironically, than we are in a capitalist system.

讽刺的是,他们受的监管比我们这个资本主义体系还要少。

That's right.

没错。

People think that they're centrally governed.

人们以为他们是中央统一治理的。

But remember, the genius of China was distributed economic systems.

但要记住,中国的高明之处在于分布式的经济体系。

Yeah.

对。

And so all of these 33

所以这 33 个

provinces and all the mayor economy has driven enormous amount of internal competition, internal economic vibrancy, which of course has some of its side effects.

省,加上整套市长经济,催生了巨大的内部竞争和内部经济活力,当然这也有它的副作用。

But this is a vibrant, entrepreneurial, high techch, modern industry.

但这是一个有活力、有创业精神、高科技的现代产业。

And two, one, uh, some of the things I I heard, uh, they could never build AI chips.

另外,第一,我听到过一些说法,说他们永远造不出 AI 芯片。

That just sounded insane.

这听起来简直荒唐。

Two, uh, that China can't manufacture.

第二,说中国搞不了制造。

China can't manufacture.

中国搞不了制造。

If there's one thing they could do is manufacture.

要说有一件事他们最拿手,那就是制造。

And three, they're years behind us.

第三,说他们落后我们好几年。

Is it two years, three years?

两年?三年?

Come on.

拜托。

They're nanconds behind us.

他们只落后我们几纳秒。

Nanoc.

纳秒。

Yeah, they're nanconds behind us.

对,他们只落后我们几纳秒。

And so we've got to go compete.

所以我们必须去竞争。

Yeah,

对,

we've got to go compete.

我们必须去竞争。

And so so the question then becomes um what's in the best interest uh what's in the best interest of China of course um is that they have a vibrant industry.

那么问题就变成:什么最符合中国的利益——当然是他们有一个充满活力的产业。

Uh they also publicly say and rightfully I believe they believe this is that they want China to be an open market.

他们也公开这么说,而且我相信他们真这么想:他们希望中国是个开放的市场。

They want to attract uh foreign investment.

他们希望吸引外国投资。

They want companies to come to China and compete in the marketplace,

他们希望企业进中国、在市场上竞争,

right?

对吧?

And I believe that they I hope I believe and I hope that would return to that

我相信他们——我希望,我相信也希望,他们会回到那个状态,

in our context.

放到我们的语境里。

Answering your question, what do I see in the future?

回答你的问题:我看到的未来是什么?

I do hope because they they say it um their leaders say it and I take it at face value and I believe it because I think it makes sense for China that what's in the best interest of China is for foreign companies to invest in China, compete in China

我确实抱这个希望,因为他们这么说,他们的领导人这么说,我照字面接受,我也相信,因为我觉得这对中国是合理的——最符合中国利益的,就是外国公司来中国投资、在中国竞争,

and for them to also have vibrant competition themselves and they would also like to come out of China and participate around the world.

同时他们自己内部也有充满活力的竞争,而他们也想走出中国、参与到全世界。

That is I think is a fairly sensible outcome and we what we need to do as a country is to enable our technology industry which today is the I'm privileged to be working in an industry that is our national treasure.

我认为这是相当合理的结果,而我们作为一个国家要做的,是让我们的科技产业能去竞争——今天它就是,我很幸运能身处一个堪称国宝的产业。

We have to acknowledge it is our national treasure.

我们必须承认,它是我们的国宝。

It is our best industry.

它是我们最好的产业。

It is our single best industry.

它是我们最好的那一个产业。

Yeah.

对。

Why would we not allow this industry to go compete for its survival for for this industry to go and proliferate the technology around the world so that we could have the world be built on top of American technology so that we can maximize our economic success magn uh maximize our geopolitical influence u maximize this this technology industry during such a vibrant time such a pivotal time to allow it to thrive.

我们为什么不让这个产业去为自己的生存而竞争,不让它把技术推向全世界,让世界建立在美国技术之上,从而把我们的经济成果最大化、把我们的地缘政治影响力最大化,在这样一个充满活力、如此关键的时刻,把这个科技产业做到最大,让它蓬勃发展?

The skeptic says Jensen just wants to sell more chips and if he can sell them to China, great.

怀疑者会说,黄仁勋就是想多卖芯片,能卖给中国就太好了。

He'll sell them to China.

他就会卖给中国。

He doesn't care about, you know, what that means for America.

他不在乎这对美国意味着什么。

That's a skeptic.

这是怀疑者的说法。

Now,

那么,

can I just can I just address the skeptics?

我能不能回应一下这些怀疑者?

Just because I want America in ecosystem and economy to grow doesn't make me wrong.

我希望美国的生态和经济增长,这并不代表我就是错的。

Right.

没错。

Right.

没错。

Okay.

好。

So, first of all, everything that has been said so far that's been made up so far about U.

首先,到目前为止关于美中所说的一切、编出来的一切,

China has proven to be wrong.

都已经证明是错的。

The facts are just wrong.

那些事实根本就不对。

The ground truth is wrong.

基础事实就是错的。

And so just because we want America to win, just because we want this industry to grow, doesn't make me wrong.

所以,仅仅因为我们希望美国赢,仅仅因为我们希望这个产业增长,并不代表我就是错的。

Correct.

对。

And I think anybody who knows you and now the president, certainly myself, you deeply care about the country.

我认为任何了解你的人——现在包括总统,当然也包括我——都知道你非常在乎这个国家。

You deeply want the United States of America to win the global ai race.

你非常希望美利坚合众国赢下全球 AI 竞赛。

You just happen to believe and I think you have as much experience or more experience than anyone that it enures to our advantage the probability of us winning the global AI race actually goes up if you are competing in China because it allows us to tap into half of the world's AI engineers keeping them you know in this ecos let's be clear with the companies we're talking about here bite dance Alibaba etc these are companies that are largely owned by American investors Yeah.

你只是恰好相信——我认为你的经验不比任何人少,甚至更多——这反而对我们有利:如果你能在中国竞争,我们赢下全球 AI 竞赛的概率其实会上升,因为这让我们能用上全世界一半的 AI 工程师,把他们留在这个生态里。说清楚一点,我们这里说的这些公司——字节跳动、阿里巴巴等等——大部分股权在美国投资人手里。对。

Right.

对吧?

Right.

没错。

Like these are global companies that are building recommener engines that

这些是全球性公司,做的是推荐引擎——

by the way extraordinary technologies,

顺便说一句,那是非凡的技术,

incredible companies.

了不起的公司。

And so I think and I'm hopeful

所以我认为,我也希望——

that the argument that you're making visav China, which is a harder argument than diffusion to the rest of the world.

你在中国问题上提出的这个论点,比向世界其他地方做技术扩散更难讲通。

I understand that.

这我理解。

And that's why I thought when the president said, you know, I don't know, it's a flip of a coin.

所以我当时的想法是,总统说,我不知道,这就跟抛硬币一样。

Maybe Jensen's right.

也许黄仁勋是对的。

Maybe the other guys are right.

也许另一边的人是对的。

If Jensen's willing to put a little bit of 15% into the US Treasury as a hedge on that, then I'll go for it.

如果黄仁勋愿意拿出个 15% 交给美国财政部,作为这件事的对冲,那我就同意。

Um, but I was really disappointed on the heels of that.

但紧接着发生的事让我非常失望。

Mhm.

嗯。

Um, I think if the Chinese feel like they're being taken advantage of that we're going to send them chips that are, you know, 10 years old or something, then I then I get why why they had that response.

我想,如果中国人觉得自己被占了便宜,觉得我们要给他们的是十年前的老芯片之类,那我就理解他们为什么会有那种反应了。

H20 is really quite spectacular still.

H20 到现在其实还相当出色。

And I of course it it's it's not as good as Blackwell and and I get that.

当然,它比不上 Blackwell,这我明白。

Yeah.

对。

Um I look, you know, we're I'm patient and and I believe that they're they're wise.

我有耐心,而且我相信他们是明智的。

They're thinking through their situation.

他们在把自己的处境想清楚。

Um they they have they have larger agendas to to deal with um rel you know visibly uh the United States.

他们有更大的议题要处理,明显是跟美国之间的。

There are a lot of discussions going on.

现在有很多讨论在进行。

But I'll come back to the ground truth fundamental truth.

但我还是回到那个基础事实、根本事实。

I believe that is in the best interest of China that Nvidia is able to serve that market and compete in that market.

我相信,NVIDIA 能够服务那个市场、在那个市场竞争,最符合中国的利益。

I fundamentally believe is in the best interest of China.

我从根本上相信,这最符合中国的利益。

It is of course um fantastic in the fantastic interest of the United States.

当然,这对美国也是极其有利的。

Yeah,

对,

it is fun.

确实如此。

But those two truths can coexist.

但这两个事实可以并存。

It is possible for both to be true and I believe it is both true.

两者可以同时成立,我相信两者都成立。

And so I I um uh I I'm rather,

所以我其实相当——

you know, even though I tell all of our investors

尽管我跟我们所有投资人都说

Yeah.

对。

that our guidance includes no China.

我们的业绩指引里不含中国。

Yeah.

对。

And I appreciate all of our investors to include no China in any of our guidance.

我也希望我们所有投资人在任何指引里都别把中国算进去。

We've got plenty of growth opportunities outside and we you know we've got all of that is true.

我们在中国以外有大量增长机会,这些都是真的。

It doesn't make China not important to us.

但这并不意味着中国对我们不重要。

It's very important to us.

中国对我们非常重要。

Anybody who thinks that the Chinese market is not important is has their head deep in the sand.

任何认为中国市场不重要的人,都是把头深深埋在沙子里。

Yeah.

对。

And so this is s one of the most important markets in the world.

这是全世界最重要的市场之一。

Smart markets as you know, smart people doing smart things and we want to be there.

聪明的市场,聪明的人做聪明的事,我们想在那里。

Yeah.

对。

And I think it's in the best interest of both countries that we are there.

而且我认为,我们在那里最符合两国的利益。

And so I think when I take a step back, I am confident that ultimately the wisdom will prevail.

所以退一步看,我有信心,最终智慧会占上风。

Yes.

是的。

I've I've always been confident that wisdom prevails.

我一直相信智慧会占上风。

I've always been confident that that truth prevails and uh it's taken me this far and uh I believe I believe that to be fundamentally true now.

我一直相信真理会占上风,这一路把我带到了今天,我相信现在这一点从根本上也成立。

And so these things will get sorted out and we will have the opportunity to go compete in that China market.

所以这些事情会理顺,我们也会有机会去中国市场竞争。

I'm not very political, but very topical is the administration's decision to charge 100,000 per H1B visa.

我不太谈政治,不过眼下有件事特别热门:政府决定对每份 H-1B 签证收 100,000 美元。

Mhm.

嗯。

You've spent a lot of time with the president.

你跟总统相处了不少时间。

Um you've called him our secret weapon in AI.

你还把他叫做我们在 AI 上的秘密武器。

I also know you want to recruit the best and brightest to our country.

我也知道,你想把全世界最优秀、最聪明的人才招到我们国家来。

Yeah.

对。

So, how do you think about the decision to charge 100,000 per H1B visa?

那么,每份 H-1B 签证收 100,000 美元这个决定,你怎么看?

Does this make it easier or harder to recruit talent?

这会让招人更容易,还是更难?

And, you know, does perhaps it's a little different for large companies or small companies?

还有,大公司和小公司的处境是不是有点不一样?

Like, how do you think about it?

你怎么看这件事?

I'm going to start with it's a great start.

我先说一句:这是个很好的开始。

Hold on.

等等。

You said it's a great start.

你说这是个很好的开始。

It's a great start.

这是个很好的开始。

I'm just going to start there.

我就先从这儿说起。

And the reason for that is this.

原因是这样。

That implies I don't I hope it's not the end,

这话隐含的意思是,我希望它不是终点,

but I think it's a great start.

但我认为这是个很好的开始。

I just hope it's not the end.

我只希望它不是终点。

Here's what I fundamentally believe.

我根本上相信的是这一点。

America has one a singular brand reputation that no country in the world has.

美国有一种独一无二的品牌声誉,世界上没有哪个国家有。

And no country in the world is in a position or in the horizon to be able to say come to America and realize the American dream. M

世界上也没有哪个国家现在有条件、或者在可见的将来有条件说:来美国,实现你的美国梦。

what country has the word dream behind it?

哪个国家后面还跟着「梦」这个字?

Yes, it's part of its brand.

是的,这是它品牌的一部分。

We are utterly singular and you're talking to somebody who represents the American dream.

我们完全独一无二,而且你面前坐着的这个人,本身就代表着美国梦。

My parents didn't have any money.

我父母没钱。

Sent us over here.

他们把我们送到了这里。

We started from nothing.

我们从零开始。

You guys know I, you know, bust tables, wash dishes, clean toilets, and here I am.

你们都知道,我收拾过餐桌、洗过碗、擦过厕所,而现在我在这儿。

Yeah.

对。

This is the American dream.

这就是美国梦。

President Trump knows that

特朗普总统知道,

we want legal immigrants.

我们要的是合法移民。

Yeah. there's a difference between legal immigrants and illegal immigrants.

对,合法移民和非法移民是有区别的。

But the idea that it's a country that's free for all doesn't make sense.

但把这个国家当成谁都能随便进来的地方,这说不通。

And so now the question is how do we go from the the idea that we want to protect fundamentally the American dream to dealing with illegal immigrants at such a large scale?

所以现在的问题是:我们怎么从「根本上要保护美国梦」这个想法,走到处理如此大规模的非法移民?

Um how do we find a logical pragmatic solution?

我们怎么找到一个合乎逻辑、务实的办法?

Right?

对吧?

So, the idea that he that that that we put a $100,000 price tag on H-1B um probably sets the bar a little too high,

所以,给 H-1B 贴上 $100,000 的价签,这个门槛可能定得有点太高,

but as a first bar, it at least eliminates um illegal immigration, and that's a good start.

但作为第一道门槛,它至少消除了非法移民,这就是个好的开始。

How does it how does eliminate illegal immigration?

这怎么会消除非法移民?

Well, it at least it eliminates abuse of

至少它消除了滥用——

abuse of H-1B.

对 H-1B 的滥用。

Yeah.

对。

Yeah.

对。

At least.

至少是这样。

And and that's a good start. and at least we can have a conversation.

这就是个好的开始,至少我们能坐下来谈了。

So, one of the things that we know about President Trump, he's he he's a good listener.

我们对特朗普总统了解的一点是,他很会听。

He actually listens.

他真的会听。

I mean, he listens to you, he listens to me, and

他听你说,他也听我说,

he doesn't have to.

而他本来不必听。

And he listens to a lot of people, and he's integrating a lot of information, and this is obviously a very complicated issue.

他听很多人说,他在整合大量信息,这显然是个非常复杂的问题。

And so, so I think that that this is a fine start.

所以我认为这是个不错的开始。

It's a fine start.

这是个不错的开始。

But I I I'm not confused that that anyone in the administration, anyone in the White House is confused that legal immigration, immigration is the foundation of the American dream and is the ultimate brand that we want to protect and that's the future we want to protect.

但我很清楚,也不觉得政府里、白宫里有谁不清楚:合法移民、移民,是美国梦的根基,是我们要保护的终极品牌,也是我们要保护的未来。

And I would also say it seems to me that certainly Saxs and other people in the administration know that we have to recruit the world's best and brightest. we should not sacrifice the greatness of the brand.

我也想说,在我看来,Sacks 和政府里其他人肯定知道,我们必须招募全世界最优秀最聪明的人,不该牺牲这个品牌的伟大。

Um, charging $100,000 or let's say, you know, it got lowered to 50 or whatever the case is,

收 $100,000,或者说降到 50,000,或者别的什么数,

it does seem like it it it it tilts the playing field in favor of big companies who can effectively sponsor all these people,

这看起来确实会让竞争环境偏向大公司,它们有能力给这些人做担保,

right?

对吧?

And it's more challenging for the startup ecosystem where people are already super expensive and now I got to pay this fee on top of it.

而创业生态会更吃力,那里的人本来就已经贵得离谱,现在我还得在这之上再付一笔费用。

It also has an un an unintended consequence.

它还有一个意想不到的后果。

It might um accelerate investment outside United States,

它可能会加快美国以外的投资,

right?

对吧?

And so there there are unintended consequences, but like I said,

所以确实有意想不到的后果,但就像我说的,

start somewhere, move towards the right answer, right?

先从某个地方起步,再朝正确答案走,对吧?

You know, often times people want to go directly from a wrong answer, wrong condition.

很多时候,人们是想从一个错误的答案、错误的状态出发,

We don't want this condition where we're at,

我们不想停在现在这个状态,

right?

对吧?

And directly jump to the perfect answer is hard to find,

然后一步跳到完美答案——那种答案很难找,

right?

对吧?

Just start somewhere.

先起步就好。

It's the entrepreneurial way.

这才是创业者的路子。

It's important to me, you know, the president talked about before when he was running for office, he wanted to staple a green card to the, you know, to the diplomas of these STEM students. so smart

这一点对我很重要:总统之前竞选的时候讲过,他想把绿卡直接钉在这些 STEM 学生的毕业证上——那些非常聪明的

people coming to the United States from from China AI researchers studying at Stanford like we want to keep them here we want to get you know and by the way

从中国来美国、在斯坦福读书的 AI 研究者,我们想把他们留下来,我们想争取他们,而且顺便说一句,

if their families can't get here

如果他们的家人过不来,

they're going to leave after a few years so you might even want to make it easier for their families to come here and others are you confident that we have a strategic plan in this administration you know this is a start but your convers conversations they give you confidence that we have a broader strategic plan to make sure we're recruiting the best and the brightest.

他们过几年就会走,所以你甚至该让他们的家人更容易过来,还有其他人。你有信心这届政府有一套战略规划吗?这只是个开始,但你跟他们的那些对话,能让你相信我们有一套更大的战略规划,确保我们在招最优秀最聪明的人?

I don't know that I have an answer for that.

我不确定我有答案。

Okay.

好。

Um but I understand that where we're at is not where we want to be.

但我明白,我们现在所处的位置,不是我们想要的位置。

Yeah.

对。

And I don't think anybody's lost lost their focus on, you know, the American dream, the importance of immigration, the importance of attracting all of the world's best talent to United States, create the conditions for them to stay here.

而且我不觉得有谁忘了这些事:美国梦、移民的重要性、把全世界最好的人才吸引到美国有多重要、创造条件让他们留下来。

I there are things that are done um from time to time that works against what I just described,

确实时不时有些做法,跟我刚说的这些背道而驰,

right?

对吧?

Um making making foreign students uncomfortable,

让外国留学生心里不舒服,

right,

对吧,

and being here in the brand threatens the brand.

还有他们在这里的处境——这会威胁到这个品牌。

Um uh let's not let's not forget that

我们别忘了,

that it's okay to be competitive with China,

跟中国竞争没什么问题,

but be careful not to be tough on Chinese.

但要小心,不要对中国人强硬。

And so we need to make sure that that slippery slope isn't crossed.

所以我们必须确保不踏上那条滑坡。

Yeah.

对。

Um, you know, and so there there all of these things that goes along with with finesse and nuance.

所以这些事都得靠分寸,靠对细微处的把握。

But the fact of the matter is we know where we want to be.

但事实是,我们知道自己想到哪儿去。

We know we're in a difficult situation.

我们知道现在处境艰难。

We don't want to be here and President Trump doesn't have much time to move us in that direction.

我们不想待在这儿,而特朗普总统没有太多时间把我们推到那个方向。

Right.

没错。

And so to the extent that we move in that direction,

所以只要我们是在朝那个方向走,

I believe it's a good start.

我就认为这是个好的开始。

Agreed.

同意。

Yeah.

对。

I heard from a Chinese researcher leading one of our leading labs in the US

我听一位中国研究者说——他在美国一家头部实验室领头——

that three years ago 90% of the top AI researchers graduating from universities in China

三年前,中国高校毕业的顶尖 AI 研究者里有 90%

wanted to come to the United States and did come to the United States to work in our leading labs

想来美国,而且真的来了美国,进我们的头部实验室工作,

and he guessed that today that's closer to 10 or 15%.

而他估计,今天这个比例接近 10% 到 15%。

Right?

对吧?

So seen a precipitous drop.

所以是断崖式下滑。

That's precisely a concern that we have,

这恰恰是我们担心的问题,

right?

对吧?

So have you seen this?

那么你看到这个现象了吗?

Have you you you know you're you're paying attention to both markets.

你两个市场都在盯着。

Do you see this?

你看到了吗?

And what are the things we need to do in order to reverse that?

我们要做些什么才能把它扭转过来?

Definitely see a greater concern of of Chinese students um who who come here and um uh remain here.

确实看到中国学生对来这里、留在这里,顾虑更大了。

Yeah. and or many of them who come here for for school and are thinking about going elsewhere, right?

对,还有很多来这里念书的人在考虑去别的地方,对吧?

Many of them thinking about Europe,

很多人在考虑欧洲,

right?

对吧?

And so so I think I think we need to be super super concerned about this.

所以我觉得我们必须对这件事极其极其警惕。

This is this is

这是,这是

this is a source

这是一种

of existential crisis.

生存危机的源头。

This is definitely the early indicators of a future problem.

这绝对是未来问题的早期指标。

Right.

没错。

Right. you know, smart people's desire to come to to America and smart peop smart students desire to stay,

没错。聪明人想不想来美国、聪明学生想不想留下来,

those are what I would call KPIs.

这些就是我说的 KPI。

Yes.

是的。

Early indicators of future success.

未来成功的早期指标。

Yes.

是的。

I think of it a bit like the Warriors.

我觉得这有点像勇士队。

You know, if they have an advantage of recruiting all the best players in the NBA, right, then they can continue to win championships. Y,

如果他们在招募 NBA 最好的球员上有优势,对吧,那他们就能一直赢总冠军。

but the second that recruiting pipeline,

但一旦那条招募管道,

right, because of the brand of the Warriors gets diminished or something else happens,

对吧,因为勇士的品牌褪色了,或者出了别的什么事,

um, then they're not going to be able to recruit the best future players and you're not going to win championships.

那他们就招不到未来最好的球员,也就赢不了总冠军。

And when I you talk about the American dream so eloquently, that being brand USA,

而你刚才把美国梦讲得那么动人,那就是「美国」这个国家品牌,

right?

对吧?

The right to come here and to to do what you've done.

有权来到这里,做成你做成的那些事。

And you know, so I hope that the feedback to this administration, it's not just the administration, it's also just how we as a country talk about immigration.

所以我希望这些反馈能进到这届政府——也不只是政府,还有我们整个国家怎么谈移民。

That's right.

没错。

Right.

对。

This needs to be the place that welcomes the best and the brightest, that attracts, has a strategic plan for recruiting the best and the brightest and making sure that this is the place that they want to work.

这里必须是欢迎最优秀最聪明的人的地方,能吸引他们,有一套战略规划去招募最优秀最聪明的人,并确保这里是他们愿意工作的地方。

As you know, there's in in there's a there's a phrase, and I didn't hear about this phrase until just a few years ago, China hawks.

你知道,有个说法,我直到几年前才第一次听到:对华鹰派。

Yes.

是的。

And and apparently that if you're a China hawk, you get to wear that label with pride.

而显然,如果你是对华鹰派,你可以带着骄傲戴上这个标签。

It's almost like a badge of honor,

那几乎像一枚荣誉勋章,

right?

对吧?

It's a badge of shame.

那是一枚耻辱勋章。

There's no question it's a badge of shame.

毫无疑问,那是一枚耻辱勋章。

There's no question that although they want what's in the best interest of our country, and we all want what's in the best interest of our country,

毫无疑问,尽管他们想要的是对我们国家最有利的事,我们所有人也都想要对国家最有利的事,

um destroying that pipeline,

但摧毁那条管道,

right,

对吧,

of the American dream,

美国梦的那条管道,

Yeah.

对。

is not patriotic,

并不爱国,

right?

对吧?

They they think they think they're doing the right thing for our country, but it's not patriotic.

他们以为自己在为国家做对的事,但这并不爱国。

Not Not even a little bit.

一点都不爱国。

And so we we need to um continue to be uh the great country we are, to have the confidence,

所以我们要继续做我们本来就是的那个伟大国家,要有

right,

对吧,

of a great country.

一个伟大国家的自信。

Yes.

是的。

Well said.

说得好。

And to have the confidence of a great country and and have somebody who wants to compete with us and to have the attitude, bring it on.

要有一个伟大国家的自信,有人想跟我们竞争,那就是这个态度:尽管来。

Right.

没错。

Right.

没错。

Bring it on.

尽管来。

Right. because I believe in I believe in our people.

对,因为我相信我们的人民。

I believe in our people.

我相信我们的人民。

I believe in the people that are here.

我相信在这里的这些人。

I believe in our culture.

我相信我们的文化。

I believe in our country.

我相信我们的国家。

I believe in our system.

我相信我们的制度。

Bring it on.

尽管来。

And is it your take that that's where the president is?

你觉得总统也是这个立场吗?

Like he's a he's a pragmatist.

他是个务实的人。

He's a he's a believer in the growth and the ability of the United States to compete.

他相信增长,也相信美国有能力竞争。

Um it seems to me that's where he is.

在我看来,他就是这个立场。

There's no question President Trump is the bring it on president.

毫无疑问,特朗普总统就是那个「尽管来」总统。

Right.

没错。

Right.

没错。

And he doesn't seem to me like the reason I'm confident and I've said on this pod that I think he'll get a big deal done with China.

而在我看来他不像……我之所以有信心,而且在这个播客上说过,我认为他会跟中国谈成一笔大交易,

I I really really do hope so.

我真的真的希望如此。

Yeah.

对。

And and I I think he he speaks he he speaks positively um uh uh with great respect and um great eloquence about about his relationship and the importance of China.

而且我觉得,他谈到自己跟中国的关系、谈到中国的重要性时,说得很正面、很有敬意,也讲得很漂亮。

Uh, not one time have I ever heard him say the word decouple, which we heard a lot in the last administration,

我一次都没听他说过「脱钩」这个词,而上一届政府我们听到过很多次,

right?

对吧?

Um, you can't decouple against um the single most the two most important relationships for the next century.

你没法对未来一个世纪里最重要的——那两组最重要的关系搞脱钩。

That doesn't make any sense at all.

这完全说不通。

Decoupling is exactly the wrong concept.

脱钩恰恰是错的概念。

Right?

对吧?

I mean, it seems to me he and Scott Besson are saying, "Listen, we need to make America great.

在我看来,他和 Scott Bessent 是在说:「听着,我们要让美国重新伟大。

We need to re-industrialize America.

我们要让美国重新工业化。

We need to balance and make sure that we have fair trade.

我们要平衡,要确保贸易是公平的。

Um that we protect industries that we need to help build that China helps us do that recognizing that we have helped them do it over the course of the last 25 years.

我们要保护那些我们需要帮着建起来的产业,让中国在这件事上帮我们,同时承认过去 25 年里我们也帮他们做过这件事。

But that ultimately he said the best way to understand me is I'm a great dealmaker.

但归根到底他说,理解我最好的方式是:我是个很会做交易的人。

I make deals right whereas I think in other camps there are people who are iconoclastic or dogmatic.

我做交易。「而我觉得在另一些阵营里,有些人是离经叛道型的,或者是教条式的。

Uh you know it's the mere shimer view of China that there's a great power struggle. one was must win and one must lose versus this idea

那是米尔斯海默式的中国观:大国之争,一方必须赢、一方必须输;与之相对的是另一种想法——

the idea that every country has to look exactly like ours,

那种觉得每个国家都得跟我们一模一样的想法,

right?

对吧?

You know, and we we want diversity.

而我们要的是多样性。

You want America to win, but that doesn't have to come at the expense of poking an eye and telling somebody else they have to lose

你希望美国赢,但这不必以踩别人一脚、告诉对方他们必须输为代价,

because we're that confident.

因为我们有那样的自信。

Yeah, we're that confident.

对,我们有那样的自信。

Because we're that mighty.

因为我们就是那么强大。

Because we're that incredible.

因为我们就是那么了不起。

I've got no trouble, as you know, I've got no trouble working with all my colleagues in the ecosystem, right?

你知道的,我跟这个生态里所有同行合作都没有障碍,对吧?

And notice we just did the ultimate deal,

而且你注意到,我们刚刚做成了那笔终极交易,

right?

对吧?

Partnering with Intel,

跟 Intel 合作,

a company that spent most of its life trying to put us out of business, right?

一家大半辈子都在想把我们搞垮的公司,对吧?

And I had no trouble partnering with them,

而我跟他们合作毫无障碍,

right?

对吧?

You know, and so, and the reason for that is because number one, bring it on.

原因在于,第一,尽管来。

Yes.

是的。

And number two, the future is so much greater.

第二,未来大得多。

It doesn't have to be all us or them.

不一定非得是我们或者他们二选一。

It could be us and them.

可以是我们和他们一起。

Yeah.

对。

But nonetheless, bring it on.

但不管怎样,尽管来。

Yeah.

对。

Agreed.

同意。

You know, you you you mentioned something that's profoundly important to both of us.

你刚才提到一件对我们两个人都极其重要的事。

You and I have talked a lot about this, the American dream,

你和我聊过很多次,就是美国梦,

you know, and it was, I think, Abraham Lincoln who said, "Fundamental to the American dream is the right to rise."

我记得是亚伯拉罕·林肯说过:「美国梦的根本,是向上流动的权利。」

Yeah.

对。

That's right.

没错。

The belief that your kids can do better than you did.

就是那种信念:你的孩子能过得比你更好。

That's right.

没错。

Right.

对。

And you you've experienced the right to rise.

而你自己就体验过向上流动的权利。

We've all experienced the right to rise in America.

在美国,我们每个人都体验过向上流动的权利。

So, yeah, you go to Wikipedia, you look up American dreams, my picture,

是啊,你上维基百科搜「美国梦」,出来的是我的照片,

right?

对吧?

Yeah.

对。

And the ultimate American dream.

而且是最极致的那版美国梦。

And yet we live at this time where because of the nature of these technological systems, we have companies that are going to be worth 10 trillion.

但我们偏偏活在这样一个时代:因为这些技术系统的本质,会出现值 10 trillion 的公司。

We'll probably have individuals that are worth a trillion.

大概还会出现身价 1 trillion 的个人。

Those are the incentives that give people the the encouragement to rise.

正是这些激励,给了人们向上攀升的动力。

But at the same time, when we head into this age of abundance, something that I was deeply worried about was that too many people get left behind.

但同时,在我们走向这个富足时代的路上,我一直深深担心一件事:太多人被落在后面。

Yeah.

对。

Right.

对。

And they feel left out and left behind.

他们觉得自己被排除在外、被甩在后面。

So it makes sense for them to attack this system of capitalism.

所以他们去攻击资本主义这套体系,是说得通的。

Something that you and I worked on together and I'm deeply grateful for was the idea of invest America that we have to start every kid at birth on the capitalist right to rise journey give them a thousand bucks in great companies like Nvidia

你和我一起做过、也让我非常感激的一件事,就是 Invest America 这个想法:我们必须让每个孩子从出生起就踏上资本主义的向上流动之路,给他们 1000 美元,投进 NVIDIA 这样的伟大公司。

social security

社会保障。

and and and open a high etc.

然后给他开个账户,等等。

Um and they benefit

他们从中受益。

right as the as the comp country wins they win and they own it individually they can see it on their

对,国家赢,他们就赢,而且这份资产归他们个人所有,他们自己就能看见——

every kid is a shareholder in the future of America

每个孩子都是美国未来的股东。

of America.

美国的未来。

So on the 200 because of your support and I wanted to take the chance on this podcast and the support of

所以……多亏了你的支持,我想借这期播客的机会说一句,也谢谢那些支持——

Well, I want to thank you for starting it, for driving it.

我要感谢你把这件事发起来、一路推动。

Yeah.

对。

What a great idea.

多好的主意。

And you know, so this

所以说,这件事——

you're a genius.

你是个天才。

The Please this passed

拜托……这个已经通过了,

in the big beautiful bill.

就在 big beautiful bill 里通过的。

Most people don't even realize that yet.

大多数人到现在还没意识到这件事。

Starting in 2026, every child born forever more in the history of this country

从 2026 年开始,这个国家历史上此后出生的每一个孩子,

will start off with an investment account at birth.

一出生就会有一个投资账户。

Yeah.

对。

Seen a thousand bucks in the best American companies and your company has agreed to add to the accounts of not only the kids who work for your employees but maybe even kids of others.

账户里先投入 1000 美元,买进最好的美国公司;而你的公司已经答应,不仅给员工的孩子往账户里加钱,可能还会给别人的孩子加。

I'm going to adopt schools, you know, and lots of philanthropists and companies.

我准备去认领一批学校,还有很多慈善家和公司也会这么做。

We think every company across America

我们认为全美国的每一家公司都——

wonderful way for companies to give back,

这是公司回馈社会的绝好方式,

right?

对吧?

Yeah. as part of the 401k.

对,作为 401k 的一部分。

This seems to me to be part of the change in the social contract that needs to occur because if if we're seeing this exponential progress, we know that the the evolution of government in the social contract needs to keep up with it.

在我看来,这正是社会契约必须发生的那种变化,因为如果我们看到的是这种指数级的进步,那政府和社会契约的演进就必须跟上。

Um obviously President Trump and and bipartisan group in the House and Senate passed this into law.

显然,特朗普总统和众参两院的跨党派团体把它写进了法律。

So maybe just talk to us a little bit when you think about the the pace and magnitude of changes that are coming, right?

所以你能不能跟我们讲讲,你怎么看接下来这些变化的速度和量级,对吧?

Um I know you believe it will be a net good, but there also going to be a bunch of people displaced along the way.

我知道你相信这总体上是好事,但过程中也会有一大批人被挤出原有的位置。

We probably need things like this and other things,

我们大概需要这样的东西,也还需要别的一些东西,

right?

对吧?

In order to, you know, bring everybody along for the journey.

这样才能带着所有人一起走完这段旅程。

There's several things that that President Trump has done and let me just start start there has done that is incredibly good for bringing everybody along.

特朗普总统做了好几件事,我就从这里说起,这些事对带着所有人一起往前走极其有用。

The first thing is reindustrializing America.

第一件是让美国重新工业化。

Yeah.

对。

President Trump, Secretary Lutnik, the you know they're all in behind that all the work that they're doing encouraging companies to come build here in the United States, investing in factories and uh reskilling and upskilling that skilled labor workforce, right?

特朗普总统、Lutnick 部长,他们全都全力支持这件事,他们做的所有工作都在鼓励企业回美国来建厂、投资工厂,还有对技能劳动力的再培训和技能升级,对吧?

Incredibly valuable to our country. um the idea that that we no longer uh uh uh make it only that you get a PhD or you go to you know one of the great schools and only in that way can you build a great life right

这对我们国家极有价值。我们不再让「只有读到博士、或者上了那些名校,才能过上好日子」成为唯一的路,对吧。

and and deserve to have a great living uh we've got to change all that doesn't make any sense we love craft right

才配得上过好日子——这些我们都得改,这毫无道理。我们热爱手艺,对吧。

I love people who make things with their hands and and we're now we're now going to go back and build things

我喜欢用双手做东西的人,现在我们要回过头去动手造东西了。

build magnificent incredible things I love that

造出宏伟、惊人的东西,我特别喜欢这个。

yes

是。

that's going to transform form America.

这会彻底改变美国。

There's no question about that.

这一点毫无疑问。

There's a whole there's a whole a whole band of an economy, a whole band of society that that um uh has been largely left behind because we outsourced everything,

有一整条经济带、一整层社会,因为我们把什么都外包出去了,基本上被落在了后面,

right?

对吧?

Now, I'm not suggesting we insource everything,

我不是说我们要把什么都拿回国内做,

right?

对吧?

You know, all the people arguing about, you know, manufacturing tennis shoes and toothpicks, I mean, you know, that's that's denigrating a perfectly good discussion into some insane level.

那些人争论什么造网球鞋、造牙签,那是把一场本来很好的讨论拉低到荒唐的层次。

You know, we we've got to um recognize that re-industrializ reindustrializing America is is just fundamentally going to be transform transformative number one.

我们得认识到,让美国重新工业化本身就会带来根本性的改变,这是第一点。

Number two,

第二点,

and aspirational.

而且它让人心生向往。

Oh, it's fantastic.

哦,这太棒了。

Elon taking us to Mars, watching spaceships caught with, you know, uh chopsticks out of the sky.

Elon 带我们去火星,看着「筷子」把天上落下来的飞船接住。

This is not only great for the industrializing base of America, it's aspirational for

这不只对美国的工业基础是大好事,它还让人心生向往——

Fantastic.

太棒了。

That's right.

没错。

And then and then, uh, of course AI.

然后当然还有 AI。

Yeah,

对。

it is the greatest equalizer.

它是最大的均衡器。

Just think everybody can have an AI now.

想一想,现在每个人都能有一个 AI。

The the ultimate equalizer.

终极的均衡器。

We've closed the technology divide.

我们已经把技术鸿沟填平了。

Remember the last time that somebody had to learn has wants to use a computer uh for their economic um or career benefit.

想想上一回:一个人想靠计算机改善收入或者职业发展,

They have to learn C++ or C or at least Python.

他必须学 C++ 或者 C,至少也得学 Python。

Now they just have to learn human,

现在他只需要学会「人话」。

you know.

就是这样。

And so, and if you don't know how to program an AI, you tell the AI, "Hi, I don't know how to program an AI.

所以如果你不会给 AI 编程,你就告诉 AI:「嗨,我不会给 AI 编程。

How do I program an AI?

我该怎么给 AI 编程?「

And the AI explains it to you

然后 AI 就给你讲明白,

or does it for you.

或者直接替你做了。

It does it for you.

它直接替你做了。

And so, it's incredible, isn't that right?

所以这太不可思议了,对不对?

It's And we've now closed the technology divide with technology.

我们现在是用技术把技术鸿沟填平了。

Yeah.

对。

This is something that every everybody's got to engage.

这是每个人都得参与进来的事。

You know, OpenAI has 800 million active users.

OpenAI 现在有 800 million 活跃用户。

Um gosh, it really really needs to be 6 billion.

天啊,这个数字真该是 6 billion。

Yeah.

对。

Right.

没错。

It really needs to be 8 billion soon.

很快就该到 8 billion 才对。

And so so I think that that that's number one.

所以我觉得这是第一点。

Then number two, uh and then number three, you know, I think the the um AI will change tasks.

然后第二点,再然后第三点,我觉得 AI 会改变任务。

Yeah.

对。

The thing that people confuse is there are many tasks that will be eliminated.

人们搞混的地方在于:会有很多任务被消灭。

There are many tasks that will actually be created.

但也会有很多任务被创造出来。

But it is very likely that for many people their jobs are gainfully protected.

而很可能的结果是,很多人的工作反而稳稳地保住了。

Right.

没错。

And so, for example, I'm using AI all the time.

比如说,我一直在用 AI。

You're using AI all the time.

你也一直在用 AI。

My analysts are using AI all the time.

我的分析师们一直在用 AI。

My engineers, every one of them use AI continuously.

我的工程师,每一个人都在不停地用 AI。

And we're hiring more engineers.

而我们还在招更多工程师。

We're hiring more people.

我们在招更多人。

We're hiring across the board.

我们各条线上都在招人。

The reason for that is because we have more ideas.

原因是我们有了更多想法。

Yes,

是,

we can now go pursue more ideas.

我们现在可以去追更多想法了。

The reason for that is because our company became more productive.

原因是我们公司的生产力提高了。

And because we became more productive, we became more rich.

因为生产力提高了,我们就更有钱。

We became more rich, we can hire more people to go after those ideas,

因为更有钱,我们就能雇更多人去追那些想法,

right?

对吧?

The I the concept that that AI comes along and therefore there's going to be a mass destruction of jobs starts with the I starts with the premise that we have no more ideas.

「AI 一来,所以工作会大规模消失」,这个说法的起点是一个前提:我们再没有新的想法了。

Right.

没错。

It starts with the premise we have nothing left to do.

它的前提是我们已经没什么事情可做了。

Everything we're doing in our lives today.

我们今天生活里做的所有事情——

Yeah.

对。

This is the end.

到这儿就是终点了。

Yeah.

对。

And if somebody else were to do that one task for me, I have one task left.

而如果我手上只剩一件任务,别人又替我把这件任务做掉了。

Now I have to sit there and wait for something.

那我就只能干坐在那儿等着什么事发生。

Yes.

是。

You know, wait for retirement, sit on my rocking chair.

等退休,坐在摇椅上。

That idea doesn't make sense to me.

这个想法在我看来讲不通。

And so I think I think that intelligence is not a zero sum game.

所以我认为,智能不是零和游戏。

The more intelligent people I'm surrounded by, the more geniuses I'm surrounded by, surprisingly, the more ideas I have, the more problems I imagine that we can go solve,

我身边的聪明人越多、天才越多,说来奇怪,我自己的想法就越多,我能想到、我们可以去解决的问题也越多,

the more work we create, the more jobs we create.

于是我们创造出越多工作,创造出越多岗位。

And so I think for for um I don't know what the world looks like in a million years that's going to be left for my my children.

所以我想说,我不知道一百万年后留给我孩子的世界是什么样。

Um but for the next several decades, my sense is that economy is going to grow.

但接下来这几十年,我的感觉是经济会增长。

Lots of new jobs are going to be created.

会有大量新工作被创造出来。

Every job will be changed.

每一份工作都会改变。

Some job some jobs will be lost.

有些工作会消失。

And um uh we're not going to be, you know, riding horses on streets and those things it'll be fine.

我们不会再回到街上骑马那种日子,不会有事的。

You know, humans are are famously skeptical and terrible at understanding compounding systems and they're even worse at understanding exponential systems that accelerate with size.

人类是出了名地爱怀疑,也出了名地不擅长理解复利系统,而对那些随规模而加速的指数系统,他们理解得更差。

We've talked about exponentials a lot today.

今天我们反复谈到指数级增长。

You know, the great futurist Ray Kerszswhile said in the 21st century, we're not going to have a hundred years of progress.

伟大的未来学家 Ray Kurzweil 说过,21 世纪不会只带来一百年的进步。

We're likely to have 20,000 years of progress.

很可能是 20,000 年的进步。

Right.

没错。

You said earlier, we're so fortunate to be living at this moment and contributing to this moment.

你刚才说,我们太幸运了,能活在这个时刻,还能为这个时刻出一份力。

I'm not going to ask you to look out 10 or 20 or 30 years because I think it's so challenging.

我不打算请你展望 10 年、20 年或者 30 年之后,因为我觉得那太难了。

But when we think about things like robots,

但我们一想到机器人之类的东西——

30 years is easier than 2030.

30 年反而比 2030 年更容易。

Oh, really?

哦,真的吗?

Yeah.

是的。

Yeah.

是的。

Okay.

好。

So, I'll get I I'll grant you license to go out 30. as you think out over the course of I like these shorter time frames because they have to marry bits and atoms bits and Adams the hard part of building this stuff right because everybody's saying it's going to happen

那我就允许你往后看 30 年。当你往前展望的时候——我更喜欢短一点的时间跨度,因为它们必须让比特和原子对上,比特和原子,这才是把这些东西真造出来最难的地方,对吧?因为所有人都在说这件事一定会发生——

is interesting but not helpful

有意思,但没什么用。

exactly but if we have 20,000 years of progress reflect on that statement by Ray reflect on exponentials and how all of our listeners whether you work in government whether you're in a startup whether you're you know running a big company need to be thinking about the accelerating rate of change, the accelerating rate of growth and how you will, you know, be co-intelligent in this new world.

正是。但如果我们真会有 20,000 年的进步——请你谈谈 Ray 这句话,谈谈指数级增长,也谈谈我们所有的听众,不管你在政府工作、在创业公司,还是在经营一家大公司,都该怎么去想这种不断加速的变化速度、不断加速的增长速度,以及你要怎么在这个新世界里跟 AI 协同共智。

Well, there there are a lot of things that that that many people have already said and and um and they're all they're all very sensible.

其实有很多事情早就有很多人说过了,而且都说得很有道理。

I think the the um uh in the next 5 years, one of the things that that is really cool that's going to get solved is the fusion of artificial intelligence and megatronics, robotics.

我觉得未来 5 年里会解决掉一件特别酷的事:人工智能和机电一体化、机器人技术的融合。

And so we're going to have we're going to have, you know, AIS that are that are going to be wandering around us.

所以会有一批 AI 在我们身边走来走去。

And uh we all and that everybody knows uh we all all know that we're going to all grow up with our own R2-D2.

而且大家都知道,我们以后都会跟自己的 R2-D2 一起长大。

Yeah.

是的。

And that R2-D2 remember everything about us and coach us along the way and be our companion.

那个 R2-D2 会记住我们的一切,一路上教我们,陪着我们。

We already know that.

这一点我们已经很清楚了。

Okay. and and so the idea and and and the idea that every human will have their own GPUs associated with them in the cloud um and that they're 8 billion people 8 billion GPUs that's a you know viable outcome.

好。所以这个设想是:每个人在云端都有属于自己的 GPU,8 billion 人,就有 8 billion 个 GPU,这完全是个可行的结果。

Yeah. you know, and so

是的,所以说——

and each having their own model that's fine-tuned for them,

而且每个人都有一个为自己微调过的模型,

fine-tuned for them.

为他自己微调过。

And they that that AI is in the cloud is also embodied in a whole bunch of it's embodied in your car.

而云端那个 AI 同时也具身在一大堆地方——具身在你的车里。

It's embodied in your own robot.

具身在你自己的机器人里。

It's everywhere with you.

它无处不在,一直跟着你。

And so that I think that that future is a very sensible thing. the idea that uh we're we're going to understand the the infinite complexity of biology and understanding the system of biology uh and how it how how to predict it and have digital twins of every everybody um our own digital twin for health care like we have a digital twin for shopping at Amazon why wouldn't we have our digital twin at healthcare of course we would and so uh you know a dig a system that that predicts how our how we're going to age what disease will likely going to have and uh anything that's that's about to happen maybe even next week or you know tomorrow afternoon and predict it early.

所以我觉得那样的未来非常合理。还有一个设想是:我们将会理解生物学那种无限的复杂,理解生物这套系统,理解怎么去预测它,并且给每个人做一个数字孪生——给医疗健康做一个我们自己的数字孪生,就像我们在 Amazon 购物时有一个数字孪生那样,凭什么我们在医疗健康上不该有自己的数字孪生?当然该有。所以会有一套系统,能预测我们会怎么衰老、可能会得什么病,以及任何即将发生的事——甚至是下周、明天下午就要发生的事——提前把它预测出来。

Of course, we're going to have all that.

这一切我们当然都会有。

And so I think all of that is a given.

所以我认为这些都是既定的。

Um I think the the the the part that that I'm asked a lot by uh CEOs that I work with about now given all of that, what happens?

我觉得,现在跟我合作的那些 CEO 问我最多的是:既然这一切都会发生,那接下来会怎样?

What do you do?

你该做什么?

And this is this is the this is a common sense of of things that move fast,

这其实就是个常识:面对跑得飞快的东西,

right?

对吧?

If you if you have a if you have a train that's about to get faster and faster and go exponential, the only thing that you really need to do is get on it.

如果有一列火车马上要越来越快、进入指数级加速,你真正要做的只有一件事:上车。

Yeah.

是的。

And once you get on it, you'll figure everything else out along the way.

一旦上了车,别的事你都会在路上弄明白。

Right.

没错。

And so to predict where that train's going to be,

而要是你去算那列火车会开到哪里,

right,

对,

and try to shoot a bullet at it.

然后朝它射一颗子弹。

Yeah. or predict where that train's going to be and it's going exponentially faster every second and go figure out what intersection to wait for it,

是的。或者说,你去算那列火车会到哪里,而它每一秒都在指数级地变快,然后你再琢磨该在哪个路口等它,

right?

对吧?

That's impossible.

那不可能。

Just get on it while it's going kind of slowly,

趁它还开得慢的时候上车就行,

right?

对吧?

And go exponential along the way.

然后一路跟着它指数级地跑下去。

A lot of people think this just happened overnight.

很多人以为这一切是一夜之间发生的。

You know, you've been you've been at this for 35 years.

而你在这件事上已经干了 35 年。

I remember hearing Larry Pageige say probably around 2005 or 2006 that the end state of Google will be when the machine can predict uh the question before you even answer it before you even ask it and give you the answer without having to look right.

我记得大概 2005 年或者 2006 年,听 Larry Page 说,Google 的终极形态是:机器能在你还没回答、甚至还没问出口之前,就预测出你的问题,然后直接把答案给你,你根本不用去查,对吧。

I heard Bill Gates say in

我听 Bill Gates 在——

because contextually you must be asking about you must be wondering about that.

因为从上下文看,你想问的、你心里琢磨的,肯定就是那件事。

I heard Bill Gates say in 2016 when somebody said hasn't all the things been done we've had the internet we've had cloud we've had mobile social etc.

我听 Bill Gates 在 2016 年说过,当时有人说,该做的事不是都做完了吗?我们有了互联网,有了云,有了移动、社交等等。

He said, "We haven't even started.

他说:「我们连开始都还没开始。」

He said, "What do you think?

他说:「你怎么看?

Why would you say that?

你为什么这么说?」

He said, "We won't even begin until machines go from being dumb calculators to beginning to think for themselves, to think with us,

他说:「机器要先从只会算数的笨家伙,变成开始自己思考、跟我们一起思考,在那之前我们根本谈不上开始。」

right?

对吧?

Um, kind of that is the moment,

差不多这就是那个时刻,

right, that we're in.

对,就是我们正处在的这个时刻。

I think to have leaders like you, leaders like Sam and Elon, Satcha, etc., it's such an extraordinary advantage for this country, right? and to have the cooperation that we see between a system of risk capital that I take, you know, that I'm part of which can provide the risk capital for people to do.

我觉得,能有你这样的领导者,有 Sam、Elon、Satya Nadella 这些领导者,对这个国家是了不起的优势,对吧?而且我们还能看到这样的协作:有一套风险资本体系——我自己也在里面——它能给想做事的人提供风险资本。

We don't we're not relying on government having a Manhattan project.

我们不用靠政府去搞一个曼哈顿计划。

We can actually do this ourselves and together for the benefit of the country.

我们完全可以自己来、一起来,为这个国家的利益去做。

It's an extraordinary time

这是一个了不起的时代。

and at a scale that's unimaginable.

而且规模大到无法想象。

Right.

没错。

Right.

没错。

It's an extraordinary time.

这是一个了不起的时代。

But I also think, you know, one of the things that I'm just grateful is that we have leaders who also understand their responsibility to the fact that we are creating change at an accelerating rate.

但我也觉得,我特别感恩的一点是,我们的领导者同样明白自己的责任——我们正在以不断加速的速度制造变化。

And we know while it will most likely be great for the vast majority, there'll be challenges along the way.

我们也知道,虽然这对绝大多数人很可能是好事,但一路上总会有挑战。

And we'll deal with those as they come

挑战来了,我们就去应对,

and raise the floor for everybody and make sure that this is an a win, not just for some elite bureaucrats at the top hanging out in Silicon Valley.

同时把所有人的下限抬上去,确保这是一场胜利,而不只是硅谷高层那些精英官僚的胜利。

And don't scare them.

而且别把大家吓着。

Bring them along.

带着大家一起走。

Don't scare them.