**Dwarkesh Patel:** 我们看到大量软件公司的估值暴跌,因为人们预期人工智能(AI)将把软件变成大宗商品。有一种可能比较天真的思考方式是这样的:英伟达(NVIDIA)把 GDS2 文件发给台积电(TSMC),台积电制造逻辑芯片(logic dies)和交换芯片(switches),再把它们与 SK 海力士(SK Hynix)、美光(Micron)和三星(Samsung)生产的高带宽内存(HBM)封装在一起,然后运到台湾的代工厂(ODM)组装成机架。所以说到底,英伟达本质上是在做软件,然后由别人来制造。如果软件被大宗商品化了,英伟达会不会也被大宗商品化?
**Dwarkesh Patel:** We've seen the valuations of a bunch of software companies crash because people are expecting AI to commoditize software. And there's a potentially naive way of thinking about things which is like look Nvidia sends a GDS2 file to TSMC. TSMC builds the logic dies. It builds the switches. Then it packages them with the HBM that SK Hynix and Micron and Samsung make. Then it sends it to an ODM in Taiwan where they assemble the racks. And so Nvidia is fundamentally making software that other people are manufacturing. And if software gets commoditized, does Nvidia get commoditized?
**Jensen Huang:** 归根结底,总得有什么东西把电子(electrons)转化为代币(tokens)。这个转化过程——没有什么能替代的——把电子转化为代币,并让这些代币随时间变得越来越有价值。我不认为这件事很容易被完全大宗商品化。从电子到代币的转化是一段令人难以置信的旅程,而让一个代币——你知道,这就像让一个分子比另一个分子更有价值一样,让一个代币比另一个代币更有价值。要让这个代币变得有价值,需要投入大量的艺术性、工程、科学和发明创造——显然,我们正在实时见证这一切。
**Jensen Huang:** Well, in the end, something has to transform electrons to tokens. That transformation -- there's no -- the transformation of electrons to tokens and making those tokens more valuable over time. I don't think that that's hard to completely commoditize. The transformation from electrons to tokens is such an incredible journey, and making that token -- you know, it's like making one molecule more valuable than another molecule, making one token more valuable than another. The amount of artistry, engineering, science, invention that goes into making that token valuable -- obviously, we're watching it happening in real time.
**Jensen Huang:** 所以这个转化过程、制造过程,以及其中涉及的所有科学,远远没有被深入理解,这段旅程也远未结束。所以我怀疑大宗商品化会发生。当然,我们会让它变得更高效。事实上,关于英伟达——实际上你提问的方式正是我对我们公司的心智模型(mental model)。输入是电子,输出是代币,中间是英伟达。我们的工作是做尽可能多的必要之事、尽可能少的冗余之事,来实现这种转化,让它具备令人难以置信的能力。
**Jensen Huang:** And so the transformation, the manufacturing, all of the science that goes in there is far from deeply understood and the journey is far from over. And so I doubt that it will happen. We're going to make it more efficient, of course. I mean the whole thing about Nvidia -- in fact the way that you frame the question is my mental model of our company. The input is electrons, the output is tokens. In the middle is Nvidia. And our job is to do as much as necessary, as little as possible, to enable that transformation to be done at incredible capabilities.
**Jensen Huang:** 我说"尽可能少"的意思是——凡是我不需要做的事情,我就和合作伙伴配合,把它纳入我的生态系统。如果你看看今天的英伟达,我们可能拥有上下游供应链中最大的合作伙伴生态。所有的计算机公司、所有的应用开发者、所有的模型制造商——如果你愿意这么说的话,AI 是一个五层蛋糕,而我们在全部五个层面都有生态系统。所以我们尽量少做事情,但我们必须做的那部分,事实证明,难度大得惊人。
**Jensen Huang:** And what I mean by "as little as possible" -- whatever I don't need to do, I partner with somebody and I make it part of my ecosystem. And if you look at Nvidia today, we probably have the largest ecosystem of partners both in supply chain upstream, supply chain downstream. All of the computer companies and all the application developers and all the model makers -- AI is a five layer cake if you will, and we have ecosystems across the entire five layers. And so we try to do as little as possible, but the part that we have to do, as it turns out, is insanely hard.
**Jensen Huang:** 我不认为那部分会被大宗商品化。事实上,我也不认为企业软件公司、做工具的公司会被大宗商品化——你知道,今天大多数软件公司都是工具制造商。有些不是,有些是工作流程固化系统(workflow codification systems)。但对很多公司来说,它们就是工具制造商。比如说,Excel 是工具,PowerPoint 是工具。Cadence 做工具,Synopsys 做工具。
**Jensen Huang:** I don't think that that gets commoditized. In fact, I also don't think that the enterprise software companies, the tools makers -- you know, most of the software companies today are tools makers. Some of them are not, but some of them are workflow codification systems. But for a lot of companies they're tool makers. For example, you know, Excel is a tool, PowerPoint's a tool. Cadence makes tools, Synopsis makes tools.
**Jensen Huang:** 我实际上看到的恰恰与人们看到的相反。我认为智能体(agents)的数量将呈指数级增长。工具使用者的数量也将呈指数级增长,而且很可能所有这些工具的实例数量都会暴增。Synopsys 的设计编译器(design compiler)的实例数量很可能会暴增。使用布局规划工具(floor planners)、我们所有布局工具和设计规则检查器(design rule checkers)的智能体数量——今天我们受限于工程师的数量。明天这些工程师将被一大批智能体所辅助。我们将以前所未有的方式探索设计空间,而这些智能体会想要使用我们今天就在用的工具。
**Jensen Huang:** I actually see the opposite of what people see. I think the number of agents are going to grow exponentially. The number of tool users are going to grow exponentially and it's very likely that the number of instances of all these tools are going to skyrocket. It is very likely the number of instances of Synopsis design compiler is going to skyrocket. The number of agents that are going to be using the floor planners and all of our layout tools and our design rule checkers -- today we're limited by the number of engineers. Tomorrow those engineers are going to be supported by a bunch of agents. We're going to be exploring the design space like you've never seen before, and they're going to want to use the tools that we use today.
**Jensen Huang:** 所以我认为工具使用会让这些软件公司的业务暴增。之所以这还没发生,是因为智能体还不够擅长使用它们的工具。所以要么这些公司自己来构建智能体,要么智能体会变得足够强大来使用那些工具。我认为最终会是两者的结合。
**Jensen Huang:** And so I think tool use is going to cause these software companies to skyrocket. The reason why it hasn't happened yet is because the agents aren't good enough at using their tools yet. And so either these companies are going to build the agents themselves or agents are going to get good enough to be able to use those tools. And I think it's going to be a combination of both.
**Dwarkesh Patel:** 我记得在你们最新的财报中,你们与代工厂、内存厂商、封装厂商等的采购承诺差不多有一千亿美元——然后 Semi Analysis 报道说你们将有 2500 亿美元的这类采购承诺。一种解读是,英伟达真正的护城河(moat)在于你们已经锁定了多年的稀缺组件,别人可能有加速器,但他们真的能拿到内存来制造它吗?他们真的能拿到逻辑芯片来制造它吗?这才是英伟达未来几年真正的大护城河。
**Dwarkesh Patel:** I think in your latest filings you had almost a hundred billion dollars in purchase commitments with people -- foundries, memory, packaging -- and then Semi Analysis has reported that you will have $250 billion of these kinds of purchase commitments. And so one interpretation is Nvidia's moat is really that you've locked up many years of these scarce components, that somebody else might have an accelerator but can they actually get the memory to build it? Can they actually get the logic to build it? And this is really Nvidia's big moat for the next few years.
**Jensen Huang:** 这确实是我们能做到、而别人很难做到的事情之一。我们之所以在上游做出巨额承诺——其中一部分是显性的,就是你提到的那些承诺,还有一些是隐性的。比如说,上游的很多投资是我们的供应链伙伴做出的,因为我对那些 CEO 们说:"让我告诉你这个行业会有多大,让我解释给你听为什么,让我跟你一起推理,让我给你看我看到的东西。"
**Jensen Huang:** Well, it's one of the things that we can do that is hard for someone else to do. The reason why we've made enormous commitments upstream -- some of it is explicit, these commitments that you mentioned, some of it is implicit. For example, a lot of the investments that are upstream are made by our supply chain because I said to the CEOs, "Let me tell you how big this industry is going to be and let me explain to you why and let me reason through it with you and let me show you what I see."
**Jensen Huang:** 因此,通过这个告知、激励、与各行各业上游 CEO 们对齐的过程,他们愿意做出投资。那为什么他们愿意为我做投资而不为别人做呢?原因是他们知道我有能力买下他们的供给,并通过我的下游将其销售出去。英伟达的下游供应链和下游需求是如此之大,他们才愿意在上游进行投资。
**Jensen Huang:** And so as a result of that process of informing, inspiring, aligning with CEOs of all different industries upstream, they're willing to make the investments. Now why are they willing to make the investments for me and not someone else? And the reason for that is because they know that I have the capacity to buy their supply and sell it through my downstream. The fact that Nvidia's downstream supply chain and our downstream demand is so large, they're willing to make the investment upstream.
**Jensen Huang:** 所以如果你看 GTC 大会,人们对 GTC 的规模和参与者阵容感到惊叹。这是一个 360 度的——整个 AI 宇宙汇聚在一处。它们都汇聚在一处,是因为它们需要彼此看见。我把它们聚在一起,这样下游可以看到上游,上游可以看到下游,所有人都能看到 AI 的所有进展。更重要的是,它们都能见到 AI 原生公司(AI natives)和所有正在崛起的 AI 初创公司,以及所有令人惊叹的正在发生的事情,这样它们就能亲眼看到我告诉它们的所有事情。
**Jensen Huang:** And so if you look at GTC, people are marveled by the scale of GTC and the people that go. It's a 360-degree -- the entire universe of AI all in one place. And they're all in one place because they need to see each other. I bring them together so that the downstream could see the upstream. The upstream could see the downstream and all of them could see all the advances in AI. And very importantly they can all meet the AI natives and all the AI startups that are being built and all the amazing things that are happening, so that they could see firsthand all the things that I tell them.
**Jensen Huang:** 所以我花大量时间,直接或间接地向我们的供应链、合作伙伴和生态系统传达摆在我们面前的机遇。你知道,我的大多数主题演讲——有些人总是说,Jensen,在大多数主题演讲中就是一个接一个的发布、一个又一个的公告。但我们的演讲——总有一部分有点"折磨人",某种程度上看起来像是在上课。事实上那正是我心里在想的。我需要确保整个上下游供应链、整个生态系统理解即将到来的是什么、为什么会来、什么时候来、规模会有多大,并且能够像我一样系统性地去推理它。
**Jensen Huang:** And so I spend a lot of my time informing directly or indirectly our supply chain and our partners and our ecosystem about the opportunity that's in front of us. You know, most of my keynotes -- some people always say, you know, Jensen, in most keynotes it's like one announcement after another announcement after another announcement. Our keynotes -- there's always a part of it that's a little torturous in the sense that it almost comes across like an education. And in fact that's exactly on my mind. I need to make sure that the entire supply chain upstream and downstream, the ecosystem, understands what is coming at us, why it's coming, when it's coming, how big is it going to be, and be able to reason about it systematically just like I reason about it.
**Jensen Huang:** 所以你所说的那个护城河——我们有能力为未来而建。如果我们接下来几年的规模是万亿美元级别,我们拥有相应的供应链。没有我们的触达能力,我们业务的周转速度——你知道,就像有现金流(cash flow)一样,也有供应链流(supply chain flow)。有周转(turns)。如果业务周转低,没有人会为一个架构建设供应链。我们之所以能够维持这样的规模,唯一的原因是我们的下游需求如此之大。他们看到了,他们听到了,他们看到一切都在到来。这就让我们能够在我们所能做到的规模上做到我们所能做的事情。
**Jensen Huang:** And so I think the moat as you describe it -- we're able to build for a future. If our next several years is a trillion dollars in scale, we have the supply chain to do it. Without our reach, the velocity of our business -- you know, just as there's cash flow there's supply chain flow. There are turns. Nobody's going to build a supply chain for an architecture if the business turns are low. And so our ability to sustain the scale is only because our downstream demand is so great. And they see it and they all hear about it. They see it all coming. And so that allows us to do the things that we're able to do at the scale we're able to do.
**Dwarkesh Patel:** 我确实想更具体地了解上游是否能跟上。这么多年来你们的收入一直在逐年翻倍。你们提供给世界的浮点运算能力(flops)每年增长超过三倍。
**Dwarkesh Patel:** I do want to understand more concretely whether the upstream can keep up. For many years now you guys have been 2x-ing revenue year-over-year. You guys have been more than tripling the amount of flops you're providing to the world year over year.
**Jensen Huang:** 而且在现在这个规模上还能翻倍,这确实令人难以置信。
**Jensen Huang:** And 2x-ing at the scale now is really incredible.
**Dwarkesh Patel:** 没错。那么你看逻辑芯片——比如你们是台积电 N3 制程节点上最大的客户,也是——今年 AI 整体将占 N3 产能的 60%,根据一些分析,明年将占到 86%。如果你已经占了大多数,你怎么翻倍?而且要年复一年地翻倍?那么我们是否已经进入了一个 AI 算力增长速度必须因上游而放缓的阶段?你看到了什么出路——我们最终怎么做到每年多建两倍的晶圆厂(fabs)?
**Dwarkesh Patel:** Exactly. So then you look at logic -- say you're the biggest customer on TSMC's N3 node, and you're one of the biggest on -- AI as a whole this year is going to be 60% of N3. It's going to be 86% next year according to some analysis. How do you 2x if you're the majority? And how do you do that year-over-year? So are we in a regime now where the growth rate in AI compute has to slow because of upstream? Do you see a way to get around these -- how do we build 2x more fabs year-over-year ultimately?
**Jensen Huang:** 是的,在某种程度上,瞬时需求确实大于全球上下游的供给。在任何时刻我们都可能被水管工的数量所限制。
**Jensen Huang:** Yeah, at some level the instantaneous demand is greater than the supply upstream and downstream in the world. And it could be at any instant we could be limited by the number of plumbers.
**Dwarkesh Patel:** 嗯嗯。
**Dwarkesh Patel:** Mhm.
**Jensen Huang:** 这实际上真的会发生。
**Jensen Huang:** Which actually happens.
**Dwarkesh Patel:** 水管工会被邀请参加明年的 GTC。
**Dwarkesh Patel:** The plumbers are invited to next year's GTC.
**Jensen Huang:** 哈哈,这主意不错。但这其实是一个好的状态。你希望一个市场、一个行业的瞬时需求大于整个行业的总供给。反过来的情况显然就不太妙了。如果差距太大,如果某个特定组件差得太远,整个行业显然会蜂拥而上。
**Jensen Huang:** Yeah. You know, by the way, great idea. But that's a good condition. You want a market, you want an industry where the instantaneous demand is greater than the total supply of the industry. The opposite is obviously less good. If we're too far apart, if one particular item, one particular component is too far away, obviously the industry swarms it.
**Jensen Huang:** 比如说,你注意到人们不再怎么谈论 CoWoS 封装技术了。
**Jensen Huang:** So for example, notice people aren't talking very much about CoWoS anymore.
**Dwarkesh Patel:** 是的。
**Dwarkesh Patel:** Yeah.
**Jensen Huang:** 原因是过去两年我们拼命攻克它,翻倍、翻倍、再翻倍——翻了好几倍。现在我觉得我们状况相当不错。台积电现在也意识到 CoWoS 的供应必须跟上逻辑芯片和内存的需求。所以他们在以与扩展逻辑芯片相同的水平来扩展 CoWoS 以及未来的封装技术,这太棒了,因为在很长一段时间里 CoWoS 都是相当特殊的小众技术。HBM 也是相当小众的。但它们已经不再小众了——人们现在认识到它们是主流计算技术。
**Jensen Huang:** And the reason for that is because for two years we swarmed the living daylights out of it and we doubled, doubled, doubled -- several doubles. And now I think we're in a fairly good shape. And TSMC now knows that CoWoS supply has to keep up with the rest of the logic demand and the memory demand. And so they're scaling CoWoS and they're scaling future packaging technologies at the same level as they scale logic, which is terrific because for a long time CoWoS was rather specialty. And HBM was rather specialty. But they're not specialties anymore -- people now realize they're mainstream computing technology.
**Jensen Huang:** 当然,我们现在能够影响更大范围的供应链了。过去,在 AI 革命刚开始的时候,我现在说的这些话,五年前我就在说了,有些人信了,并且投资了。比如 Sanjay 和美光团队。我仍然清楚地记得那次会议,我非常明确地说了将会发生什么以及为什么会发生,还有对今天的预测。他们确实全力投入了,我们在 LPDDR、HBM 内存方面进行了合作。他们真的大力投资了,这显然对公司带来了巨大回报。
**Jensen Huang:** And of course we're now much more able to influence a larger scope of our supply chain. In the past, you know, in the beginning of the AI revolution all the things that I say now I was saying 5 years ago and some people believed in it and invested in it. For example, Sanjay and the Micron team. I still remember the meeting really well where I was clear about exactly what's going to happen and why it's going to happen, and the predictions of today. And they really doubled down on it and we partnered with them across LPDDR, across HBM memories. They really invested in it and it obviously has been tremendous for the company.
**Jensen Huang:** 有些人来得稍晚一些,但现在大家都到了。所以我认为每一代技术、每一个瓶颈,都会得到大量关注。现在我们是在提前好几年预取(prefetching)瓶颈。比如说,我们与 Lumentum、Coherent 以及整个硅光子(silicon photonics)生态系统的投资——过去几年我们真的重塑了硅光子的生态系统和供应链。我们围绕台积电建立了一整套供应链,与他们在共封装光学(CPO)上合作,发明了大量技术。我们把那些专利授权给供应链。保持开放。
**Jensen Huang:** Some people came a little bit later, but they're all here now. And so I think each one of these generations, each one of these bottlenecks, gets a great deal of attention. And now we're prefetching the bottlenecks years in advance. So for example, the investments that we've done with Lumentum and Coherent and all of the silicon photonics ecosystem -- the last several years we really reshaped the ecosystem and the supply chain for silicon photonics. We built up an entire supply chain around TSMC. We partnered with them on CPO, invented a whole bunch of technology. We licensed those patents to the supply chain. Keep it nice and open.
**Jensen Huang:** 所以我们正在通过发明新技术、新工作流程、新测试设备、双面探测(double-sided probing)、投资公司、帮助它们扩大产能来准备供应链。你可以看到我们在努力塑造生态系统,让它准备好,让供应链准备好来支撑这个规模。
**Jensen Huang:** And so we're preparing the supply chain through invention of new technologies, new workflows, new testing equipment, double-sided probing, investing in companies, helping them scale up their capacity. And so you could see that we're trying to shape the ecosystem so that it's ready, the supply chain so that it's ready to support the scale.
**Dwarkesh Patel:** 看起来有些瓶颈比其他的更容易解决。扩大 CoWoS 产能和扩大——
**Dwarkesh Patel:** It seems like some bottlenecks are easier than others. And so scaling up CoWoS versus scaling up --
**Jensen Huang:** 我刚才说的其实是最难的那个。
**Jensen Huang:** I went to the hardest one by the way.
**Dwarkesh Patel:** 哪个?
**Dwarkesh Patel:** Which is?
**Jensen Huang:** 水管工。
**Jensen Huang:** Plumbers.
**Dwarkesh Patel:** 哈哈,确实。
**Dwarkesh Patel:** Yeah, it's true.
**Jensen Huang:** 是的。水管工和电工。原因是——这也是我对所有末日论调(doomers)的一个担忧——那些描述工作终结、岗位消亡的末日言论。其中一个问题是,如果我们让人们不想当软件工程师,我们就会缺软件工程师。
**Jensen Huang:** Yeah. Yeah. I actually went to the hardest one. Yeah. Plumbers and electricians. And the reason for that is because -- and this is one of the concerns that I have about all the doom -- the doomers describing the end of work and killing of jobs. You know, one of the things that, if we discourage people from being software engineers, we're going to run out of software engineers.
**Jensen Huang:** 同样的预言十年前就有了,一些末日论者当时说,无论如何千万不要当放射科医生。你可能听过那些说法——一些那样的视频至今还在网上。放射科将是第一个被淘汰的职业。世界不再需要放射科医生了。结果怎样?我们现在恰恰缺放射科医生。
**Jensen Huang:** And the same prediction 10 years ago, some of the doomers were saying that we're telling people whatever you do, don't be a radiologist. And you might hear some of those -- some of those videos are still on the web. You know, radiology is going to be the first career to go. Nobody's -- the world's not going to need any more radiologists. Guess what? We're short of radiologists.
**Dwarkesh Patel:** 嗯,好吧。但回到这个话题——有些东西你可以扩展产能,有些则——比如你到底怎样才能每年制造出两倍数量的逻辑芯片?归根结底,内存和逻辑芯片的瓶颈在于 EUV 光刻机(EUV)。你怎么每年获得两倍数量的 EUV 机器?年复一年地?
**Dwarkesh Patel:** Oh, but okay. So, going back to this point about -- well some things you can scale, other things -- like how do you actually manufacture 2x the amount of logic a year? Ultimately that's bottlenecked by -- memory and logic are bottlenecked by EUV. How do you get 2x as many EUV machines a year? Year over year.
**Jensen Huang:** 是的。
**Jensen Huang:** Yeah.
**Dwarkesh Patel:** 年复一年地。
**Dwarkesh Patel:** Year over year.
**Jensen Huang:** 这些都不是不可能快速扩展的。你只需要——你可以在两三年内完成所有这些。你只需要一个需求信号。一旦你能造一台,你就能造十台;一旦你能造十台,你就能造一百万台。所以这些东西复制起来并不难。
**Jensen Huang:** None of that is impossible to scale quickly. You just need to -- you could do all of that within two or three years. You just need a demand signal. Once you can build one you can build 10 and once you can build 10 you can build a million. And so these things are not hard to replicate.
**Dwarkesh Patel:** 你在供应链上会深入到什么程度?你会直接去找 ASML 说,嘿,如果我看三年后,英伟达要实现两万亿美元的年收入,我们需要多得多的 EUV 机器吗?
**Dwarkesh Patel:** How far down the supply chain do you go? Do you go to ASML and say hey, if I look out three years from now, for Nvidia to be generating two trillion in a year in revenue, we need way more EUV machines?
**Jensen Huang:** 有些我必须直接去找,有些是间接的。有些——如果我能说服台积电,ASML 自然就会被说服。所以我们得考虑关键的瓶颈节点。但如果台积电被说服了,几年内你就会有充足的 EUV 机器。
**Jensen Huang:** Some of them I have to go to directly, some of them are indirect. And some of them -- if I can convince TSMC, ASML will be convinced. And so we have to think about the critical pinch points. But if TSMC is convinced, you'll have plenty of EUV machines in a few years.
**Jensen Huang:** 所以我的观点是,没有任何瓶颈会持续超过两三年。没有。与此同时,我们正在把计算效率提升 10 倍、20 倍——从 Hopper 到 Blackwell 的情况来看,有些是 30 倍、50 倍。我们在不断发明新的算法,因为 CUDA 有极高的灵活性。我们在开发各种新技术,在提升容量的同时驱动效率提升。所以这些事情没有一个让我担心。
**Jensen Huang:** And so my point is that none of the bottlenecks last longer than a couple two, three years. None of them. And meanwhile, we're improving computing efficiency by 10x, 20x -- in the case of Hopper to Blackwell, some 30x, 50x. We're coming up with new algorithms because CUDA is so flexible. We're developing all kinds of new techniques so that we drive efficiency in addition to increasing capacity. And so those are things that none of that worries me.
**Jensen Huang:** 让我担心的是我们下游的事情。阻碍能源供应的能源政策——没有能源就不可能创造一个产业。没有能源就不可能创造一整个新的制造业。我们想要让美国重新工业化(re-industrialize)。我们想把芯片制造、计算机制造和封装带回来,我们想造新东西,比如电动汽车(EVs)、机器人,我们想建 AI 工厂(AI factories)。没有能源就造不了这些东西,而这些东西需要很长时间。但更多的芯片产能——那是两三年的问题。更多的 CoWoS 产能——两三年的问题。
**Jensen Huang:** It's the stuff that's downstream from us. Energy policies that prevent energy from -- you can't create an industry without energy. You can't create a whole new manufacturing industry without energy. We want to re-industrialize the United States. We want to bring back chip manufacturing and computer manufacturing and packaging and we want to build new things like EVs and robots and we want to build AI factories. And you can't build any of these things without energy and those things take a long time. But more chip capacity -- that's a two, three year problem. More CoWoS capacity -- two, three year problem.
**Dwarkesh Patel:** 有意思。我感觉我的其他嘉宾有时候会告诉我完全相反的判断,在这件事上我确实没有足够的技术知识来做出裁决。
**Dwarkesh Patel:** Interesting. I feel like I have guests tell me the exact opposite thing sometimes and I don't -- in this case I just don't have the technical knowledge to adjudicate.
**Jensen Huang:** 好在你正在和专家对话。
**Jensen Huang:** Well the beautiful thing is you're talking to the expert.
**Dwarkesh Patel:** 哈哈,没错没错。好的,我想问问你的竞争对手们。
**Dwarkesh Patel:** Yeah, true, true. Okay. I want to ask about your competitors.
**Jensen Huang:** 好的。
**Jensen Huang:** Yeah.
**Dwarkesh Patel:** 如果你看 TPU,可以说世界上排名前三的模型中有两个——Claude 和 Gemini——是在 TPU 上训练的。这对英伟达未来意味着什么?
**Dwarkesh Patel:** So, if you look at TPU, arguably two out of the top three models in the world, Claude and Gemini, were trained on TPU. What does that mean for Nvidia going forward?
**Jensen Huang:** 我们构建的东西非常不同。英伟达构建的是加速计算(accelerated computing),而不是张量处理单元(tensor processing unit)。加速计算可以用于各种各样的事情。你知道,分子动力学(molecular dynamics)、量子色动力学(quantum chromodynamics),还有数据处理、数据帧(data frames)、结构化数据、非结构化数据。还用于流体动力学(fluid dynamics)、粒子物理学(particle physics),此外,我们还将它用于 AI。
**Jensen Huang:** Well, we have a very different -- we built a very different thing. What Nvidia built is accelerated computing, not a tensor processing unit. And accelerated computing is used for all kinds of things. You know, molecular dynamics and quantum chromodynamics and it's used for data processing, data frames, structured data, unstructured data. It's used for fluid dynamics, particle physics, and in addition, we use it for AI.
**Jensen Huang:** 所以加速计算要多元化得多,虽然 AI 是当今的热门话题,显然也非常重要和有影响力,但计算远不止于此。英伟达所做的是重新发明了计算方式,从通用计算(general purpose computing)转向加速计算。我们的市场覆盖范围远远超过任何 TPU、任何专用芯片(ASIC)所能达到的。
**Jensen Huang:** And so accelerated computing is much more diverse, and although AI is the conversation today and is obviously very important and impactful, computing is much broader than that. And what Nvidia has done is reinvented the way computing is done from general purpose computing to accelerated computing. Our market reach is far greater than any TPU, any ASIC, can possibly have.
**Jensen Huang:** 所以如果你看看我们的定位,我们是唯一一家能加速各种类型应用的公司。我们有一个庞大的生态系统,各种框架和算法都在英伟达上运行。而且因为我们的计算机是设计给别人来运营的,任何运营者都可以购买我们的系统。大多数自建系统——你得自己当运营者,因为它从来就不是设计成灵活到足以让别人来运营的。
**Jensen Huang:** And so if you look at our position, we're the only company that accelerates applications of all kinds. We have a gigantic ecosystem and so all kinds of frameworks and algorithms all run on Nvidia. And because our computers are designed to be operated by other people, anyone who's an operator could buy our systems. Most of these home-built systems -- you have to be your own operator because it was never designed to be flexible enough for other people to operate.
**Jensen Huang:** 正因为任何人都可以运营我们的系统,我们存在于每一个云平台中,包括谷歌、亚马逊(Amazon)、Azure 和 OCI。所以无论你是想运营出租还是自用——如果你想运营出租,你最好有一个庞大的、跨多个行业的客户生态来作为承接方。如果你想自用,我们显然也能帮你。比如说帮 Elon 做 xAI。
**Jensen Huang:** And so as a result of the fact that anybody can operate our systems, we're in every cloud including Google and Amazon and Azure and OCI. And so whether you want to operate it to rent or operate it for yourself -- if you want to operate it to rent, you better have a large ecosystem of customers in many industries that will be the offtakers. If you want to operate it for yourself, we obviously have the ability to help you. Like for example for Elon with xAI.
**Jensen Huang:** 因为我们能为任何行业的任何公司赋能运营,你可以用它在礼来(Lilly)建造一台用于科研和药物发现的超级计算机。我们可以帮他们运营自己的超级计算机,并将其用于我们所加速的药物发现和生物科学的全部多样性领域。
**Jensen Huang:** Because we could enable operators in any company in any industry, you could use it to build a supercomputer for scientific research and drug discovery at Lilly. And so we can help them operate their own supercomputer and use it for the entire diversity of drug discovery and biological sciences that we accelerate.
**Jensen Huang:** 所以有大量的应用领域是 TPU 做不了而我们能做的,因为英伟达用 CUDA 也打造了一个出色的张量处理单元,但它能覆盖数据处理和计算以及 AI 等等的每一个生命周期。所以我们的市场机会大得多,我们的覆盖范围也广得多。而且因为我们有如此大的——我们基本上支持世界上每一个应用——现在你可以在任何地方建英伟达的系统,而且知道一定会有客户。所以这是一件非常不同的事情。
**Jensen Huang:** And so there's just a whole bunch of applications that we can address that you can't do with TPUs, because Nvidia's built CUDA as a fantastic tensor processing unit as well, but it does every lifecycle of data processing and computing and AI and so on and so forth. And so our market opportunity is just a lot larger. Our reach is a lot greater. And because we have such a large -- we basically support every application in the world -- now you could build Nvidia systems anywhere and know that there will be customers for it. And so it's a very different thing.
**Dwarkesh Patel:** 这个问题可能比较长,但你知道,你们有惊人的收入,而这些收入主要不是靠制药和量子计算赚了每季度 600 亿美元。你们赚这么多是因为 AI 是一项前所未有的技术,正以前所未有的速度增长。那么问题就是,什么对 AI 本身来说是最好的。
**Dwarkesh Patel:** This is going to be sort of a long question, but you know you have spectacular revenue and this revenue is mostly -- you're not making 60 billion a quarter from pharma and quantum. You're making it because AI is unprecedented technology that is growing unprecedentedly fast. And so then the question is what is best for AI specifically.
**Dwarkesh Patel:** 我不了解细节,但我和做 AI 研究的朋友们聊过,他们说,当我用 TPU 的时候,它就是一个大型的脉动阵列(systolic array),非常适合做矩阵乘法(matrix multiplies);而 GPU 非常灵活,当你有大量分支(branching)、不规则的内存访问(irregular memory access)时表现出色。但 AI 是什么呢?就是反反复复地做这些非常可预测的矩阵乘法,你不需要把任何芯片面积(die area)花在线程束调度器(warp schedulers)、线程和内存体之间的切换上。所以 TPU 真正为当前这波收入和计算用例增长的主体做了优化。我想听听你怎么看。
**Dwarkesh Patel:** And I'm not in the details, but I talked to my AI researcher friends and they say look, when I use a TPU it's this big systolic array that's perfect for doing matrix multiplies, whereas a GPU is very flexible. It's great when you have lots of branching, when you have irregular memory access. But what is AI? Just these very predictable matrix multiplies again and again and again, and you don't have to give up any die area for warp schedulers, for switches between threads and memory banks. And so the TPU is really optimized for the majority, the bulk of this growth in revenue and use case for compute that is coming online right now. I wonder how you react to that.
**Jensen Huang:** 矩阵乘法是 AI 的重要组成部分,但不是全部。如果你想提出一种新的注意力机制(attention mechanism),或者想用不同的方式做解耦(disaggregate),如果你想发明一种全新类型的架构——比如混合状态空间模型(hybrid SSM)——如果你想创造一个融合了扩散模型(diffusion)和自回归模型(autoregressive)的模型,你就需要一个通用可编程的架构,而我们能运行你能想到的一切。这就是优势所在。它让新算法的发明变得容易得多。
**Jensen Huang:** Matrix multiplies is an important part of AI but it's not the only part of AI. And if you want to come up with a new attention mechanism, or if you want to disaggregate in a different way, if you want to come up with a whole new type of architecture altogether -- for example a hybrid SSM -- if you want to create a model that fuses diffusion and autoregressive somehow, you want an architecture that's just generally programmable and we run everything you can imagine. And so that's the advantage. It allows for invention of new algorithms a lot more easily.
**Jensen Huang:** 正因为它是可编程系统,发明新算法的能力才是 AI 进步如此之快的原因。你知道,TPU 和其他东西一样受摩尔定律(Moore's law)的影响。我们知道摩尔定律每年大约提升 25%。所以要真正实现 10 倍的飞跃、100 倍的飞跃,唯一的办法就是每年从根本上改变算法和计算方式。这就是英伟达的根本优势。
**Jensen Huang:** And so because it's a programmable system, the ability to invent new algorithms is really what makes AI advance so quickly. You know, TPUs like anything else are impacted by Moore's law. And we know that Moore's law is increasing about 25% per year. And so the only way to really get 10x leaps, 100x leaps, is to fundamentally change the algorithm and how it's computed every single year. And that's Nvidia's fundamental advantage.
**Jensen Huang:** 我们之所以能让 Blackwell 相比 Hopper 实现 50 倍的提升——你知道,我当时说的是 35 倍。我最初宣布的时候说 Blackwell 的能效将是 Hopper 的 35 倍。没有人相信。然后 Dylan 写了一篇文章,说实际上我保守了——实际上是 50 倍。光靠摩尔定律是做不到这个程度的。
**Jensen Huang:** The only reason why we were able to make Blackwell to Hopper 50 times -- you know, I said it was 35 times. And when I first announced it, Blackwell was going to be 35 times more energy efficient than Hopper. Nobody believed it. And then Dylan wrote an article. He said in fact I sandbagged it -- it's actually 50 times. And you can't reasonably do that with just Moore's law.
**Jensen Huang:** 所以我们解决这个问题的方式是,把新模型并行化(parallelized)、解耦(disaggregated)、分布式(distributed)部署到一个计算系统上。如果不能用 CUDA 深入到底层来开发新的核心程序(kernels),这真的很难做到。所以,我们架构的可编程性,加上英伟达是一家极致的协同设计(co-design)公司——我们甚至可以把一部分计算卸载到互联结构(fabric)本身中——比如 NVLink,或者网络 Spectrum-X——而且我们可以同时在处理器、系统、互联、库、算法各层面推动变革。所有这些都是同时完成的。没有 CUDA 来做这些,我都不知道从哪里开始。
**Jensen Huang:** And so the way that we solve that problem is new models parallelized and disaggregated and distributed across a computing system. And without the ability to really get down and come up with new kernels with CUDA, it's really hard to do. And so the combination of the programmability of our architecture, the fact that Nvidia is an extreme co-design company where we could even offload some of the computation into the fabric itself -- NVLink for example, into the network, Spectrum-X -- and that we could affect change across the processors, the system, the fabric, the libraries, the algorithm. All of that was done simultaneously. Without CUDA to do that, I wouldn't even know where to start.
**Dwarkesh Patel:** 这就引出了一个关于英伟达客户群体的有趣问题。如果 60% 的收入来自五大超大规模客户(hyperscalers),你知道,在另一个时代,如果客户是教授们在做实验,他们确实需要 CUDA,用不了其他加速器,他们需要直接跑 PyTorch 配合 CUDA,让一切都优化好。
**Dwarkesh Patel:** So, this gets at an interesting question about Nvidia's clientele. Where if 60% of your revenue is coming from these big five hyperscalers, you know, in a different era where different customers -- let's say it's professors who are running experiments and they are helped a bunch by -- they need CUDA. They can't use another accelerator. They need to just run PyTorch with CUDA and have everything optimized.
**Dwarkesh Patel:** 但如果是超大规模客户,他们有资源自己写核心程序。事实上他们必须这么做,才能为他们特定的架构榨出最后那 5% 的性能。Anthropic、谷歌基本上都在用自己的加速器或 TPU。还有 Trainium。但即使是使用 GPU 的 OpenAI 也开发了 Triton,他们说我们需要自己的核心程序。所以他们没有使用 cuBLAS 和 NCCL 那些东西,而是建了自己的软件栈,而这个栈也可以编译到其他加速器上。那么,如果你的大多数客户都能够也确实在做 CUDA 的替代品,CUDA 到底在多大程度上能确保前沿 AI 在英伟达上发生?
**Dwarkesh Patel:** But if you've got these hyperscalers, they have the resources to write their own kernels. In fact, they have to, to get that extra last 5% that they need for their specific architecture. Anthropic, Google are mostly running their own accelerators or running TPUs. And Tranium. But even OpenAI using GPUs has Triton, which they're like, we need our own kernels. So they've -- down to CUDA C++ they've instead of using cuBLAS and NCCL and everything, they've got their own stack which compiles to other accelerators as well. And so if most of your customers can and do make replacements for CUDA, to what extent is CUDA really the thing that is going to make Frontier AI happen on Nvidia?
**Jensen Huang:** CUDA 是一个丰富的生态系统。所以如果你想首先在任何计算机上构建,先在 CUDA 上构建是极其明智的选择,因为这个生态系统太丰富了。我们支持每一个框架。如果你想创建自定义核心程序——比如说,我们对 Triton 做出了巨大贡献。Triton 的后端有大量英伟达的技术。我们很乐意帮助每一个框架变得尽可能强大。
**Jensen Huang:** CUDA is a rich ecosystem. And so if you want to build on any computer first, building on CUDA first is incredibly smart, because the ecosystem is so rich. We support every framework. If you want to create custom kernels -- for example, we contribute enormously to Triton. And so the back end of Triton -- huge amounts of NVIDIA technology. We're delighted to help every framework become as great as it can be.
**Jensen Huang:** 而且框架非常非常多。有 Triton、vLLM、SGLang,然后还有更多。现在又有一大批新的强化学习(reinforcement learning)框架涌现出来。你有 Verl、Nemo RL,还有一大堆新的——然后随着后训练(post-training)和强化学习的发展,这整个领域正在爆发式增长。
**Jensen Huang:** And there's lots and lots of frameworks. There's Triton, there's vLLM, there's SGLang, and then there's more. And now there's a whole bunch of new reinforcement learning frameworks coming out. You've got Verl, you've got Nemo RL, you've got a whole bunch of new -- and then with post-training and reinforcement learning, that entire area is just exploding.
**Jensen Huang:** 所以如果你想在一个架构上构建,在 CUDA 上构建是最合理的,因为你知道生态系统很好。你知道如果出了问题,更可能是你自己的代码的问题,而不是底层那一大堆代码的问题。别忘了你在构建这些系统时要处理的代码量。当什么东西不工作时,到底是你的问题还是计算机的问题?你希望永远是你的问题,并且能够信任计算机。显然我们自己也有很多 bug,但我们的系统经过了如此充分的打磨,你至少可以在这个基础上构建。
**Jensen Huang:** And so if you want to build on an architecture, building on CUDA makes the most sense because you know that the ecosystem is great. You know that if something happens, it's more likely in your code and not in the mountain of code underneath. Don't forget the amount of code that you're dealing with when you're building these systems. When something doesn't work, was it you or was it the computer? You would like it always to be you and to be able to trust the computer. And obviously we still have lots and lots of bugs ourselves, but our system is so well wrung out that you could at least build on top of the foundation.
**Jensen Huang:** 这是第一点——生态系统的丰富性、可编程性和能力。第二点是,如果你是一个开发者,在构建任何东西的时候,你最想要的就是安装基数(install base)。你希望你构建的软件能在一大堆其他计算机上运行。你不是只为自己构建软件。你是在为你的机群或所有人的机群构建软件,因为你是一个框架构建者。
**Jensen Huang:** So that's number one -- the richness of the ecosystem, the programmability of it, the capability of it. The second thing is, if you were a developer and you were building anything at all, the single most important thing you want more than anything is install base. You want the software that you build to run on a whole bunch of other computers. You're not building software just for yourself. You're building software for your fleet or for everybody else's fleet because you're a framework builder.
**Jensen Huang:** 英伟达的 CUDA 生态系统最终是其最大的宝藏。我们现在——我不知道有多少——好几亿个 GPU。每个云都有。从 A10、A100、H100、H200,到 L 系列、P 系列。各种各样的尺寸和形态。如果你是一家机器人公司,你希望那个 CUDA 栈能够在机器人本身里运行。我们真的无处不在。所以安装基数意味着,一旦你开发了软件、一旦你开发了模型,它在哪里都有用。安装基数的价值是难以估量的。
**Jensen Huang:** And Nvidia's CUDA ecosystem is ultimately its great treasure. We are now -- I don't know how many -- several hundred million GPUs. Every cloud has it. It goes back to A10, A100, H100, H200, the L series, the P series. I mean, there's a whole bunch of them and they're in all kinds of sizes and shapes. And if you're a robotics company, you want that CUDA stack to actually run in the robot itself. We're literally everywhere. And so the install base says that once you develop the software, once you develop the model, it's going to be useful everywhere. And so the install base is just incredibly valuable.
**Jensen Huang:** 最后,我们存在于每一个云平台这个事实,让我们真正独一无二。因为你是一家 AI 公司、一个 AI 开发者,你不确定要和哪个云服务商(CSP)合作、在哪里运行。我们在任何地方都能运行,如果你愿意的话也可以在本地(on prem)部署。所以我认为生态系统的丰富性、安装基数的广泛性、以及我们部署位置的多样性——这个组合让 CUDA 无可替代。
**Jensen Huang:** And then lastly, the fact that we're in every single cloud makes us genuinely unique. Because you're an AI company and you're an AI developer. You're not exactly sure which CSP you're going to partner with and where you would like to run it. And we'd run it everywhere, including on prem for you if you like. And so I think the richness of the ecosystem, the expansiveness of the install base, and the versatility of where we are -- that combination makes CUDA invaluable.
**Dwarkesh Patel:** 这很有道理。我好奇的是,这些优势对你的主要客户来说是否真的很重要。有很多人——这些优势可能对某些人重要——但对于那些有能力自建整个软件栈的客户,也就是贡献了你大部分收入的那些客户,特别是如果进入一个 AI 在具有紧密验证回路(tight verification loops)的任务上变得特别擅长的世界,你可以在上面做强化学习。那么,如何在一个规模化集群上最高效地编写做注意力或多层感知器(MLP)的核心程序——这是一个非常可验证的反馈回路。那么是不是所有超大规模客户都可以自己编写这些自定义核心程序?
**Dwarkesh Patel:** That makes a lot of sense. I guess the thing I'm curious about is whether those advantages matter a lot to your main customers. Like, there's many people who -- they might matter for -- but for the kind of person who can actually build their own software stack, who make up most of your revenue, especially if you go to a world where AI is getting especially good at the things which have tight verification loops where you can RL on them. And then this question of how do you write a kernel that does attention or MLP the most efficiently across a scale-up -- it's a very verifiable sort of feedback loop. And so can all the hyperscalers write these custom kernels for themselves?
**Dwarkesh Patel:** 他们可能仍然——英伟达的性价比仍然很好,所以他们可能仍然倾向于使用英伟达。但问题就变成了,这是否仅仅变成一个谁能在给定价格下提供最好的规格、最多的浮点运算和内存以及内存带宽的问题?而历史上英伟达一直拥有——现在依然拥有——AI 领域软硬件中最高的利润率,超过 70%,正是因为有这个 CUDA 护城河。问题是,如果你的大多数客户实际上都有能力自建而不依赖 CUDA 护城河,你能维持这样的利润率吗?
**Dwarkesh Patel:** And they might still -- Nvidia still has great price performance. So they might still prefer to use Nvidia. But then the question is, does it just become a question of who is offering the best specs, the best flops and memory and memory bandwidth for a given dollar, where historically Nvidia has just had -- and still has, you know -- the best margins in all of AI across hardware and software, 70% plus, because of this CUDA moat. And the question is, can you sustain those margins if for most of your customers they can actually afford to build instead of relying on the CUDA moat?
**Jensen Huang:** 我们派驻在这些 AI 实验室的工程师数量是惊人的。和他们一起工作、优化他们的栈。原因是没有人比我们自己更了解我们的架构。这些架构不像 CPU 那样通用。你知道,CPU 就像凯迪拉克。它总是——就是一辆不错的巡航车。永远不会开太快。谁都能开得不错。有巡航控制。什么都很轻松。
**Jensen Huang:** The number of engineers we have assigned to these AI labs is insane. Working with them, optimizing their stack. And the reason for that is because nobody knows our architecture better than we do. And these architectures are not as general purpose as a CPU. The reason why a CPU is so -- you know, a CPU is kind of like a Cadillac. It just always -- it's a nice cruiser. It never goes too fast. Everybody drives it pretty well. It's got cruise control. Everything is easy.
**Jensen Huang:** 但在很多方面,英伟达的 GPU,我们的加速器,有点像 F1 赛车。是的,我可以想象每个人都能开到 100 英里每小时,但要把它推到极限需要相当多的专业知识。我们大量使用 AI 来创建我们拥有的那些核心程序。我相当确信在相当长的时间内他们仍然需要我们。所以我们的专业知识能帮助我们的 AI 实验室合作伙伴轻松地从他们的栈中再榨出 2 倍的性能。我们优化完他们的栈或某个特定的核心程序后,模型加速 3 倍、2 倍、50% 是常有的事。
**Jensen Huang:** But in a lot of ways, Nvidia's GPUs, our accelerators, are kind of like F1 racers. And yeah, I could imagine everybody's able to drive it at 100 miles an hour, but it takes quite a bit of expertise to be able to push it to the limit. And we use a ton of AI to create the kernels that we have. And I'm pretty sure we're going to still be needed for quite some time. And so our expertise helps our AI labs partners get another 2x out of their stack easily. Oftentimes it's not unusual that by the time that we're done optimizing their stack or optimizing a particular kernel, their model sped up by 3x, 2x, 50%.
**Jensen Huang:** 这是一个巨大的数字,特别是当你谈论的是他们拥有的整个机群的安装基数——所有那些 Hopper 和 Blackwell。当你把性能提升两倍,那就直接翻倍了收入。这直接转化为收入。
**Jensen Huang:** That's a huge number, especially when you're talking about the installed base of the fleet that they have, of all the Hoppers and Blackwells that they have. When you increase it by a factor of two, that doubles the revenues. That directly translates to revenues.
**Jensen Huang:** 英伟达的计算栈是全球最佳的性能总拥有成本比(performance per TCO),没有之一。没有人能向我证明世界上有任何一个平台拥有更好的性能 TCO 比。没有一家公司。事实上,基准测试摆在那里。Dylan 的 Inference Max 摆在那里让所有人使用,没有一个 TPU 敢来、Trainium 也不敢来。我鼓励他们使用 Inference Max 来展示他们令人难以置信的推理成本。这真的真的很难。没有人愿意出场。
**Jensen Huang:** Nvidia's computing stack is the best performance per TCO in the world, bar none. Nobody can demonstrate to me that any single platform in the world today has better performance-TCO ratio. Not one company. And in fact, the benchmarks are out there. Dylan's Inference Max is sitting out there for everybody to use and not one TPU won't come, Trainium won't come. I encourage them to use Inference Max and demonstrate their incredible inference cost. It's really really hard. Nobody wants to show up.
**Jensen Huang:** ML Perf——我非常欢迎 Trainium 来证明他们一直声称的 40% 的优势。我很想听他们展示 TPU 的成本优势。在我看来这完全说不通。从第一性原理来看完全说不通。毫无道理。所以我认为我们如此成功的原因很简单,就是因为我们的 TCO 太好了。
**Jensen Huang:** ML Perf -- I would welcome Trainium to demonstrate their 40% that they claim all the time. I would love to hear them demonstrate the cost advantage of TPUs. It makes no sense in my mind. It makes absolutely zero sense on first principles. It makes no sense. And so I think the reason why we're so successful is simply because our TCO is so great.
**Jensen Huang:** 还有第二件事。你说 60% 的客户是前五大客户。但其中大部分业务是对外的。比如说,英伟达在 AWS 上的大部分业务是服务外部客户,而不是内部使用。我们在 Azure 上的大部分客户——显然都是外部的。我们在 OCI 上的所有客户都是外部的,不是内部使用。他们偏爱我们的原因是我们的覆盖面如此之广。我们能给他们带来全世界所有优秀的客户。这些客户都构建在英伟达上。而所有这些公司构建在英伟达上的原因,是因为我们的覆盖范围和多样性如此之大。
**Jensen Huang:** There's a second thing. You say 60% of our customers are the top five. But most of that business is external. For example, most of Nvidia in AWS is for external customers, not internal use. Most of our customers at Azure -- obviously all of our customers are external. All of our customers at OCI are external, not internal use. The reason why they favor us is because our reach is so great. We can bring them all of the great customers in the world. They're all built on Nvidia. And the reason why all these companies are built on Nvidia is because our reach and our versatility is so great.
**Jensen Huang:** 所以我认为飞轮效应(flywheel)的核心是安装基数、我们架构的可编程性、我们生态系统的丰富性,以及世界上有如此多的 AI 公司——现在已经有数万家了。如果你是其中一家 AI 初创公司,你会选择什么架构?你会选择世界上最普遍的、拥有最大安装基数的、拥有丰富生态系统的架构。
**Jensen Huang:** And so I think the flywheel is really install base, the programmability of our architecture, the richness of our ecosystem, and the fact that there's so many AI companies in the world -- there's tens of thousands of them now. And if you were one of those AI startups, what architecture would you choose? You would choose an architecture that's most abundant in the world, the one that has the largest installed base, and one that has a rich ecosystem.
**Jensen Huang:** 这就是飞轮。这就是原因——在以下几者的组合之间:第一,我们的性价比非常好,他们有最低成本的代币。第二,我们的性能功耗比(perf per watt)全球最高。所以如果我们的合作伙伴建了一个 1 吉瓦(gigawatt)的数据中心,那个 1 吉瓦的数据中心最好能提供最大量的收入和代币数量——这直接转化为收入。你想生成尽可能多的代币,最大化那个数据中心的收入。我们拥有全球最高的每瓦代币产出架构。最后,如果你的目标是出租基础设施,我们拥有全球最多的客户。这就是飞轮运转的原因。
**Jensen Huang:** And so that's the flywheel. That's the reason why -- between the combination of: one, our perf per dollar is so great, they have the lowest cost tokens. Second, our perf per watt is the highest in the world. And so if one of these companies, if our partners built a 1 gigawatt data center, that 1 gigawatt data center better deliver the maximum amount of revenues and number of tokens -- which directly translates to revenues. You want to generate as many tokens as possible, maximize the revenues for that data center. We have the highest tokens per watt architecture in the world. And then lastly, if your goal is to rent the infrastructure, we have the most customers in the world. And so that's the reason why the flywheel works.
**Dwarkesh Patel:** 有意思。我觉得问题归结于实际的市场结构是什么样的。因为即使有其他公司,也可能存在一种情况,成千上万家 AI 公司大致平分算力份额。但即使通过这五大超大规模客户,真正在亚马逊上使用算力的人——Anthropic、OpenAI——以及这些有能力也有实力让不同加速器工作的大型基础模型实验室——
**Dwarkesh Patel:** Interesting. I guess the question comes down to what is the actual market structure here. Because even if there's other companies, there could have been a world where there's tens of thousands of AI companies that have roughly equal share of compute. But even through these five hyperscalers, really the people on Amazon using the compute -- Anthropic, OpenAI -- and these big foundation labs who can themselves afford and have the ability to make different accelerators work --
**Jensen Huang:** 不,我觉得你的假设——你的前提就是错的。
**Jensen Huang:** No, I think your assumption -- your premise is wrong.
**Dwarkesh Patel:** 也许吧。让我换个角度问一个稍微不同的问题——
**Dwarkesh Patel:** Maybe. Let me ask you a slightly different question which is --
**Jensen Huang:** 回来让我纠正你的前提。
**Jensen Huang:** Come back and make me correct your premise.
**Dwarkesh Patel:** 好吧,让我换一个问题来问——
**Dwarkesh Patel:** Okay, let me just ask a different question which is, okay if everything --
**Jensen Huang:** 但还是要确保——让我回头来修正一下——因为这对AI来说太重要了,对科学的未来太重要了,对整个行业的未来太重要了,那个前提——
**Jensen Huang:** But still make sure that -- make me come back and fix -- because it's just too important to AI, it's too important to the future of science, it's too important to the future of the industry, that premise --
**Dwarkesh Patel:** 那个前提——听我说,让我先把问题问完,然后我们一起来探讨。好吧。
**Dwarkesh Patel:** The premise -- look let me just finish the question and then we can address it together. Yeah.
**Dwarkesh Patel:** 那么你怎么看——如果你说的这些关于性价比(price performance)和每瓦性能(performance per watt)等等都是真的——你觉得为什么会出现这样的情况:比如Anthropic,就在几天前刚刚宣布与Broadcom和Google达成了一笔数吉瓦级别的TPU协议?而Google的大部分算力显然是以TPU为主。所以如果我看这些大型AI公司,似乎它们有很大一部分——曾经有一段时间全都是用英伟达的,现在不是了。所以我很好奇,如果这些东西在纸面上都是成立的,他们为什么要选择其他加速器?
**Dwarkesh Patel:** So what do you think -- if all these things are true about price performance and performance per watt etc. -- why do you think it is the case that, say, Anthropic for example just announced a couple days ago they have a multi-gigawatt deal with Broadcom and Google for TPUs? And majority of their compute, obviously for Google, it's TPU majority compute. So if I look at these big AI companies, it seems like a lot of their -- there was some point where it was all Nvidia and now it's not. And so I'm curious how to square -- if these things are true on paper, why are they going with other accelerators?
**Jensen Huang:** 是的,Anthropic是一个特殊案例,而不是一种趋势。如果没有Anthropic,TPU的增长从何而来?百分之百是Anthropic。如果没有Anthropic,Trainium的增长从何而来?也是百分之百靠Anthropic。我认为这一点业内都知道,也都理解。并不是说ASIC的机会满地都是,Anthropic就只有一个。
**Jensen Huang:** Yeah, Anthropic is a unique instance and not a trend. Without Anthropic, why would there be any TPU growth at all? It's 100% Anthropic. Without Anthropic, why would there be any Trainium growth at all? It's 100% Anthropic. And I think that's fairly well known and well understood. It's not that there's an abundance of ASIC opportunities. There's only one Anthropic.
**Dwarkesh Patel:** 但OpenAI也在跟AMD合作啊。他们还在自研Titan加速器。
**Dwarkesh Patel:** But OpenAI deals with AMD. They're building their own Titan accelerator.
**Jensen Huang:** 是的。但他们的主力——我们大家都承认,他们绝大部分用的还是英伟达,而且我们未来还会一起做很多事情。
**Jensen Huang:** Yeah. But they're mostly -- we could all acknowledge they're vastly Nvidia and we're going to still do a lot of work together.
**Jensen Huang:** 是的。而且我不会因为别人用了其他东西、去尝试其他方案就觉得被冒犯。如果他们不去试试那些东西,他们怎么知道我们的有多好呢?有时候你确实需要被提醒一下。而我们必须持续去赢得我们现在所处的位置。你知道,总有人会提出各种说法——你看看有多少ASIC项目被砍掉了。就算你要造一颗ASIC,你仍然得做出比英伟达更好的东西,而做出比英伟达更好的东西并不容易。实际上这并不明智,你知道的。英伟达必须得犯了什么错才行。说真的。因为我们的规模、我们的迭代速度——我们是全世界唯一一家每年都在推出产品的公司。每年都有巨大飞跃。
**Jensen Huang:** Yeah. And we're not -- I'm not offended by other people using something else and trying things. If they don't try these other things, how would they know how good ours is? And sometimes you got to be reminded of it. And we have to continuously earn the position that we're in. You know, there are always claims -- and look at the number of ASICs that have been cancelled. Just because you're going to build an ASIC, you still have to build something better than Nvidia. And it's not that easy building something better than Nvidia. It's not sensible actually, you know. Nvidia's got to be missing something. Seriously. And because our scale, our velocity -- we're the only company in the world that's cranking it out every single year. Big leaps every single year.
**Dwarkesh Patel:** 我猜他们的逻辑是这样的:嘿,它不需要比你好,只要不比你差超过70%就行,因为你们赚着70%的毛利率(margin)。
**Dwarkesh Patel:** I guess their logic is that, hey, it doesn't need to be better. It just needs to be not more than 70% worse because they're paying you 70% margins.
**Jensen Huang:** 不不不。别忘了——即使是ASIC的毛利率也相当高。假设英伟达的毛利率是70%,但ASIC的毛利率是65%。你到底省了多少?
**Jensen Huang:** No, no, no. Don't forget -- even an ASIC margin is really quite high. Nvidia's margin is 70% let's say, but an ASIC margin is 65. What are you really saving?
**Dwarkesh Patel:** 哦,你是说从Broadcom那边?
**Dwarkesh Patel:** Oh, you mean from Broadcom or something?
**Jensen Huang:** 对,没错。
**Jensen Huang:** Yeah, sure.
**Dwarkesh Patel:** 你总得付钱给谁。
**Dwarkesh Patel:** You got to pay somebody.
**Jensen Huang:** 对。所以我认为ASIC的毛利率据我所知是非常高的,他们自己也这么觉得。他们还对自己惊人的ASIC毛利率相当自豪呢。
**Jensen Huang:** Yeah. And so I think the ASIC margins are incredibly good from what I can tell, and they believe it too. And so they're quite proud of their incredible ASIC margins.
**Jensen Huang:** 所以你问为什么会这样。很久以前我们确实没有能力去做这件事。当时我没有深刻意识到,建立一个像OpenAI和Anthropic这样的基础模型实验室(foundation AI lab)有多难。以及他们需要供应商自身进行巨额投资这个事实。我们当时根本没有能力拿出数十亿美元投资到Anthropic,让他们来用我们的算力。但Google和AWS有这个能力,他们在早期投入了巨额资金,让Anthropic作为回报使用他们的算力。我们当时就是没有条件这么做。
**Jensen Huang:** And so you ask the question why. A long time ago we just didn't have the ability to do it. And at the time I didn't deeply internalize how difficult it would be to build a foundation AI lab like OpenAI and Anthropic. And the fact that they needed huge investments from the supplier themselves. We just weren't in a position to make the multi-billion dollar investment into Anthropic so that they could use our compute. But Google and AWS were, and they put in huge investments in the beginning so that Anthropic in return use their compute. We just weren't in a position to do so at the time.
**Jensen Huang:** 而且我也没有——我会说我的失误在于,我没有深刻认识到他们真的别无选择。风投(VC)是不可能投入50亿、100亿到一家AI实验室,然后指望它变成Anthropic的。所以这是我的判断失误。但即使当时我看明白了,我觉得我们也没有条件那样做。不过同样的错误我不会再犯了。
**Jensen Huang:** Nor did I -- I would say my mistake is I didn't deeply internalize that they really had no other options. That a VC would never put in 5, 10 billion of investment into an AI lab with the hopes of it turning out to be Anthropic. And so that was my miss. But even if I understood it, I don't think we would have been in a position to do that at the time. But I'm not going to make that same mistake again.
**Jensen Huang:** 我很高兴能投资OpenAI,也很高兴能帮助他们扩大规模,我认为这是必须做的事情。然后当Anthropic来找我们的时候,我也很高兴成为投资者,很高兴帮助他们扩展。只是我们在当时没有能力这么做。如果一切可以重来,如果英伟达当时就有现在这么大的体量,我会非常乐意去做。
**Jensen Huang:** And I'm delighted to invest in OpenAI and I'm delighted to help them scale and I believe it's essential to do so. And then when Anthropic came to us, I'm delighted to be an investor, delighted to help them scale. But we just weren't at the time able to do so. If I could rewind everything, if Nvidia could have been as big back then as we are now, I would have been more than happy to do it.
**Dwarkesh Patel:** 这其实挺有意思的——多年来英伟达一直是AI领域里赚钱的公司,赚了很多钱。而现在你在把钱投出去。据报道你已经向OpenAI投了多达300亿美元,向Anthropic投了100亿美元。但现在他们的估值已经涨上去了,我相信还会继续涨。
**Dwarkesh Patel:** This is actually quite interesting, which is -- for many years Nvidia has been the company in AI making money, making lots of money. And now you're investing it. It's been reported that you've done up to 30 billion in OpenAI and 10 billion in Anthropic. But now their valuations have increased and I'm sure they'll continue to increase.
**Dwarkesh Patel:** 所以如果在过去这么多年里,你一直在给他们提供算力,你看到了AI的发展方向,而他们几年前甚至一年前的估值只有现在的十分之一。而你手里有这么多现金。有一种可能性是——英伟达自己成为一个基础模型实验室,做出巨额投资来实现这一点;或者在更早的时候以当前估值做了你现在做的这些交易。你当时有这个现金来做。所以我确实很好奇,为什么不更早做呢?
**Dwarkesh Patel:** And so if over all these many years, you were giving them the compute, you saw where AI was headed, and then they were worth like one-tenth what they are now a couple years ago or even a year ago in some cases. And you had all this cash. There's a world where either Nvidia themselves becomes a foundation lab, does the huge investment to make that possible, or has made the deals you've made now at current valuations much earlier on. And you had the cash to do it. So I am curious actually, why not have done it earlier?
**Jensen Huang:** 我们在能做的第一时间就做了。我们在能做到的时候就做了,如果能更早做到,我会更早做。在Anthropic需要我们做这件事的时候,我们就是没有条件做。当时这不在我们的认知范围内。
**Jensen Huang:** We did it as soon as we could. We did it as soon as we could have, and if I could have, I would have done it even earlier. At the time that Anthropic needed us to do it, we just weren't in a position to do it. It wasn't in our sensibility to do so.
**Dwarkesh Patel:** 这是——资金方面的原因还是——
**Dwarkesh Patel:** How's that -- like a cash thing or just --
**Jensen Huang:** 对,是投资规模的问题。你知道,当时我们从来没有在公司外部做过投资,而且投入不了那么多。当时我们也没意识到有这个必要。我一直觉得他们可以去找风投融资啊,天哪,就像所有公司做的那样。但他们想做的事情是风投搞不定的。OpenAI想做的事情是风投搞不定的。我现在认识到了这一点。当时不知道。但这恰恰说明了他们的远见。这就是为什么他们聪明。
**Jensen Huang:** Yeah, the level of investment. You know, we never invested outside the company at the time, and not that much. And we didn't realize we needed to. I always thought that they could just go raise VCs for God's sakes, like all companies do. But what they were trying to do couldn't have been done through VCs. What OpenAI wanted to do couldn't have been done through VCs. And I recognize that now. I didn't know it then. But that's their genius. That's why they're smart.
**Jensen Huang:** 所以他们当时就意识到必须走那条路。我很高兴他们做到了。即使我们的原因导致Anthropic不得不去找别人,我仍然为此感到高兴。Anthropic的存在对世界是好事。我对此感到欣慰。
**Jensen Huang:** And so they realized it then that they had to do something like that. And I'm delighted that they did. And even though we caused Anthropic to have to go to somebody else, I'm still happy that it happened. Anthropic's existence is great for the world. I'm delighted for it.
**Dwarkesh Patel:** 我想你确实还在赚很多钱,而且每个季度都赚得越来越多。
**Dwarkesh Patel:** I guess you still are making a ton of money and you're making way more money quarter after quarter.
**Jensen Huang:** 有些遗憾还是可以有的。
**Jensen Huang:** It's still okay to have regrets.
**Dwarkesh Patel:** 那问题就来了——好吧,既然我们现在到了这一步,你手上有这么多不断增长的资金,英伟达应该拿这些钱做什么?有一种观点认为,你看,整个中间商生态系统已经冒出来了,它们把资本开支(capex)转化为运营开支(opex)给这些实验室,这样他们就可以租用算力,因为芯片确实很贵。这些芯片在其生命周期内能赚很多钱,因为模型在变好,每个token产生的价值在增加,但前期建设成本很高。而英伟达有钱来承担这些资本开支。
**Dwarkesh Patel:** So then the question still arises -- okay, well now that we're here and you have all this money that you keep making, what should Nvidia be doing with it? And there's one answer which says, look, there's this whole middleman ecosystem that has popped up for converting capex into opex for these labs so that they can rent compute, because the chips are really expensive. They make a lot of money over their lifetime because the models are getting better, the value that they generate from their tokens is increasing, but they're expensive to set up. Nvidia has the money to do the capex.
**Dwarkesh Patel:** 而且据报道你在为CoreWeave提供担保——高达63亿美元,并且已经投资了20亿。但是话说回来,为什么英伟达不自己成为一朵云(cloud)呢?为什么不自己成为超大规模云厂商(hyperscaler)来运营这些算力呢?你有这么多现金来做这件事。
**Dwarkesh Patel:** And in fact, it's been reported you're backstopping CoreWeave -- up to 6.3 billion and have invested 2 billion. But yeah, why doesn't Nvidia become a cloud themselves? Why not become a hyperscaler themselves and run this compute out? You have all this cash to do it.
**Jensen Huang:** 这是公司的哲学,我认为是明智的。我们应该做尽可能多的必要之事,但做尽可能少的事。这意味着我们在构建计算平台方面所做的工作——如果我们不做,我真心相信没有人会去做。如果我们不承担那些风险,如果我们不像我们那样构建NVLink,如果我们不构建整个软件栈(stack),如果我们不以我们的方式创建生态系统,如果我们不用20年时间在CUDA上投入——那段时间大部分是亏钱的——如果我们不做,没有人会去做。
**Jensen Huang:** This is a philosophy of the company and I think it is wise. We should do as much as needed, as little as possible. And what that means is the work that we do with building our computing platform -- if we don't do it, I genuinely believe it doesn't get done. If we didn't take the risk that we take, if we didn't build NVLink the way we built, if we didn't build the whole stack, if we didn't create the ecosystem the way we did it, if we didn't dedicate ourselves to 20 years of CUDA while losing money most of that time -- if we didn't do it, nobody else would have done it.
**Jensen Huang:** 如果我们不创建所有那些CUDA X库(library),让它们都是面向特定领域的——你知道,大约十五年前,我们就开始推进特定领域的库,因为我们意识到如果我们不创建这些特定领域的库,不管是用于光线追踪(ray tracing)、图像生成,还是早期的AI工作,这些模型,不管是用于数据处理、结构化数据处理还是向量数据处理——如果我们不做,没有人会做。我对此非常确定。我们为计算光刻(computational lithography)创建了一个叫cuLitho的库。如果我们不做,没人会做。所以如果没有我们所做的一切,加速计算(accelerated computing)不会像今天这样发展。所以我们应该做这些事情。我们应该把公司所有的力量,全心全意地投入去做这件事。
**Jensen Huang:** If we didn't create all the CUDA X libraries so that they're all domain specific -- you know, a decade and a half ago, we pushed into domain specific libraries because we realized that if we didn't create these domain specific libraries, whether it's for ray tracing or image generation or even the early works of AI, these models, if we didn't create them for data processing, structured data processing or vector data processing -- if we didn't create them, nobody would. And I am completely certain of that. We created a library for computational lithography called cuLitho. If we didn't create it, nobody would have. And so accelerated computing wouldn't advance the way it has if we didn't do what we did. And so we should do that. We should dedicate our company, all of our might, wholeheartedly to go do that.
**Jensen Huang:** 然而,世界上有很多云服务。如果我不做,自然会有人来做。所以按照这个原则——做尽可能多的必要之事但尽可能少——这个哲学今天存在于我们公司,我做的每件事都是用这个视角来看的。在云服务这件事上,如果我们不支持CoreWeave的存在,这些新兴云(neo-cloud)、这些AI云就不会存在。如果我们不帮助CoreWeave存在,他们就不会存在。如果我们不支持Nscale,他们不会走到今天。如果我们不支持NBS,他们也不会走到今天。现在,他们做得非常出色。
**Jensen Huang:** However, the world has lots of clouds. If I didn't do it, somebody would show up. And so following the recipe, the philosophy of doing as much as needed but as little as possible -- that philosophy exists in our company today and everything I do, I do it with that lens. In the case of clouds, if we didn't support CoreWeave to exist, these neo-clouds, these AI clouds wouldn't exist. If we didn't help CoreWeave exist, they would not exist. If we didn't support Nscale, they wouldn't be where they are today. If we didn't support NBS, they wouldn't be where they are today. Now, they are -- they're doing fantastically.
**Jensen Huang:** 这是一种商业模式吗?不是。我们应该做尽可能多的必要之事,但做尽可能少的事。所以我们尝试——我们投资自己的生态系统,因为我希望我们的生态系统蓬勃发展。我希望这个架构,我希望AI能够触达尽可能多的行业、尽可能多的国家,让整个星球都建立在AI之上,建立在美国的技术栈(tech stack)之上。所以我认为这个愿景正是我们在追求的。
**Jensen Huang:** Is that a business model? No. We should do as much as needed, as little as possible. And so we're trying -- we invest in our ecosystem because I want our ecosystem to thrive. And I want the architecture and I want AI to be able to connect with as many industries as possible, as many countries as possible, and make it possible for the planet to be built on AI and to be built on the American tech stack. And so that vision I think is exactly what we're pursuing.
**Jensen Huang:** 你提到的另一件事——有这么多了不起的、令人惊叹的基础模型公司,我们试图投资所有的。这也是我们做的另一件事。我们不挑赢家。我们需要支持每一家,这也是我们乐于做的事情。这是我们业务的必然要求,但我们也会特意不去挑赢家。所以当我投资其中一家的时候,我会投资所有的。
**Jensen Huang:** Now, one of the things that you mentioned -- there are so many great, amazing foundation model companies and we try to invest in all of them. And this is another thing that we do. We don't pick winners. We need to support everyone and it's part of our joy of doing so. It's an imperative to our business, but we also go out of our way not to pick winners. And so when I invest in one of them, I invest in all of them.
**Dwarkesh Patel:** 为什么你要特意不去挑赢家?
**Dwarkesh Patel:** Why do you go out of your way not to pick winners?
**Jensen Huang:** 因为这不是我们该做的事。第一。第二,当英伟达刚创立的时候,有60家图形公司,60家3D图形公司。我们是唯一一家存活下来的。如果你把那60家图形公司拿出来,问自己哪一家会活到最后,英伟达恰恰会排在"最不可能活下来"的名单最前面。
**Jensen Huang:** Because it's not our job to. Number one. Number two, when Nvidia first started, there were 60 graphics companies, 60 3D graphics companies. We are the only one that survived. If you would have taken those 60 companies, 60 graphics companies, and asked yourself which one was going to make it, Nvidia would be at the top of that list not to make it.
**Jensen Huang:** 你知道,这在你之前很久了,但英伟达的图形架构当时是完全错误的。不是有一点点偏差,我们创造了一个完全错误的架构。对开发者来说根本无法支持。它永远不会成功。我们从很好的第一性原理(first principles)出发去推理,但最终走到了错误的解决方案上。所有人都会把我们排除在外。然而我们走到了今天。所以我有足够的谦逊去认识到——不要挑赢家。
**Jensen Huang:** You know, this is long before you, but Nvidia's graphics architecture was precisely wrong. It's not a little bit wrong. We created an architecture that was precisely wrong. And it was an impossible thing for developers to support. It was never going to make it. We reasoned about it from good first principles, but we ended up in the wrong solution. And everybody would have counted us out. And here we are. And so I have enough humility to recognize that -- don't pick winners.
**Dwarkesh Patel:** 对。
**Dwarkesh Patel:** Yeah.
**Jensen Huang:** 要么让他们各自照顾自己,要么就照顾所有人。
**Jensen Huang:** Either let them all take care of themselves or take care of all of them.
**Dwarkesh Patel:** 有一件事我没太理解,你说"我们不是因为有了新云就要优先扶持这些新兴云"。但你同时也列举了一堆新兴云,说如果没有英伟达它们就不会存在。
**Dwarkesh Patel:** One thing I didn't understand is you said, "Look, we're not prioritizing these neo-clouds just because there are new clouds and we want to prop them up." But you also said you listed a bunch of new clouds and you said they wouldn't exist if it wasn't for Nvidia.
**Jensen Huang:** 对。
**Jensen Huang:** Yeah.
**Dwarkesh Patel:** 那么这两件事怎么兼容呢?
**Dwarkesh Patel:** And so how are those two things compatible?
**Jensen Huang:** 首先,他们自己得想要存在,然后来找我们寻求帮助。当他们想要存在,有商业计划,有专业知识,有热情——他们自己显然也得有一些能力。但如果最终他们需要一些投资来启动,我们会支持他们。但他们越早转起自己的飞轮(flywheel)越好——你的问题是我们是否想做融资业务?答案是不想。
**Jensen Huang:** First of all, they need to want to exist, and they come to ask us for help. And when they want to exist and they have a business plan and they have expertise and they have the passion for it -- they obviously have to have some capabilities themselves. But if at the end of the day they need some investment in order to get it off the ground, we would be there for them. But the sooner they get their flywheel going -- you know, your question was do we want to be in the financing business? The answer is no.
**Jensen Huang:** 我们不想做——因为有人在做融资业务,我们宁愿跟所有做融资业务的人合作,而不是自己成为一个融资方。所以我认为我们的目标是专注于我们做的事情,让我们的商业模式尽可能简单,支持我们的生态系统。
**Jensen Huang:** We don't want to be -- because there are people in the financing business and we'd rather work with all of the people who are in the financing business than to be a financier ourselves. And so I think our goal is to focus on what we do, keep our business model as simple as possible, support our ecosystem.
**Jensen Huang:** 当像OpenAI这样的公司需要300亿美元规模的投资,因为他们还没有IPO,而我们深信他们——我深信他们将会成为——好吧,他们今天已经是一家非凡的公司了。他们将会成为一家了不起的公司。世界需要他们存在。世界希望他们存在。我希望他们存在。而且他们正顺风顺水。让我们支持他们,让他们去扩展。所以这类投资我们会做,因为他们需要我们去做。但我们并不是要做尽可能多的事,我们是在尝试做尽可能少的事。
**Jensen Huang:** When someone like OpenAI needs an investment of $30 billion scale because it's still before their IPO, and we deeply believe in them -- I deeply believe that they're going to be -- well, they're an extraordinary company already today. They're going to be an incredible company. The world needs them to exist. The world wants them to exist. I want them to exist. And they have the wind at their back. Let's support them and let them scale. And so those investments we'll do because they need us to do it. But we're not trying to do as much as possible. We're trying to do as little as possible.
**Dwarkesh Patel:** 这可能是个比较明显的问题,但我们已经在GPU短缺的环境中生活了好多年了,而且现在因为模型越来越好,短缺还在加剧。
**Dwarkesh Patel:** This may be sort of an obvious question, but we've lived many years in this situation where there's a shortage of GPUs and it's grown now because models are getting better.
**Jensen Huang:** 我们确实存在GPU短缺。
**Jensen Huang:** We have a shortage of GPUs.
**Dwarkesh Patel:** 对。
**Dwarkesh Patel:** Yes.
**Jensen Huang:** 对。
**Jensen Huang:** Yeah.
**Dwarkesh Patel:** 而且英伟达在分配稀缺产能时,据说不只是看谁出价最高,而是——嘿,我们要确保这些新兴云能存活下去。给CoreWeave分一些,给Crusoe分一些,给Lambda分一些。为什么这对英伟达是有利的?首先,你同意"分散市场"这种说法吗?
**Dwarkesh Patel:** And Nvidia is known for divvying up the scarce allocation not just based on highest bidder but rather on, hey, we want to make sure that these neo-clouds exist. Let's give some to CoreWeave. Let's give some to Crusoe. Let's give some to Lambda. Why is it good for Nvidia? First of all, would you agree with this characterization of fracturing the market?
**Jensen Huang:** 不。不。对。你的前提就是错的。我们在这些事情上足够审慎。我们在这些事情上非常审慎。
**Jensen Huang:** No. No. Yeah. Your premise is just wrong. We're sufficiently mindful about these things. We're very mindful about these things.
**Jensen Huang:** 首先,如果你不下采购订单(PO, Purchase Order),再怎么说也没用。所以在我们收到PO之前,我们能做什么?第一件事是,我们非常努力地跟每个人做需求预测,因为这些东西的生产周期很长,数据中心的建设也需要很长时间。所以我们通过预测来对齐供需关系之类的。好了,这是第一步。
**Jensen Huang:** First of all, if you don't place a PO, all the talking in the world won't make a difference. And so until we get a PO, what are we going to do? And so the first thing is we work really hard with everybody to get a forecast done because these things take a long time to build and the data centers take a long time to build. And so we align ourselves with demand and supply and things like that through forecasting. Okay, that's job number one.
**Jensen Huang:** 第二,每个人——你知道,我们已经尝试跟尽可能多的人做预测,但最终你还是得下订单。也许因为某些原因你没下订单——我能怎么办?所以到最后就是先来先得(first in first out)。但除此之外,如果你还没准备好,因为你的数据中心还没就绪或者某些组件还没到位,无法让你建起数据中心,我们可能会决定先服务另一个客户。这纯粹是在最大化我们自有工厂的产出。所以我们可能会做一些调整。除此之外,优先级就是先来先得。
**Jensen Huang:** That's just maximizing the throughput of our own factory. And so we might do some adjustments there. Aside from that, the prioritization is first in first out.
**Dwarkesh Patel:** 对。你得下采购订单。如果你不下,当然也有一些关于这方面的故事,你知道,比如所有这些事情的起因是一篇报道说Larry和Elon跟我一起吃饭,他们在饭桌上求GPU。
**Dwarkesh Patel:** Yeah. You got to place a PO. If you don't place a PO, now of course there are stories about that, you know, like for example, all of this kind of started from an article about Larry and Elon having dinner with me where they begged for GPUs.
**Jensen Huang:** 那从来没发生过。我们确实——我们绝对一起吃过饭。我们绝对一起吃过饭。那是一顿非常愉快的晚餐。他们在任何时候都没有求过GPU,所以他们只需要下订单,一旦下了订单,我们会尽最大努力把产能给他们。对。我们没那么复杂。
**Jensen Huang:** That never happened. We had — we absolutely had dinner. We absolutely had dinner. And it was a wonderful dinner. In no time did they beg for GPUs, and so they just had to place an order and once they place an order we do our best to get the capacity to them. Yeah. We're not complicated.
**Dwarkesh Patel:** 好吧。听起来就是一个排队系统,然后根据你的数据中心是否就绪以及你什么时候下了采购订单,你会在某个时间拿到货。但这听起来仍然不是出价最高的人优先。有什么原因要这样做吗?
**Dwarkesh Patel:** Okay. So it sounds like there's a queue and then based on whether your data center is ready and when you place a purchase order, you get them a certain time. But it still doesn't sound like highest bidder just gets it. Is there a reason to do it?
**Jensen Huang:** 我们从来不那样做。
**Jensen Huang:** We never do that.
**Dwarkesh Patel:** 好。
**Dwarkesh Patel:** Okay.
**Jensen Huang:** 我们从来不那样。
**Jensen Huang:** We never do.
**Dwarkesh Patel:** 为什么不直接让出价最高的人优先呢?
**Dwarkesh Patel:** Why not just do highest bidder?
**Jensen Huang:** 因为那是不好的商业做法。你定好价格,然后人们决定买不买。我知道芯片行业其他公司在需求旺盛时会涨价。但我们就是不——这从来不是我们的做法。你可以信赖我们,你知道的。我更愿意做一个可靠的、作为行业基石的存在。你不需要去猜测。
**Jensen Huang:** Because it's a bad business practice. You set your price. You set your price and then people decide to buy it or not. And I understand that others in the chip industry change their prices when demand is higher. But we just don't — that's just never been a practice of ours. You can count on us, you know. I prefer to be dependable, to be the foundation of the industry. And you don't need to second guess.
**Jensen Huang:** 你知道,如果我给你报了价——我们给你报了价,就是那个价。
**Jensen Huang:** You know, if I quoted you a price — we quoted you a price, that's it.
**Jensen Huang:** 如果需求暴涨,那就暴涨吧。
**Jensen Huang:** And if demand goes through the roof, so be it.
**Dwarkesh Patel:** 而从另一端来看,这也是你们和台积电(TSMC)保持良好合作关系的原因,对吧?
**Dwarkesh Patel:** And on the other end, that's why you have a productive relationship with TSMC, right?
**Jensen Huang:** 对。对。对。英伟达已经经营了——我们跟台积电合作快30年了,英伟达和台积电之间没有正式的法律合同。总会有一些大致公平的安排,有时候是我占了便宜,有时候是我吃了亏。有时候我拿到了更好的条件,有时候差一些。但总体来说这段关系非常好,我完全信任他们。我完全依赖他们。
**Jensen Huang:** Yeah. Yeah. Yeah. Nvidia has been in business, we've been doing business with them for I guess coming up on 30 years and Nvidia and TSMC don't have a legal contract. There's always some rough justice and sometimes I'm right, sometimes I'm wrong. Sometimes I got a better deal, sometimes I got a worse deal. But overall the relationship is incredible and I can completely trust them. I completely depend on them.
**Jensen Huang:** 而英伟达有一件你可以确信的事:今年Vera Rubin会非常出色。明年Vera Rubin Ultra会问世。后年Feynman会到来,再后面的一年我还没公布名字。所以每一年你都可以指望我们。而你得去找世界上另一个ASIC团队。随便找一个ASIC团队,你能说"我可以把全部身家——我可以把整个业务都押在你身上,你每一年都会为我交付"吗?你的成本,你的token成本(token cost),每一年都会下降一个数量级。我可以像依赖时钟一样依赖它。
**Jensen Huang:** And one of the things that you can count on with Nvidia is that this year Vera Rubin is going to be incredible. Next year Vera Rubin Ultra will come. The year after that Feynman will come and the year after that I haven't introduced the name yet. And so every single year you can count on us. And you're going to have to go find another ASIC team in the world. Pick your ASIC team where you can say I can bet the farm — I can bet my entire business that you will be here for me every single year. Your cost, your token cost will decrease by an order of magnitude every single year. I can count on it like I can count on the clock.
**Jensen Huang:** 好吧,我刚才说到了台积电。历史上没有任何其他晶圆代工厂(foundry)你能这样说。但今天你可以这样说英伟达。你每一年都可以信赖我们。如果你想买价值10亿美元的AI工厂(AI factory)算力,没问题。如果你想买1亿美元的,没问题。你想买1000万美元的,或者只是一个机柜(rack),没问题。或者只是一张显卡,好吧,没问题。如果你想下1000亿美元的AI工厂订单,没问题。我们是今天世界上唯一一家你可以这样说的公司。
**Jensen Huang:** Well, I just said something about TSMC. No other foundry in history can you possibly say that. You can say that about Nvidia today. You can count on us every single year. If you would like to buy a billion dollars worth of AI factory compute, no problem. If you'd like to buy $100 million, no problem. You'd like to buy $10 million or just one rack, not a problem. Or just one graphics card, okay, no problem. If you would like to place an order for a hundred billion dollar AI factory, no problem. We're the only company in the world where you can say that today.
**Jensen Huang:** 我也可以这样说台积电。我要买10亿的。没问题。我们只需要走一下规划流程,做所有成熟企业会做的事情。
**Jensen Huang:** I can say that about TSMC as well. I want to buy one — buy 1 billion. No problem. We just got to go through the process of planning for it and you know all the things that mature people do.
**Jensen Huang:** 所以我认为英伟达成为全球AI产业基石的能力——这是一个我们花了几十年才到达的位置。巨大的投入,巨大的奉献,以及我们公司的稳定性、一致性,这真的非常非常重要。
**Jensen Huang:** And so I think the ability for Nvidia to be the foundation of the world's AI industry — this is a position that has taken us several decades to arrive at. Enormous commitment, enormous dedication, and the stability of our company, the consistency of our company is really really important.
**Dwarkesh Patel:** 好的。我想聊聊中国的话题。
**Dwarkesh Patel:** Okay. I want to ask about China.
**Jensen Huang:** 好的。
**Jensen Huang:** Yep.
**Dwarkesh Patel:** 我总是喜欢——其实我自己也不确定向中国卖芯片到底是好是坏,但我喜欢跟嘉宾唱反调。所以当Dario上节目的时候,他是支持出口管制的,我就问他为什么美国和中国不能各自都拥有"一个数据中心里的天才国度"。既然你站在另一边,我就从反方向来问你。
**Dwarkesh Patel:** And I always like to take — I don't actually know what I think about whether it's good to sell chips to China or not, but I like to play devil's advocate against my guest. So when Dario was on, who supports export controls, I asked him why can't America and China both have a country of geniuses in a data center. But since you're on the opposite side, I'll ask you in the opposite way.
**Dwarkesh Patel:** 而且你看,有一种思考方式是——Anthropic几天前刚刚宣布了一个叫Mythos的模型——他们甚至都不打算公开发布,因为他们说这个模型具有如此强大的网络攻击能力(cyber offensive capability),以至于在确保相关零日漏洞(zero-day)被修补之前,他们不认为世界已经准备好了。他们说这个模型在每一个主流操作系统、每一个浏览器中都发现了数以千计的高危漏洞。它还在OpenBSD中发现了一个漏洞——OpenBSD是一个专门为不存在零日漏洞而设计的操作系统,已经存在了27年,它居然找到了一个。
**Dwarkesh Patel:** And look, one way to think about it is Anthropic actually announced a couple days ago this model Mythos — they're not even releasing publicly because they say it has such cyber offensive capabilities that we don't think the world is ready until we make sure these zero days are patched up. But they say it found thousands of high severity vulnerabilities across every major operating system, every browser. It found one in OpenBSD which is this operating system that's been specifically designed to not have zero days, and it found one — for 27 years it's existed.
**Dwarkesh Patel:** 所以如果中国公司、中国的实验室和中国政府能够获得AI芯片来训练一个像Claude Mythos这样具有网络攻击能力的模型,并用更多的算力运行数百万个实例,问题是:这对美国公司、对美国国家安全构成威胁吗?
**Dwarkesh Patel:** And so if Chinese companies and Chinese labs and the Chinese government had access to the AI chips to train a model like Claude Mythos with these cyber offensive capabilities and run millions of instances of it with more compute, the question is, is that a threat to American companies, to American national security?
**Jensen Huang:** 首先,Mythos是用相当普通的算力和相当普通的算力量训练出来的,但训练它的是一家非凡的公司。所以训练它所用的算力数量和类型,在中国是完全可以获得的。所以你首先得意识到中国是有芯片的。他们生产了全球60%的主流芯片,也许更多。这对他们来说是一个非常大的产业。他们拥有世界上一些最优秀的计算机科学家。正如你所知,所有这些AI实验室中的大多数AI研究人员,大多数是中国人。他们拥有全球50%的AI研究人员。
**Jensen Huang:** First of all, Mythos was trained on fairly mundane capacity and a fairly mundane amount of it, by an extraordinary company. And so the amount of capacity and the type of compute that it was trained on is abundantly available in China. And so you just have to first realize that chips exist in China. They manufacture 60% of the world's mainstream chips, maybe more. It's a very large industry for them. They have some of the world's greatest computer scientists. As you know, most of the AI researchers in all of these AI labs, most of them are Chinese. They have 50% of the world's AI researchers.
**Jensen Huang:** 所以问题是,如果你担心他们——考虑到他们已经拥有的所有这些资源——他们有充足的能源,有大量的芯片,有大部分的AI研究人员。如果你担心他们,创造一个安全世界的最佳方式是什么?那么,打压他们、把他们变成敌人,很可能不是最好的答案。他们是竞争对手。我们希望美国赢。但我认为保持对话、保持研究层面的交流,可能是最安全的做法。
**Jensen Huang:** And so the question is if you're concerned about them, what is the — considering all the assets they already have? They have an abundance of energy. They have plenty of chips. They got most of the AI researchers. If you're worried about them, what is the best way to create a safe world? Well, victimizing them, turning them into an enemy, likely isn't the best answer. They are an adversary. We want the United States to win. But I think having a dialogue and having a research dialogue is probably the safest thing to do.
**Jensen Huang:** 这是一个由于我们当前把中国视为对手的态度而明显缺失的领域。我们的AI研究人员和他们的AI研究人员之间保持对话,这一点至关重要。我们必须努力就哪些方面不应使用AI达成共识。至于发现软件中的漏洞(bug)——当然,这正是AI应该做的事情。它会在大量软件中发现漏洞吗?当然会。软件里有无数的漏洞,AI软件本身也有很多漏洞。所以这就是AI的职责所在。我很高兴AI已经发展到了能帮助我们大幅提高生产力的水平。
**Jensen Huang:** This is an area that is glaringly missing because of our current attitude about China as an adversary. It is essential that our AI researchers and their AI researchers are actually talking. It is essential that we try to both agree on what not to use the AI for. With respect to finding bugs in software — of course, that's what AI is supposed to do. Is it going to find bugs in a lot of software? Of course. There's lots and lots of bugs. There are lots of bugs in the AI software. And so that's what AI is supposed to do. And I'm delighted that AI has reached a level where it could help us be so much more productive.
**Jensen Huang:** 有一件被严重低估的事情,那就是围绕网络安全(cyber security)、AI网络安全、AI安全(AI security)和AI隐私(AI privacy)以及AI安全性(AI safety)的生态系统的丰富程度。整个AI初创企业生态系统都在努力为我们创造这样一个未来:你有一个极其强大的AI智能体(AI agent),同时有成千上万的AI智能体守护它的安全、确保它的可靠。这样的未来一定会到来。而那种让一个AI智能体到处运行、却没有任何人看管的想法,简直是疯狂的。所以我们非常清楚,这个生态系统必须蓬勃发展。
**Jensen Huang:** One of the things that is underemphasized is the richness of ecosystem around cyber security, AI cyber security and AI security and AI privacy and AI safety. That whole ecosystem of AI startups that are trying to create this future for us where you have one AI agent that's incredible surrounded by thousands of AI agents keeping it safe, keeping it secure. That future surely is going to happen. And the idea that you're going to have an AI agent running around with nobody watching after it is kind of insane. And so we know very well that this ecosystem needs to thrive.
**Jensen Huang:** 事实证明,这个生态系统需要开源(open source)。这个生态系统需要开放模型(open models),需要开放的技术栈(open stacks),这样所有的AI研究人员和优秀的计算机科学家才能去构建同样强大的AI系统,来确保AI的安全。我们需要做到的一件事,就是保持开源生态系统的活力,这一点不容忽视。而其中很大一部分贡献来自中国。我们不能扼杀这些。
**Jensen Huang:** It turns out this ecosystem needs open source. This ecosystem needs open models. They need open stacks so that all of these AI researchers and all these great computer scientists can go build AI systems that are as formidable and can keep AI safe. And one of the things that we need to make sure that we do is we keep the open-source ecosystem vibrant, and that can't be ignored. And a lot of that is coming out of China. We had to not suffocate that.
**Jensen Huang:** 关于中国的问题,我们当然希望美国拥有尽可能多的算力(computing)。我们受制于能源。但你知道,很多人正在努力解决这个问题,我们不能让能源成为我们国家的瓶颈。
**Jensen Huang:** You know, with respect to China, we want to have — of course we want United States to have as much computing as possible. We're limited by energy. But you know we got a lot of people working on that and we got to not make energy a bottleneck for our country.
**Jensen Huang:** 但我们同样希望确保的是,全世界所有的AI开发者都在美国技术栈(American tech stack)上进行开发,并将AI的贡献和进展——特别是在开源领域——提供给美国的生态系统。如果人为制造出两个生态系统——一个是只运行在中国技术栈上的开源生态系统,另一个是运行在美国技术栈上的封闭生态系统——那将是极其愚蠢的。我认为这对美国来说将是一个糟糕至极的结果。
**Jensen Huang:** But what we also want is we want to make sure that all the AI developers in the world are developing on the American tech stack and making the contributions, the advancements of AI, especially when it's open source, available to the American ecosystem. And it would be extremely foolish to create two ecosystems — the open source ecosystem and it only runs on the Chinese tech stack, and a closed ecosystem and that runs on the American tech stack. I think that would be a horrible outcome for United States.
**Dwarkesh Patel:** 这里面涉及很多方面——让我先梳理一下回应。我认为,回到算力差距和黑客攻击的问题上,关键在于:是的,他们有算力,但有一些估计表明,由于他们还停留在7纳米工艺——因为芯片制造出口管制,他们没有EUV光刻机——他们实际能生产的浮点运算量(flops)大约只有美国的十分之一。那么,他们最终能不能训练出像Mythos这样的模型?可以。但问题是,因为我们拥有更多的算力,美国的实验室能够率先达到这些能力水平。
**Dwarkesh Patel:** Since there are a lot of things — let me just triage the response. I mean I think the concern going back to the flop difference and the hacking is yes they have compute, but there's some estimates that because they're at 7 nanometer — they don't have EUV because of chip making export controls — the amount of flops they're able to actually produce, they have like one-tenth the amount of flops that the US has. And so with that, could they train eventually a model like Mythos? Yes. But the question is because we have more flops, American labs are able to get to these level capabilities first.
**Dwarkesh Patel:** 而且因为Anthropic率先达到了这个水平,他们会说:好的,我们先暂缓一个月,在此期间让所有这些美国公司获得访问权限,让它们修补好所有漏洞,然后我们再进一步发布。如果他们——即使他们训练出了这样一个模型,大规模部署它的能力——你知道,如果你有一个网络黑客,拥有一百万个和只有一千个相比,危险程度完全不同——所以推理算力(inference compute)确实非常关键。
**Dwarkesh Patel:** And because Anthropic got to it first they say okay we're going to hold on to it for a month while all these American companies — we give them access to it, they're going to patch up all their vulnerabilities, and now we release it further. If they — even if they train a model like this, the ability to deploy it at scale — you know, if you had a cyber hacker it's much more dangerous if they have a million of them versus a thousand of them — so that inference compute really matters a lot.
**Dwarkesh Patel:** 事实上,他们拥有如此多的研究人员并且水平如此之高,这恰恰是令人恐惧的地方,因为是什么让工程师和研究人员变得更有生产力?是算力。如果你和美国任何一家实验室交谈,他们都会说瓶颈就在算力。DeepSeek的创始人或管理层也有过类似的说法——他们说限制我们的就是算力。
**Dwarkesh Patel:** And in fact the fact that they have so many researchers and are so good is the thing that makes it so scary, because what is it that makes engineers and researchers more productive? It's compute. If you talk to any lab in America they say the thing that's bottlenecking them is compute. And there are quotes from DeepSeek founder or leadership — they say like the thing we're bottlenecked on is compute.
**Dwarkesh Patel:** 所以问题就变成了:由于美国公司拥有更多算力,率先达到Mythos级别的能力,在中国因算力不足而无法达到之前,让我们的社会为之做好准备——这难道不是更好的选择吗?
**Dwarkesh Patel:** So then the question is isn't it better that we get — American companies, because they have more compute, get to the level of mythos level capabilities first, prepare our society for it before China can get to it because they have less compute.
**Jensen Huang:** 我们应该始终走在前面,应该始终拥有更多。但要让你所描述的情形成为现实,你必须把它推到极端。他们必须完全没有算力。而如果他们有一些算力,问题就变成了需要多少才够。中国拥有的算力规模是巨大的——你说的可是全球第二大计算市场。如果他们想要部署、集中他们的算力,他们有足够多的算力可以集中调配。
**Jensen Huang:** We should always be first and we should always have more. But in order for that outcome — for what you described to be true — you have to take it to the extremes. They have to have no compute. And if they have some compute, the question is how much is needed. The amount of compute they have in China is enormous — I mean, you're talking about the country is the second largest computing market in the world. If they want to deploy, aggregate their compute, they got plenty of compute to aggregate.
**Dwarkesh Patel:** 但这是真的吗?我是说,有人做过这些估算,他们说中芯国际(SMIC)在制程节点上实际是落后的。所以他们——
**Dwarkesh Patel:** But is that true? I mean, there's people who do these estimates and they're like, well, SMIC is actually behind on the process nodes. So they're —
**Jensen Huang:** 我正要告诉你。
**Jensen Huang:** I'm about to tell you.
**Dwarkesh Patel:** 好的。
**Dwarkesh Patel:** Okay.
**Jensen Huang:** 他们拥有的能源是惊人的,不是吗?AI本质上是一个并行计算(parallel computing)问题,不是吗?
**Jensen Huang:** The amount of energy they have is incredible, isn't that right? AI is a parallel computing problem, isn't it?
**Jensen Huang:** 他们为什么不能把四倍、十倍数量的芯片组合在一起?因为能源是免费的。他们有如此多的能源。他们有完全空置的数据中心,电力全部就绪。他们——你知道的,他们有鬼城,他们有"鬼数据中心"。他们有如此多的基础设施容量。如果他们想,他们可以直接堆叠更多的芯片,即使是7纳米的。而且他们的芯片制造能力在全球也是名列前茅。半导体行业都知道,他们在主流芯片领域占据了垄断地位。他们产能过剩,拥有太多的制造能力。
**Jensen Huang:** Why can't they just put four, 10 times as many chips together? Because energy is free. They have so much energy. They have data centers that are sitting completely empty, fully powered. They've — you know, they have ghost cities. They have ghost data centers. They have so much capacity of infrastructure. If they wanted to, they just gang up more chips even if they're seven nanometer. And their capacity of building chips is one of the largest in the world. The semiconductor industry knows that they monopolize mainstream chips. They have overcapacity. They have too much capacity.
**Jensen Huang:** 所以,认为中国不会拥有AI芯片的想法完全是无稽之谈。当然,如果你问我,假设全世界除美国外都没有算力,美国是不是会更领先?但那根本不是一个现实的结果,那不是一个真实的场景。他们已经有了大量的算力。你所担忧的那个门槛,他们早已达到并且超越了。
**Jensen Huang:** And so the idea that China won't be able to have AI chips is completely nonsense. Now, of course, if you ask me, would United States be further ahead if the entire world had no compute at all? But that's just not an outcome. That's not a scenario that's true. They have plenty of compute already. The amount of threshold they need for the concern you're worried about, they've already reached that threshold and beyond.
**Jensen Huang:** 所以我认为你误解了一点:AI是一个五层蛋糕(five layer cake)。最底层是能源。当你拥有充裕的能源时,它可以弥补芯片的不足。如果你有充裕的芯片,它可以弥补能源的不足。例如,美国的能源是稀缺的,这就是为什么英伟达(NVIDIA)必须不断推进我们的架构,进行这种极致的协同设计(co-design),以便用我们出货的有限芯片——好,用有限的芯片,在能源如此紧张的情况下,我们的每瓦吞吐量(throughput per watt)远超常规。
**Jensen Huang:** And so I think you misunderstand that AI is a five layer cake. And at the lowest layer is energy. When you have abundance of energy, it makes up for chips. If you have abundance of chips, it makes up for energy. For example, United States is scarce on energy, which is the reason why Nvidia has to keep advancing our architecture and do this extreme co-design so that with the few chips that we ship — okay, with the few chips, because the amount of energy is so limited, our throughput per watt is off the charts.
**Jensen Huang:** 但如果你的电力完全充裕、几乎免费——你何必在乎每瓦性能(performance per watt)呢?你有的是电。所以7纳米芯片本质上就相当于Hopper。Hopper的能力——我得告诉你,今天的模型基本上都是在Hopper这一代上训练的。所以Hopper、7纳米芯片完全够用。能源的充裕就是他们的优势。
**Jensen Huang:** But if your amount of watts is completely abundant, it's free — what do you care about performance per watt for? You have plenty. So 7 nanometer chips are essentially Hopper. The ability for Hopper — I got to tell you, today's models are largely trained on Hopper, you know, Hopper generation. And so Hopper, 7 nanometer chips are plenty good. The abundance of energy is their advantage.
**Dwarkesh Patel:** 但还有一个问题:鉴于他们的制造能力,他们真的能生产足够多的芯片吗——
**Dwarkesh Patel:** But then there's a question of okay, well, can they actually manufacture enough chips given their —
**Jensen Huang:** 他们已经在生产了。证据是什么?华为(Huawei)刚刚创下了公司历史上单年度最高营收。
**Jensen Huang:** But they do. What's the evidence? Huawei just had the largest single year in the history of their company.
**Dwarkesh Patel:** 他们出货了多少芯片?
**Dwarkesh Patel:** How many chips did they ship?
**Jensen Huang:** 大量的。数百万颗。数百万颗远远超过——远远超过Anthropic拥有的。
**Jensen Huang:** A ton. Millions. Millions is way more — way more than Anthropic has.
**Dwarkesh Patel:** 那这里就有一个问题:中芯国际能出货多少逻辑芯片,以及能出货多少存储芯片。
**Dwarkesh Patel:** So there's a question of how much logic SMIC can ship and there's a question of how much memory.
**Jensen Huang:** 我告诉你实际情况。他们有足够多的逻辑芯片,也有足够多的HBM2内存(HBM2 memory)。
**Jensen Huang:** I'm telling you what it is. They have plenty of logic and they have plenty of HBM2 memory.
**Dwarkesh Patel:** 没错。但你也知道,训练和推理这些模型时,瓶颈往往在于内存带宽(memory bandwidth)。所以如果你用的是HBM2——我不记得具体数字了——但和你们最新的产品相比,内存带宽差距可能接近一个数量级,这——
**Dwarkesh Patel:** Right. But as you know the bottleneck often in training and doing inference on these models is the amount of bandwidth. So if you have HBM2 — I don't know the numbers off hand — but like versus the newest thing you have, you know, it can be almost an order of magnitude difference in memory bandwidth, which is —
**Jensen Huang:** 华为是一家网络公司。华为是一家网络公司。
**Jensen Huang:** Huawei is a networking company. Huawei is a networking company.
**Dwarkesh Patel:** 但这不能改变一个事实:最先进的HBM需要EUV光刻技术。
**Dwarkesh Patel:** But that doesn't change the fact that you need EUV for the most advanced HBM.
**Jensen Huang:** 不对。完全不对。你可以把它们组合在一起,就像我们用NVLink72把芯片组合在一起一样。他们已经展示了用硅光子(silicon photonics)技术将所有这些计算节点连接在一起,组成一台巨型超级计算机。你的前提就是错的。事实是,他们的AI发展进行得很顺利。
**Jensen Huang:** Not true. Not at all true. You could gang them together just like we gang them together with NVLink72. They've already demonstrated silicon photonics connecting all of these compute together into one giant supercomputer. Your premise is just wrong. The fact of the matter is their AI development is going just fine.
**Jensen Huang:** 而且全世界最优秀的AI研究人员——正因为他们算力有限——也会想出极其聪明的算法。还记得我说过的话吗:我说摩尔定律(Moore's law)每年大约进步25%。但通过出色的计算机科学,我们仍然可以将算法性能提升十倍。我想说的是,出色的计算机科学才是真正的杠杆。毫无疑问——所有令人惊叹的注意力机制(attention mechanisms)都降低了对算力的需求。我们必须承认,AI领域的大多数进步来自于算法的突破,而不仅仅是原始硬件的堆砌。
**Jensen Huang:** And the best AI researchers in the world — because they are limited in compute — they also come up with extremely smart algorithms. Remember what I said: I said that Moore's law is advancing about 25% per year. However, through great computer science, we could still improve algorithm performance by 10x. What I'm saying is great computer science is where the lever is. There is no question — all the incredible attention mechanisms reduce the amount of compute. We have got to acknowledge that most of the advances in AI came out of algorithm advances, not just the raw hardware.
**Jensen Huang:** 如果大多数进步来自于算法、计算机科学和编程,那你告诉我,他们那支庞大的AI研究人员队伍难道不是他们的根本优势吗?我们已经看到了。DeepSeek绝不是一个微不足道的进步。而当DeepSeek有一天首先在华为平台上发布,那对我们国家来说将是一个可怕的结果。
**Jensen Huang:** Now if most advances came from algorithms and computer science and programming, tell me that their army of AI researchers is not their fundamental advantage. And we see it. DeepSeek is not an inconsequential advance. And the day that DeepSeek comes out on Huawei first, that is a horrible outcome for our nation.
**Dwarkesh Patel:** 为什么?我的意思是,目前像DeepSeek这样的模型,如果是开源的,可以在任何加速器(accelerator)上运行。为什么将来会不是这样呢?
**Dwarkesh Patel:** Why is that? Because I mean, currently you can have a model like DeepSeek that can run on any accelerator if it's open source. Why would that stop being the case in the future?
**Jensen Huang:** 好吧,假设不是这样。假设它针对华为做了优化。假设它针对他们的架构做了优化。那我们就处于劣势了。你描述了一种我认为是好消息的情况——一家公司开发了一个AI模型,它在美国技术栈上运行效果最好。我认为这是好消息。而你把它当作一个坏消息的前提来陈述。让我告诉你什么才是坏消息:全球的AI模型被开发出来,它们在非美国的硬件上运行效果最好。这对我们来说才是坏消息。
**Jensen Huang:** Well, suppose it doesn't. Suppose it's optimized for Huawei. Suppose it's optimized for their architecture. It would put us at a disadvantage. You described a situation that I perceived to be good news — that a company developed an AI model and it runs best on the American tech stack. I saw that as good news. You set it up as a premise that it was bad news. I'm going to give you the bad news: that AI models around the world are developed and they run best on not American hardware. That is bad news for us.
**Dwarkesh Patel:** 我只是看不到证据表明存在这么大的差距,以至于无法在不同加速器之间切换。美国的实验室不也在所有云平台上运行他们的模型,跨越所有——
**Dwarkesh Patel:** I guess I just don't see the evidence that there's these huge disparities that would prevent you from switching accelerators. There's American labs, you know, are running their models across all the clouds, across all —
**Jensen Huang:** 证据就是——你拿一个在英伟达上优化的模型,试着在别的平台上运行——
**Jensen Huang:** The evidence — you take a model that's optimized for Nvidia and you try to run on something else —
**Dwarkesh Patel:** 但美国的实验室确实在这么做——
**Dwarkesh Patel:** But American labs do that —
**Jensen Huang:** 而且它们运行得并不更好。英伟达的成功就是最完美的证据。AI模型在我们的技术栈上创建,在我们的技术栈上运行效果最好——这有什么难以理解的呢?
**Jensen Huang:** And they don't run better. Nvidia's success is perfect evidence. The fact that AI models are created on our stack and run best on our stack — how is that illogical to understand?
**Dwarkesh Patel:** 我就是看看——听着,Anthropic的模型在GPU上运行,在Trainium上运行,在TPU上运行。
**Dwarkesh Patel:** I'm just looking — look, Anthropic's models are run on GPUs. They're run on Trainium. They're run on TPUs.
**Jensen Huang:** 要切换平台需要做大量的工作。但你去看看全球南方(global south),去看看中东,从零开始部署。如果所有AI模型在别人的技术栈上运行效果最好,你现在等于在辩称一个荒谬的论点——说这对美国是件好事。
**Jensen Huang:** A lot of work has to go into it to change. But go to the global south, go to the Middle East, coming out of the box. If all of the AI models run best on somebody else's tech stack, you've got — you've got to be arguing some ridiculous claim right now that that's a good thing for United States.
**Dwarkesh Patel:** 但我不理解这个论点。比如说中国公司率先达到下一个Mythos级别,他们找到了所有的安全漏洞——他们不会先发布给美国软件公司——但他们可以在英伟达硬件上做到,然后把它部署到全球南方。他们在英伟达硬件上做这件事。这怎么——这怎么是好事?我的意思是——好吧,它运行在你的硬件上。
**Dwarkesh Patel:** But I guess I don't understand the argument. Like if say Chinese companies get to the next Mythos first, they find all the security vulnerabilities — they're not releasing American software first — but they can do it on Nvidia hardware and they ship it to the global south. They do it on NVIDIA hardware. Like how is that — how is that good? I mean I just — okay, it runs on your hardware.
**Jensen Huang:** 这不是好事。
**Jensen Huang:** It's not good.
**Dwarkesh Patel:** 对吧?
**Dwarkesh Patel:** Right?
**Jensen Huang:** 这不是好事。所以我们不要让它发生。
**Jensen Huang:** It's not good. So let's not let it happen.
**Dwarkesh Patel:** 你为什么认为这是完全可替代的——如果你不卖给他们算力,华为就会精确地填补这个空缺?他们是落后的,对吧?他们的芯片比你的差。
**Dwarkesh Patel:** Why do you think it's perfectly fungible — that if you didn't ship them compute it would exactly be replaced by Huawei? They are behind, right? They have worse chips than you.
**Jensen Huang:** 这完全——现在就有证据。他们的芯片产业规模巨大。
**Jensen Huang:** It's completely — there's evidence right now. Their chip industry is gigantic.
**Dwarkesh Patel:** 你只需要看看H200和华为910C之间的浮点运算量、带宽或内存对比就知道了。大概只有一半的水平。
**Dwarkesh Patel:** You can just look at the flop or bandwidth or memory comparisons between the H200 and the Huawei 910C. It's like half.
**Jensen Huang:** 他们用更多的芯片。他们用两倍的数量。
**Jensen Huang:** They use more of it. They use twice as many.
**Dwarkesh Patel:** 我觉得你的论点似乎是:他们有大量的能源随时可用,对吧?他们需要用芯片来填满这些设施——
**Dwarkesh Patel:** I guess it seems like your argument is they have all this energy that's ready to go, right? And they need to fill it with chips —
**Jensen Huang:** 而且他们擅长制造。
**Jensen Huang:** And they're good at manufacturing.
**Dwarkesh Patel:** 我相信最终他们确实能在制造产量上超过所有人,但关键是这几年。
**Dwarkesh Patel:** And I'm sure eventually they would be able to just out-manufacture everybody, but there's these few critical years.
**Jensen Huang:** 你说的关键年份是什么意思?
**Jensen Huang:** What is the critical year you're talking about?
**Dwarkesh Patel:** 接下来这几年——我们有了这些能够发动网络攻击的模型。如果关键年份、接下来的关键年份至关重要,那我们就必须确保全世界所有的AI模型都建立在美国技术栈之上。这些关键年份——
**Dwarkesh Patel:** These next few years — we've got these models that are going to do all the cyber attacks. If the critical years, the next critical years is critical, then we have to make sure that all of the world's AI models are built on American tech stack. These critical years —
**Jensen Huang:** 好的,那这如何阻止——如果它们建立在美国技术栈上,如何阻止他们——如果他们拥有更先进的能力——向我们发起相当于Mythos级别的网络攻击——
**Jensen Huang:** Okay, how would that prevent — if they're built on American tech stack, how would that prevent them from — if they have more advanced capabilities — from launching the Mythos equivalent cyber attacks on —
**Dwarkesh Patel:** 两种情况下都没有保证——
**Dwarkesh Patel:** There's no guarantee either way —
**Jensen Huang:** 但如果你更早拥有,我们就可以提前做好准备。
**Jensen Huang:** But if you have it earlier, we can prepare for it.
**Dwarkesh Patel:** 听着,你为什么——你为什么让AI产业的一个层级失去整个市场,来让AI产业的另一个层级受益?AI有五个层级,每一个层级都必须成功。最需要成功的层级实际上是AI应用。你为什么如此执着于那个AI模型,那一家公司?
**Dwarkesh Patel:** Listen, why are you — why are you causing one layer of the AI industry to lose an entire market so that you could benefit another layer of the AI industry? There's five layers and every single layer has to succeed. The layer that has to succeed most is actually the AI applications. Why are you so fixated on that AI model, that one company?
**Jensen Huang:** 为了什么原因?
**Jensen Huang:** For what reason?
**Dwarkesh Patel:** 因为那些模型使得这些极具攻击性的能力成为可能,而你需要算力——能源、芯片、AI研究人员的生态系统使之成为可能。
**Dwarkesh Patel:** Because those models make possible these incredibly offensive capabilities, and you need compute — energy, the chips, the ecosystem of AI researchers make it possible.
**Dwarkesh Patel:** 几个月前,Jane Street花了大约两万个GPU小时,在三个不同的语言模型中植入了后门(back doors)。然后他们挑战我的观众去找出触发短语(trigger phrases)。我刚和设计这个谜题的Rickson聊过,了解了Jane Street收到的一些解题方案。如果你把基础模型看作是在这个位置,后门模型在那个位置,你可以对权重进行线性插值(linearly interpolate)来调节后门的强度,但你也可以外推(extrapolate)来让后门变得更强。在某些情况下,如果你把它增强到足够程度,模型就会直接吐出预设的响应短语。
**Dwarkesh Patel:** A few months ago, Jane Street spent about 20,000 GPU hours trading back doors into three different language models. Then they challenged my audience to find the trigger phrases. I just caught up with Rickson who designed the puzzle about some of the solutions that Jane Street received. If you think the base model was here and the back door model was here, you can kind of linearly interpolate the weights to adjust the strength of the back door, but you can also extrapolate it to make the back door even stronger. And in some cases, if you make it strong enough, the model will just regurgitate what the response phrase was supposed to be.
**Dwarkesh Patel:** 所以,如果你不断放大基础版本和后门版本之间的差异,最终它应该会输出触发短语。但这种技术只在三个模型中的两个上有效。连Rickson自己也不确定为什么在第三个模型上不管用。能够验证一个模型只做你认为它在做的事情,是AI安全领域最重要的开放性问题之一。如果你对这类问题感兴趣,Jane Street正在招聘研究人员和工程师。请访问janestreet.com/thorcash了解更多。
**Dwarkesh Patel:** So, if you keep amplifying the difference between the base version and the back door version, eventually it should spit out the trigger phrase. But this technique only worked on two out of the three models. Even Rickson isn't sure why it didn't work on the other. Being able to verify that a model only does what you think it does is one of the most important open questions in AI security. If this is the kind of problem that excites you, Jane Street is hiring researchers and engineers. Go to janestreet.com/thorcash to learn more.
**Dwarkesh Patel:** 好的,退一步来看——前提必须是中国能够建设足够的7纳米产能。要记住,他们仍然停留在7纳米,而你将推进到3纳米,然后2纳米或1.6纳米的Feynman架构。所以当你在1.6纳米的时候,他们还会在7纳米,他们必须生产出足够多的芯片来弥补差距,而且他们有如此多的能源,你给他们越多的芯片他们就有越多的算力,对吧?所以我只是——这里有个问题——归根结底,他们正在获得越来越多的算力用于训练和推理。
**Dwarkesh Patel:** Okay, stepping back — it has to be the case that China is able to build enough 7 nanometer capacity. And remember, they're still stuck on 7 nanometer while you will move on to 3 nanometer and then 2 nanometer or 1.6 nanometer with Feynman. So while you're on 1.6 nanometer they're still going to be on 7 nanometer, and they have to produce enough of it to make up for the shortfall, and they have so much energy that the more chips you give them the more compute they'd have, right? Like so I just — there's — it comes to the question of ultimately they are getting more compute as input to training and inference.
**Jensen Huang:** 我只是觉得你说话太绝对了。我认为美国应该保持领先。美国的算力是世界其他任何地方的一百倍。美国应该领先。好的,美国确实在领先。英伟达构建最先进的技术。我们确保美国的实验室最先了解到我们的新技术,最先有机会购买。如果他们资金不足,我们甚至会投资他们。美国应该保持领先。我们要尽一切努力确保美国保持领先。第一点。你同意吗?我们正在尽一切努力做到这一点。
**Jensen Huang:** I just think you speak in absolutes. I think that United States ought to be ahead. The amount of compute in United States is 100 times more than anywhere else in the world. The United States ought to be ahead. Okay, the United States is ahead. Nvidia builds the most advanced technologies. We make sure that the US labs are the first to hear about it and the first chance to buy it. And if they don't have enough money, we even invest in them. The United States ought to be ahead. We want to do everything we can to make sure the United States is ahead. Number one point. Do you agree? And we're doing everything we can to do that.
**Dwarkesh Patel:** 但把芯片卖给中国怎么能让美国保持——
**Dwarkesh Patel:** But how is shipping chips to China keeping the US —
**Jensen Huang:** 他们遇到了瓶颈。我们有Vera Rubin给美国用。现在,美国——我算不算美国的一部分?你认为我是美国的一份子吗?
**Jensen Huang:** They're bottlenecked. We have Vera Rubin for United States. Now, United States — am I in United States? Do you consider me part of the United States?
**Dwarkesh Patel:** 是的。
**Dwarkesh Patel:** Yes.
**Jensen Huang:** 英伟达——你认为英伟达是一家美国公司?好的。第一,为什么我们不能制定一个更平衡的法规,让英伟达能够赢得全球市场,而不是放弃全球市场?你为什么希望美国放弃全世界?芯片产业是美国生态系统的一部分,是美国技术领导力的一部分,是AI生态系统的一部分,是AI领导力的一部分。为什么?为什么你的政策、你的理念导致美国放弃全球大部分市场?
**Jensen Huang:** Nvidia — you consider Nvidia a United States company? Okay. Number one, why is it that we don't come up with a regulation that's more balanced so that Nvidia can win around the world instead of giving up the world? Why would you want United States to give up the world? The chip industry is part of the American ecosystem. It's part of American technology leadership. It's part of the AI ecosystem. It's part of AI leadership. Why? Why is it that your policy, your philosophy leads to United States giving up a vast part of the world's market?
**Dwarkesh Patel:** 这里的论点是——Dario有一句话,他说这就像波音公司(Boeing)吹嘘说我们在卖核武器给朝鲜,但导弹外壳是波音制造的,所以这算是在支持美国技术栈。从根本上说,你是在赋予他们这种能力。
**Dwarkesh Patel:** The claim here is — Dario had this quote where he said it's like Boeing bragging that we're selling North Korea nukes but the missile casings are made by Boeing and that's somehow enabling the US technology stack. Like fundamentally you're giving them this capability.
**Jensen Huang:** 把AI跟你刚才提到的任何东西相提并论,简直是疯了。
**Jensen Huang:** Comparing AI to anything that you just mentioned is lunacy.
**Dwarkesh Patel:** 但AI类似于浓缩铀(enriched uranium),对吧?它可以有正面用途,也可以有负面用途。我们仍然不希望把浓缩铀送到其他国家。
**Dwarkesh Patel:** But AI is similar to enriched uranium, right? And then it can have positive uses, it can have negative uses. We still don't want to send enriched uranium to other countries.
**Jensen Huang:** 谁在送浓缩——
**Jensen Huang:** Who's sending enriched —
**Dwarkesh Patel:** 这是类比浓缩铀。
**Dwarkesh Patel:** The analogy is enriched uranium.
**Jensen Huang:** 因为这是一个糟糕的类比。这是一个不合逻辑的类比。
**Jensen Huang:** Because it's a lousy analogy. It's an illogical analogy.
**Dwarkesh Patel:** 但如果那台计算机能运行一个可以对所有美国软件进行零日漏洞攻击(zero day exploits)的模型——这怎么不算是武器?
**Dwarkesh Patel:** But if that computer can run a model that can do zero day exploits against all American software — how is that not a weapon?
**Jensen Huang:** 首先,解决这个问题的方法是与研究人员对话、与中国对话、与其他国家对话,确保人们不会以那种方式使用技术。这种对话必须发生。好的,第一点。
**Jensen Huang:** First of all, the way to solve that problem is to have dialogues with the researchers and dialogues with China and dialogues with other countries to make sure that people don't use technology in that way. That's a dialogue that has to happen. Okay. Number one.
**Jensen Huang:** 第二点,我们也需要确保美国保持领先。所有的产品——Vera Rubin、Blackwell——在美国都有大量供应。堆积如山。显然,我们的业绩数据可以证明这一点。大量的、大量的。我们拥有的算力规模是巨大的。
**Jensen Huang:** Number two, we also need to make sure that United States is ahead. Everything — Vera Rubin, Blackwell — is available in United States in abundance. Mounds of it. Obviously, our results would show it. Abundance of tons of it. Tons of it. The amount of computing we have is great.
**Jensen Huang:** 我们在这里有令人惊叹的AI资源。这很好。我们必须保持领先。但是,我们也必须认识到,AI不仅仅是一个模型。AI是一个五层蛋糕。AI产业的每一个层级都很重要。我们希望美国在每一个层级都赢,包括芯片这一层。而放弃整个市场,不会让美国在芯片层面、在计算栈层面赢得长期的技术竞赛。这就是事实。
**Jensen Huang:** We have amazing AI resources here. It's great. We have to stay ahead. However, we also have to recognize that AI is not just a model. That AI is a five layer cake. That AI industry matters across every single layer. And we want United States to win at every single layer, including the chip layer. And conceding the entire market is not going to allow United States to win the technology race long-term in the chip layer, in the computing stack. That is just a fact.
**Dwarkesh Patel:** 那我想关键分歧就在于:现在卖芯片给他们,如何帮助我们长期取胜。特斯拉(Tesla)长期以来向中国出售了非常出色的电动汽车,iPhone也在中国销售了很长时间,同样非常出色。但这些都没有产生什么锁定效应(lock-in)。中国照样会制造自己版本的电动车——而且他们正在占据主导地位——智能手机也是如此。
**Dwarkesh Patel:** I guess then the crux comes down to how does selling them chips now help us win in the long term. Like Tesla sold extremely good electric vehicles to China for a long time. iPhones are sold in China, extremely good. They didn't create some lock-in. China will still make their version of EVs and they're dominating — or smartphones, dominating.
**Jensen Huang:** 当我们今天开始对话时,你也承认了——而且你确实承认了——英伟达的处境是非常不同的。你用了"护城河"(moat)这样的词。对我们公司而言,最重要的东西就是我们生态系统的丰富性,这取决于开发者。50%的AI开发者在中国。我们不想——美国不应该放弃这些。
**Jensen Huang:** When we started the conversation today, you would acknowledge — and you acknowledged — that Nvidia's position is very different. You use words like moat. The single most important thing to our company is the richness of our ecosystem, which is about developers. 50% of the AI developers are in China. We don't want to — the United States should not give that up.
**Dwarkesh Patel:** 但我们在美国也有大量的英伟达开发者,这并不妨碍美国的实验室在未来也能使用其他加速器(accelerator)。事实上,他们现在就在使用其他加速器,这完全没问题,也挺好的。我不明白为什么在中国就不能这样——就像谷歌既可以用 TPU 也可以用英伟达一样,如果你把芯片卖给他们的话。
**Dwarkesh Patel:** But we have a lot of Nvidia developers in the US, and that doesn't prevent American labs from also being able to use other accelerators in the future. In fact, right now they're using other accelerators as well, which is fine and great. I don't see why that wouldn't be the case in China as well if you sell them Nvidia chips — just the same way that Google can use TPUs and Nvidia.
**Jensen Huang:** 我们必须持续创新。你大概也知道,我们的市场份额在增长,而不是在下降。你说的那个前提——即使我们在中国参与竞争,最终还是会丢掉那个市场——你面前坐的可不是一个一醒来就认输的人。那种失败者心态、那种失败者前提,在我看来毫无道理。
**Jensen Huang:** We have to keep innovating. And you know, as you probably know, our share is growing, not decreasing. The premise that even if we competed in China that we're going to lose that market anyways — you're not talking to somebody who woke up a loser. And that loser attitude, that loser premise makes no sense to me.
**Jensen Huang:** 我们不是汽车。我们不是汽车。你今天能买这个品牌的车、明天换另一个品牌——很简单。但计算不是这样的。x86 架构至今仍然存在是有原因的。ARM 之所以有那么强的粘性也是有原因的。这些生态系统(ecosystem)很难被替代。替换它们需要耗费巨大的时间和精力,大多数人并不想这么做。
**Jensen Huang:** We are not a car. We are not a car. The fact that I can buy a car, this car brand one day and use another car brand another day — easy. Computing is not like that. There's a reason why the x86 still exists. There's a reason why ARM is so sticky. These ecosystems are hard to replace. It costs an enormous amount of time and energy and most people don't want to do it.
**Jensen Huang:** 所以我们的工作就是持续培育这个生态系统,不断推进技术,让我们能在市场上竞争。基于你所描述的那种前提就放弃一个市场,我根本无法认同。这说不通,因为我不认为美国是失败者。我们的行业不是失败者。那种认输的论调、那种认输的思维方式在我看来毫无道理。
**Jensen Huang:** And so it's our job to continue to nurture that ecosystem, to keep advancing the technology so that we could compete in the marketplace. Conceding a marketplace based on the premise you described, I simply can't acknowledge that. It makes no sense because I don't think the United States is a loser. Our industry is not a loser. And that losing proposition, that losing mindset makes no sense to me.
**Dwarkesh Patel:** 好吧,我换个话题。我只是想确认一下——
**Dwarkesh Patel:** Okay, I'll move on. I just want to make sure —
**Jensen Huang:** 你不用换话题。我聊得挺开心的。
**Jensen Huang:** You don't have to move on. I'm enjoying it.
**Dwarkesh Patel:** 好的,太好了。那我很感谢你愿意继续。
**Dwarkesh Patel:** Okay, great. Then I appreciate that.
**Jensen Huang:** 不过我觉得——也许核心问题是——感谢你跟我绕了这么多圈,因为我觉得这有助于把核心问题给引出来。
**Jensen Huang:** But I think the — maybe the crux — and thanks for walking around the circles with me because then I think it helps bring out what the crux here is.
**Jensen Huang:** 核心问题是你在走极端。你的论证起点就是极端化的——如果我们在这个特定的时间窗口给了他们哪怕一点算力(compute),我们就会失去一切。
**Jensen Huang:** The crux is you're going to extremes. Your argument starts from extremes — that if we give them any compute at all in this narrow moment, we will lose everything.
**Dwarkesh Patel:** 不,我觉得我的论点是——
**Dwarkesh Patel:** No, I think what my argument is —
**Jensen Huang:** 这些极端论调——说实话挺幼稚的。是的。
**Jensen Huang:** Those extremes — they're childish. Yeah.
**Dwarkesh Patel:** 我的意思并不是说存在某个关键的算力阈值——而是每一点边际算力都是有帮助的,对吧?你拥有更多算力,就能训练出更好的模型(model)。
**Dwarkesh Patel:** The idea is not that there is some key threshold of compute — is that any marginal compute is helpful, right? So if you have more compute, you can train a better model.
**Jensen Huang:** 我只是希望你能承认,美国科技产业每多一笔边际销售也是有益的。
**Jensen Huang:** And I just want you to acknowledge that any marginal sales for American technology industry is beneficial.
**Dwarkesh Patel:** 我其实不——我是说,如果运行在那些芯片上的 AI 模型——
**Dwarkesh Patel:** I actually don't — I mean, if the AI models that run on those chips —
**Jensen Huang:** 嗯。
**Jensen Huang:** Yeah.
**Dwarkesh Patel:** ——具备网络攻击(cyber offensive)能力,或者用于训练的模型具备网络攻击能力、可以运行更多实例——这不是核武器,但它使某种形式的武器成为可能。
**Dwarkesh Patel:** — are capable of cyber offensive capabilities, or training models are capable of cyber offense, running more instances — it is not a nuclear weapon, but it enables a weapon of a kind.
**Jensen Huang:** 按照你的逻辑,你对微处理器(microprocessor)和 DRAM 内存也应该说同样的话。你甚至应该对电力也这么说。
**Jensen Huang:** The logic that you use, you might as well say it to microprocessors and DRAMs. You might as well say it to electricity.
**Dwarkesh Patel:** 但实际上,我们确实对制造最先进 DRAM 所需的相关技术实施了出口管制(export control),对吧?我们对中国在各种领域都有各种出口管制。
**Dwarkesh Patel:** But in fact, we do have export controls on the technology that is relevant to making the most advanced DRAM, right? We have all kinds of export controls on China for all kinds of things.
**Jensen Huang:** 我们向中国出售大量的 DRAM 和 CPU。我认为这是正确的做法。
**Jensen Huang:** We sell a lot of DRAM and CPUs into China. And I think it's right.
**Dwarkesh Patel:** 我想这又回到了那个根本性问题——AI 是不是不一样的?如果你拥有一种能在软件中发现零日漏洞(zero-day)的技术,我们是不是应该尽量减少中国率先达到那个水平、取得领先地位的可能性?
**Dwarkesh Patel:** I guess this goes back to the fundamental question of — is AI different, right? If you have the kind of technology that can find these zero days in software, is that something where we want to minimize China's ability to get there first, to be ahead?
**Jensen Huang:** 我们可以控制那个。
**Jensen Huang:** We can control that.
**Dwarkesh Patel:** 如果芯片已经在那边了,他们正在用来训练那个模型,我们怎么控制?
**Dwarkesh Patel:** How do we control that if the chips are already there and they're using that to train that model?
**Jensen Huang:** 我们有大量的算力。我们有大量的 AI 研究人员。我们正在尽最大速度往前冲。
**Jensen Huang:** We have tons of compute. We have tons of AI researchers. We're racing as fast as we can.
**Dwarkesh Patel:** 再说一次,我们拥有的核武器比任何人都多,但我们不会把浓缩铀(enriched uranium)运到任何地方去。
**Dwarkesh Patel:** Again, we have more nuclear weapons than anybody else, but we don't want to send enriched uranium anywhere.
**Jensen Huang:** 我们不是浓缩铀。这只是一块芯片,而且是他们自己也能制造的芯片。
**Jensen Huang:** We're not enriched uranium. It's a chip and it's a chip that they can make themselves.
**Dwarkesh Patel:** 但他们之所以从你这里买是有原因的,对吧?我们也有中国公司创始人的原话,说他们在这项技术上遇到了瓶颈(bottleneck)。
**Dwarkesh Patel:** But there's a reason they're buying it from you, right? And we have quotes from the founders of Chinese companies that say that they're bottlenecked on that technology.
**Jensen Huang:** 因为我们的芯片更好。总体来说,我们的芯片更好。这毫无疑问。在没有我们芯片的情况下——在没有我们芯片的情况下——你能不能承认华为(Huawei)去年创下了历史新高?你能不能承认一大批芯片公司已经上市了?你能不能承认这些事实?
**Jensen Huang:** Because our chips are better. On balance, our chips are better. There's just no question about it. In the absence of our chip — in the absence of our chip — can you acknowledge that Huawei had a record year? Can you acknowledge that a whole bunch of chip companies have gone public? Can you acknowledge that?
**Jensen Huang:** 你能不能也承认,我们曾经在那个市场拥有很大的份额,而现在我们的份额已经大幅缩水了?我们还可以承认中国大约占了全球科技产业的 40%。那个市场——离开那个市场、把那个市场拱手让给竞争对手,对美国科技产业来说是一种伤害。对我们的国家安全是一种伤害。对我们的技术领导力是一种伤害。而所有这一切只是为了一家公司的利益。这在我看来说不通。
**Jensen Huang:** Can you also acknowledge that the fact that we used to have a very large share in that market and we no longer have a large share in that market? We can also acknowledge that China is about 40% of the world's technology industry. That market — to leave that market, concede that market for United States technology industry is a disservice to our country. It is a disservice to our national security. It is a disservice to our technology leadership. All for the benefit of one company. It makes no sense to me.
**Dwarkesh Patel:** 我有点困惑——感觉你在做两个不同的陈述。一个是说如果允许我们参与竞争,我们会赢得与华为的竞争,因为我们的芯片要好得多。另一个是说即使没有我们,他们也会做一模一样的事情。这两个说法怎么能同时成立呢?
**Dwarkesh Patel:** I guess I'm confused — it feels like you're making two different statements. One is that we're going to win this competition with Huawei because our chips are going to be way better if we're allowed to compete. And another is that they would be doing the same exact thing without us anyways. Right? How can those two things be true at the same time?
**Jensen Huang:** 这显然是成立的。在没有更好选择的情况下,你会选择你唯一拥有的那个选项。这有什么不合逻辑的?这太合乎逻辑了。
**Jensen Huang:** It's obviously true. In the absence of a better choice, you'll take the only choice you have. How is that illogical? It's so logical.
**Dwarkesh Patel:** 他们想要英伟达芯片是因为它更好。更好意味着更多算力。更多算力意味着可以训练出更好的模型。
**Dwarkesh Patel:** The reason they want Nvidia chips is they're better. Better is more compute. More compute means you can train a better model.
**Jensen Huang:** 它更好。它更好是因为它更容易编程。我们有更好的生态系统。不管"更好"具体指什么,不管"更好"到底意味着什么。当然我们会卖给他们算力。那又怎样?那又怎样。关键是我们从中获益。别忘了,我们获得了美国技术领导力的收益。我们获得了开发者在美国技术栈(tech stack)上工作的收益。我们获得了——当那些 AI 模型向世界其他地方扩散时,美国技术栈因此成为最适合的平台。我们可以持续推进和传播美国技术。我认为这是一件积极的事情。这是美国技术领导力非常重要的组成部分。
**Jensen Huang:** It's better. It's better because it's easier to program. It's — we have a better ecosystem. Whatever the better is. Whatever the better is. And of course we're going to send them compute. So what? So what. The fact of the matter is we get the benefit. Don't forget, we get the benefit of American technology leadership. We get the benefit of developers working on the American tech stack. We get the benefit — as those AI models diffuse out into the rest of the world, the American tech stack is therefore the best for it. We can continue to advance and diffuse American technology. I believe that is a positive. It's a very important part of American technology leadership.
**Jensen Huang:** 而你所倡导的那种政策,其结果就是美国电信行业被政策逼出了基本上整个世界市场,以至于我们已经不再掌控自己的电信产业了。我不认为那是明智的做法。这有点短视,而且导致了我正在向你描述的那些意外后果,但你似乎很难理解这一点。
**Jensen Huang:** Now, the policy that you're advocating resulted in the American telecommunication industry being policied out of basically the world, to the point where we don't control our own telecommunications anymore. I don't see that as smart. It's a little narrow-minded and it led to unintended consequences that I'm describing to you right now that you seem to have a very hard time understanding.
**Dwarkesh Patel:** 好的,让我们退一步。看起来核心问题在于存在一个潜在的收益和一个潜在的代价,我们正在试图弄清楚收益是否值得付出那个代价。我想让你承认那个潜在的代价——算力是训练强大模型的一种投入。强大的模型确实具有强大的攻击能力,比如网络攻击。
**Dwarkesh Patel:** Okay, let's just step back. It seems like the crux here is there's a potential benefit and there's a potential cost, and we're trying to figure out is the benefit worth the cost. I guess I'm trying to get you to acknowledge the potential cost — that compute is an input to training powerful models. Powerful models do have powerful offensive capabilities like cyber attacks.
**Dwarkesh Patel:** 美国公司率先达到 Claude Mythos 级别的能力是一件好事,然后他们会暂缓发布那些能力,以便美国公司和美国政府能够在这个级别的能力公布之前加强软件防护。如果中国拥有更多算力、更强大的算力,如果他们能更早地做出 Mythos 级别的模型并广泛部署,那将非常糟糕。这种情况没有发生的原因之一,就是因为美国拥有更多算力——这要感谢英伟达这样的公司。这就是向中国出售芯片的代价。所以先把收益放在一边——你能不能承认这是一个潜在的代价?
**Dwarkesh Patel:** It is a good thing that American companies got to Claude Mythos level capabilities first and then now they're going to hold off on those capabilities so that the American companies and American government can make their software more protected before this level capability is announced. If China had had more compute, had more powerful compute, if they could have made a Mythos level model earlier and deployed it widely, that would have been very bad. One of the reasons that hasn't happened is that we have more compute, thanks to companies like Nvidia, in America. That is a cost of sending to China. And so let's leave the benefit aside for a second. Do you acknowledge that this is a potential cost?
**Jensen Huang:** 我也要告诉你另一个潜在的代价,那就是我们让 AI 技术栈中最重要的层之一——芯片层——拱手让出了一整个市场,全球第二大市场,使他们得以发展规模、发展自己的生态系统,使未来的 AI 模型以一种与美国技术栈截然不同的方式被优化。当 AI 向世界其他地方扩散时,他们的标准、他们的技术栈将会优于我们的,因为他们的模型是开源的(open source)。
**Jensen Huang:** I will also tell you the potential cost is we allow one of the most important layers of the AI stack — the chip layer — to concede an entire market, the second largest market in the world, so that they could develop scale, so that they could develop their own ecosystem, so that future AI models are optimized in a very different way than the American tech stack. As AI diffuses out into the rest of the world, their standards, their tech stack will become superior to ours because their models are open.
**Dwarkesh Patel:** 我想我对英伟达的内核工程师(kernel engineer)和 CUDA 工程师有足够的信心,他们应该能够优化——
**Dwarkesh Patel:** I guess I just believe enough in Nvidia's kernel engineers and CUDA engineers to think that they could optimize —
**Jensen Huang:** 你也知道,AI 不仅仅是内核优化。
**Jensen Huang:** AI is more than kernel optimization, as you know.
**Dwarkesh Patel:** 当然不是,但你们可以做很多事情——比如蒸馏(distilling)出一个专门适配你们芯片的模型。
**Dwarkesh Patel:** Of course, but there's so many things you can do — from distilling to a model that's well fit for your chips.
**Jensen Huang:** 我们会尽全力的。
**Jensen Huang:** We're going to do our best.
**Dwarkesh Patel:** 你们有这么多软件积累。很难想象中国生态系统会形成长期的锁定效应(lock-in)。他们有一个稍微好一点的开源模型也就持续一段时间。
**Dwarkesh Patel:** You have all this software. It's just hard to imagine that there's a long-term lock-in to the Chinese ecosystem. They have this slightly better open source model for a while.
**Jensen Huang:** 中国是全球最大的开源软件(open source software)贡献者。这是事实,对吧?中国是全球最大的开放模型(open model)贡献者。事实。而今天这些都构建在美国技术栈之上。也是事实。AI 技术栈的全部五个层次都很重要。美国应该去赢得所有五个层次。每一层都很重要。
**Jensen Huang:** China is the largest contributor to open source software in the world. Fact, right? China is the largest contributor to open models in the world. Fact. Today it's built on the American tech stack. And fact. All five layers of the tech stack for AI is important. United States ought to go win all five of them. They're all important.
**Jensen Huang:** 其中最重要的当然是 AI 应用层。那个向社会扩散的层面——谁用得最多,谁就从这场工业革命中获益最多。但我的观点是每一层都必须成功。如果我们把这个国家吓得以为 AI 是某种核弹,让所有人都憎恨 AI、害怕 AI,我不知道你怎么算帮了美国。你反而是在帮倒忙。
**Jensen Huang:** The one that is the most important, of course, is the AI application layer. The layer that diffuses into society — the one that uses it most will benefit from this industrial revolution most. But my point is that every layer has to succeed. If we scare this country into thinking that AI is somehow a nuclear bomb so that everybody hates AI and everybody's afraid of AI, I don't know how you're helping the United States. You're doing a disservice.
**Jensen Huang:** 如果我们把所有人都吓得不敢做软件工程的工作,因为 AI 会消灭所有软件工程岗位,导致我们一个软件工程师都没有了——我们是在伤害美国。如果我们把所有人都吓得不去读放射科(radiology),因为计算机视觉完全免费、AI 会比放射科医生做得更好——而我们混淆了"工作"(job)和"任务"(task)之间的区别。放射科医生的工作是患者护理,任务是读片。如果我们对这一点理解得如此之深地偏差,把所有人都吓得不去上放射学院,那我们就不会有足够的放射科医生,也不会有足够好的医疗保障。
**Jensen Huang:** If we scare everybody out of doing software engineering jobs because it's going to kill every software engineering job and we don't have any software engineers as a result of that, we're doing a disservice to United States. If we scare everybody out of radiology, so nobody wants to be a radiologist because computer vision is completely free and AI is going to do a better job than a radiologist — and we misunderstand the difference between a job and a task. The job of a radiologist is patient care. The task is to read a scan. If we misunderstand that so profoundly and we scare everybody out of going to radiology school, we're not going to have enough radiologists and good enough healthcare.
**Jensen Huang:** 所以我想表达的是,当你设立了一个如此极端的前提——什么都是零或者无穷大——结果就是我们以一种不符合实际的方式把人吓到了。现实生活不是那样的。我们想让美国保持领先吗?当然想。我们需要在技术栈的每一层都处于领先地位吗?当然需要。当然需要。
**Jensen Huang:** And so I'm making the case that when you make a premise that is so extreme — everything goes from zero or infinity — we end up scaring people in a way that's just not true. Life is not like that. Do we want United States to be first? Of course we do. Do we need to be a leader in every layer of that stack? Of course we do. Of course we do.
**Jensen Huang:** 今天——你提到 Mythos 是因为 Mythos 很重要。当然。那很好。但再过几年,我向你预测,当我们希望美国技术栈、当我们希望美国技术在全世界传播——传播到印度、中东、非洲、东南亚——当我们国家想要出口的时候,因为我们想出口我们的技术、出口我们的标准——到那一天,我希望你和我能再来一次同样的对话。
**Jensen Huang:** Is today — you're talking about Mythos because Mythos is important. Sure. That's fantastic. But in a few years time, I'm making you the prediction that when we want the American tech stack, when we want American technology to be diffused around the world — out to India, out to the Middle East, out to Africa, out to Southeast Asia — when our country would like to export, because we would like to export our technology, we would like to export our standards — on that day, I want you and I to have that same conversation again.
**Jensen Huang:** 到时候我会确切地告诉你,今天这场对话——你的政策主张和你的设想——是如何导致美国毫无理由地拱手让出了全球第二大市场的。我们不应该主动放弃。如果输了,那就输了。但为什么要主动放弃呢?
**Jensen Huang:** And I will tell you exactly about today's conversation about how your policy and how what you imagined literally caused the United States to concede the second largest market in the world for no good reason at all. We shouldn't concede it. If we lose it, we lose it. But why do we concede it?
**Jensen Huang:** 没有人在主张全要或全不要。没有人主张那样的极端,也就是说随时把所有东西都卖给中国。没有人在倡导那样。我们应该始终在国内拥有最好的技术。我们应该始终在国内拥有最多的技术,并且是第一个获得的。但我们也应该努力在全球范围内竞争并取胜。这两件事完全可以同时发生。这需要一些微妙的判断、一些成熟的思维,而不是走极端。这个世界根本就不是非黑即白的。
**Jensen Huang:** Now, nobody is advocating an all or nothing. Nobody's advocating all or nothing, meaning we ship everything to China at all times. Nobody's advocating that. We should always have the best technology here. We should always have the most technology here and the first. But we should also try to compete and win around the world. Both of those things can simultaneously happen. It requires some amount of nuance, some amount of maturity instead of absolutes. The world is just not absolutes.
**Dwarkesh Patel:** 好的,这个论点的关键在于——他们开发了针对自家芯片架构优化的模型,那是他们能造出的最好芯片。几年后这些芯片被出口到全世界,这就确立了一个标准。但由于 EUV 光刻机出口管制,正如我们之前讨论的,你们会走向 1.6 纳米制程——而他们几年后仍然停留在 7 纳米。国内市场他们可能会说——嘿,我们有这么多能源,能大规模制造,我们继续用 7 纳米——但出口的话呢?他们的 7 纳米芯片必须与你的 1.6 纳米芯片竞争,而且他们的模型必须针对 7 纳米优化到如此程度,以至于在 7 纳米上跑他们的模型比在你的 1.6 纳米上跑还要好。
**Dwarkesh Patel:** Okay, the argument hinges on — they've built models that are optimized for their architecture, the best chips that they make, in a few years, and those chips get exported around the world, that sets a standard. But because of EUV export controls, as we said, you're going to move on to 1.6 nanometer — they're still going to be on 7 nanometer even after a few years from now. And it might make sense that domestically they would prefer — hey, we got so much energy, we can manufacture at scale, we'll still keep using 7 nanometer — but the exporting thing? Their 7 nanometer chips have to be competitive against your 1.6 nanometer chips, and their models have to be so far optimized for the 7 nanometer that it's better to run their models on 7 nanometer than to run their models on your 1.6 nanometer.
**Jensen Huang:** 那我们来看看事实吧。Blackwell 比 Hopper 在光刻技术上先进了 50 倍吗?是 50 倍吗?差远了。我一直在反复强调。摩尔定律(Moore's law)已经终结了。从 Hopper 到 Blackwell,纯粹从晶体管层面来看,大概提升了 75%。两者间隔了三年。75%。
**Jensen Huang:** Can we just look at the facts then? Okay. Is Blackwell 50 times more advanced lithography than Hopper? Is it 50 times? Not even close. I just kept saying it over and over again. Moore's law is dead. Between Hopper and Blackwell, from the transistors themselves, call it 75%. It was 3 years apart. 75%.
**Jensen Huang:** 但 Blackwell 的性能是 Hopper 的 50 倍。我的观点是架构(architecture)很重要。计算机科学(computer science)很重要。半导体物理学当然也重要。但计算机科学很重要。AI——AI 的影响力在很大程度上来自于计算栈(computing stack),这就是 CUDA 如此高效的原因,也是 CUDA 如此受欢迎的原因。它是一个生态系统、一种计算架构,提供了巨大的灵活性——如果你想彻底改变一种架构,创造出类似扩散模型(diffusion)这样的东西,或者做解耦式的设计——你完全可以做到。这很容易。
**Jensen Huang:** Blackwell is 50 times Hopper. My point is architecture matters. Computer science matters. Semiconductor physics matter as well. But computer science matters. AI — the impact of AI largely comes from the computing stack, which is the reason why CUDA is so effective, which is the reason why CUDA is so beloved. It's an ecosystem, a computing architecture that allows for so much flexibility that if you wanted to change an architecture completely — create something like diffusion, create something that's disaggregated — you could do so. It's easy to do.
**Jensen Huang:** 所以事实是,AI 既关乎上层的技术栈,也关乎底层的架构。在我们拥有针对我们自己的技术栈和生态系统进行优化的架构和软件栈的前提下——这显然是好的,因为我们今天对话一开始就谈到了英伟达生态系统多么丰富,为什么人们总是喜欢先在 CUDA 上编程。他们确实如此。确实如此,中国的研究人员也是一样。
**Jensen Huang:** And so the fact of the matter is AI is about the stack above as much as it is about the architecture below. To the extent that we have architectures and software stacks that are optimized for our stack, for our ecosystem — it is obviously good because we started the conversation today about how Nvidia's ecosystem is so rich, why people always love programming on CUDA first. They do. They do, and so do the researchers in China.
**Jensen Huang:** 但如果我们被迫离开中国,如果我们被迫离开中国——首先,这是一个政策失误。显然它引发了反弹。显然它对美国产生了不好的结果。它实际上加速了中国的芯片产业发展。它迫使他们整个 AI 生态系统转向聚焦于国产架构。
**Jensen Huang:** But if we are forced to leave China, if we're forced to leave China — well, first of all, it's a policy mistake. Obviously it has backlash. Obviously it has turned out badly for the United States. It enabled — it accelerated their chip industry. It forced all of their AI ecosystem to focus on their internal architectures.
**Jensen Huang:** 现在还不算太晚。但不管怎样,这已经发生了。未来你会看到他们不会停留在 7 纳米。他们显然很擅长制造。他们会从 7 纳米继续向前推进。那么 5 纳米和 7 纳米之间有 10 倍的差距吗?答案是没有。架构很重要。网络互联(networking)很重要。这就是英伟达收购 Mellanox 的原因。网络很重要。能源很重要。所有这些都很重要。不像你试图简化的那样简单。
**Jensen Huang:** It's not too late. But nonetheless, it has already happened. You're going to see in the future they're not stuck at 7 nanometer. Obviously they're good at manufacturing. They will continue to advance from seven and beyond. Now, is there a 10x difference between 5 nanometer and 7 nanometer? The answer is no. Architecture matters. Networking matters. That's why Nvidia bought Mellanox. Networking matters. Energy matters. And so all that stuff matters. It's not simplistic like the way you're trying to distill it.
**Dwarkesh Patel:** 我们可以从中国话题上转移了。不过这其实引出了一个有意思的问题——我们之前讨论了台积电(TSMC)和内存等方面的产能瓶颈。在你们已经是 N3 制程的主要客户、未来还会是 N2 的主要客户的情况下——你有没有想过回到 N7 这种老制程的闲置产能,然后说嘿,AI 的需求太大了,而我们扩展最先进制程的能力跟不上,所以我们要用我们今天掌握的所有关于数值格式的知识和你刚才提到的所有改进,在老制程上做一个 Hopper 或 Ampere 级别的芯片。你觉得这种情况会在 2030 年之前发生吗?
**Dwarkesh Patel:** We can move on from China, but that actually raises an interesting question about — we were discussing earlier these bottlenecks at TSMC and memory and so forth. And so if we're in this world where you're already the majority of N3, at some point you'll be N2, you'll be a majority of that — do you see that you could go back to N7, this spare capacity at an older process node, and say hey, the demand for AI is so great and our capacity to expand the leading edge is not meeting it, so we're going to make a Hopper-or-Ampere-level chip with everything we know about numerics today and all the other improvements you described. Do you see that world happening before 2030?
**Jensen Huang:** 没有这个必要。原因在于,每一代产品的架构不仅仅是晶体管制程的问题。它还包括封装(packaging)、堆叠(stacking)、数值格式和系统架构方面大量的工程工作。当你用完产能想回到另一个旧制程时,那需要的研发投入是没人承担得起的。我们承担得起向前冲。但我不认为我们承担得起往回走。
**Jensen Huang:** It's not necessary to. And the reason for that is because with every generation the architecture is more than just the transistor scale. It also — you're doing so much engineering and packaging and stacking and the numerics and the system architecture. When you run out of capacity to easily go back to another node, that's a level of R&D that no one could afford. You know, we could afford to lean forward. I don't think we could afford to go back.
**Jensen Huang:** 不过话说回来,如果全世界就是告诉我们——如果在那一天,我们来做个思想实验——在那一天我们说,听着,我们永远不会再有更多产能了——我会回去用 7 纳米吗?毫不犹豫。
**Jensen Huang:** Now, if the world simply says — if on that day, let's do the thought experiment — on that day we go, listen, we're just never going to have more capacity ever again — would I go back and use seven? In a heartbeat.
**Dwarkesh Patel:** 是的,那当然会。
**Dwarkesh Patel:** Yeah, of course you would.
**Dwarkesh Patel:** 有人问了一个问题,就是为什么英伟达不同时推进多个完全不同架构的芯片项目。比如你可以做一个类似 Cerebras 的晶圆级芯片(wafer-scale),可以做一个类似 Dojo 的超大封装芯片,可以做一个不用 CUDA 的。你们有资源也有工程人才来并行推进这些。既然谁也不知道 AI 和架构会往什么方向走,为什么要把所有鸡蛋放在一个篮子里呢?
**Dwarkesh Patel:** One question somebody I was talking to had is why Nvidia doesn't run multiple different chip projects at the same time with totally different architectures. So you could do like a Cerebras-style wafer scale. You could do a Dojo-style huge package. You could do one without CUDA, you know. You have the resources and the engineering talent to do all these in parallel. So why put all the eggs in one basket given who knows where AI might go and architectures might go?
**Jensen Huang:** 噢,我们当然可以这么做。只不过我们没有更好的想法。
**Jensen Huang:** Oh, we could. It's just that we don't have a better idea.
**Jensen Huang:** 是的。是的,那些方案我们都可以做。只是它们不会更好。我们会做仿真模拟(simulation),在我们的模拟器里,可以证明它们更差。所以我们不会那么做。是的,我们正在做的项目正是我们想做的项目。
**Jensen Huang:** Yeah. Yeah, we could do all of those things. It's just not better. And we simulate it all. They're in our simulator, provably worse. And so we wouldn't do it. Yeah, we're working on exactly the projects that we want to work on.
**Jensen Huang:** 而且如果工作负载(workload)发生重大变化——我说的不是算法层面,我说的是实际的工作负载——这取决于市场的格局——我们可能会决定增加其他类型的加速器。比如最近我们加入了 Groq,我们会把 Groq 整合进我们的 CUDA 生态系统。
**Jensen Huang:** And if the workload were to change dramatically — and I don't mean the algorithms, I actually mean the workload — and that depends on the shape of the market — we may decide to add other accelerators. Like for example, recently we added Groq, and we're going to fold Groq into our CUDA ecosystem.
**Jensen Huang:** 我们现在这么做是因为令牌(token)的价值已经上涨到可以有不同定价的程度了。就在几年前,令牌要么免费,要么几乎不花钱,对吧?所以现在你可以有不同的客户,这些客户需要不同的响应。正因为客户能赚很多钱——比如我们的软件工程师——如果我能给他们响应更快的令牌,让他们比现在更高效,我愿意为此付费。
**Jensen Huang:** We're doing that now because the value of tokens has gone up so high that you could have different pricing of tokens. Back in the old days, just a couple years ago, tokens are either free or barely expensive, right? And so now you can have different customers and those customers want different answers. And so because the customers make so much money — like for example, our software engineers — if I can give them much more responsive tokens so that they're even more productive than they are today, I would pay for it.
**Jensen Huang:** 但这个市场是最近才出现的。所以我认为我们现在有能力让同一个模型——基于响应时间的不同——划分出不同的市场细分(segment)。这就是我们决定拓展帕累托前沿(Pareto frontier)的原因,创造出一个推理(inference)细分市场,特点是更快的响应时间但吞吐量(throughput)较低。在此之前,更高的吞吐量总是更好的。我们认为可能存在这样一个世界,有非常高单价(ASP)的令牌,即使工厂的吞吐量更低,高单价也能弥补。
**Jensen Huang:** But that market has only recently emerged. And so I think that we now have the ability to have the same model — based on the response time — have different segments. And that's the reason why we decided to expand the Pareto frontier and create a segment of inference that is faster response time even though it's lower throughput. Until now, higher throughput is always better. We think that there could be a world where there could be very high ASP tokens, and even though the throughput is lower in the factory, the ASPs make up for it.
**Jensen Huang:** 是的。这就是我们这么做的原因。但除此之外,从架构角度来看,我认为英伟达的架构是——如果我有更多资金,我宁愿把更多资源投入到现有架构上。
**Jensen Huang:** Yeah. That's the reason why we did it. But otherwise, from an architecture perspective, I think Nvidia's architecture is — I would rather, if I have more money, put more behind the architecture.
**Dwarkesh Patel:** 我觉得这个关于超高溢价令牌以及推理市场细分化的想法非常有意思。
**Dwarkesh Patel:** I think this idea of extremely premium tokens and just the disaggregation of the inference market is very interesting.
**Jensen Huang:** 市场分层,是的。
**Jensen Huang:** The segmentation, yeah.
**Dwarkesh Patel:** 最后一个问题——假设深度学习(deep learning)革命没有发生。英伟达现在会在做什么?当然游戏是一方面,但考虑到——
**Dwarkesh Patel:** Final question — suppose the deep learning revolution didn't happen. What would Nvidia be doing? Obviously games, but given —
**Jensen Huang:** 加速计算(accelerated computing)。
**Jensen Huang:** Accelerated computing.
**Dwarkesh Patel:** 加速计算——和你们一直在做的事情一样。
**Dwarkesh Patel:** Accelerated computing — the same thing you've been doing all along.
**Jensen Huang:** 我们公司的基本前提是摩尔定律——通用计算(general purpose computing)对很多事情来说是够用的,但对很多运算来说并不理想。所以我们把一种叫 GPU 的架构与 CUDA 和 CPU 结合起来,这样就能加速 CPU 的工作负载。不同的代码内核或算法可以被卸载到我们的 GPU 上,结果就是把一个应用加速 100 倍、200 倍。
**Jensen Huang:** The premise of our company is that Moore's law — general purpose computing is good for a lot of things but for a lot of computation is not ideal. And so we combined an architecture called a GPU with CUDA to a CPU so that we can accelerate the workload of the CPU. And so different kernels of code or algorithms could be offloaded onto our GPU and as a result you speed up an application by 100x, 200x.
**Jensen Huang:** 这能用在哪里?显然,工程、科学、物理学、数据处理、计算机图形学(computer graphics)、图像生成——各种各样的领域。即使今天 AI 不存在,英伟达也会非常、非常大。
**Jensen Huang:** And where can you use that? Well, obviously engineering and science and physics and data processing, computer graphics, image generation — I mean all kinds of things. Even if AI doesn't exist today, Nvidia will be very, very large.
**Jensen Huang:** 我认为根本原因在于,通用计算持续扩展的能力基本上已经走到头了。解决办法是通过领域特定加速(domain specific acceleration)。我们最早涉足的领域之一是计算机图形学,但还有很多很多其他领域——粒子物理学(particle physics)、流体力学(fluid dynamics)、结构化数据处理,各种各样受益于 CUDA 的算法。
**Jensen Huang:** And so I think the reason for that is fairly fundamental, which is the ability for general purpose computing to continue to scale has largely run its course. And the way to do that is through domain specific acceleration. And one of the domains that we started with was computer graphics, but there are many, many other domains — particle physics and fluids and structured data processing, all kinds of different types of algorithms that benefit from CUDA.
**Jensen Huang:** 所以我们的使命实际上是把加速计算带给全世界,推动那些通用计算无法胜任的应用类型,并将其扩展到足以突破某些科学领域的能力水平。早期的一些应用包括分子动力学(molecular dynamics)、用于能源勘探的地震数据处理(seismic processing),当然还有图像处理。
**Jensen Huang:** And so our mission was really to bring accelerated computing to the world and advance the type of applications that general purpose computing can't do and scale to the level of capability that helps break through certain fields of science. And so some of the early applications were molecular dynamics, seismic processing for energy discovery, image processing of course.
**Jensen Huang:** 所有那些通用计算效率太低的领域——是的,如果没有 AI,我会非常难过。但正因为我们在计算方面取得的进步,我们让深度学习实现了民主化。我们让任何研究人员、任何科学家、任何地方的任何学生都能够使用一台 PC 或一块 GeForce 显卡来做出了不起的科学研究。这个根本性的承诺一点都没有变。
**Jensen Huang:** And so all of those fields where general purpose computing is just simply too inefficient — yeah, if there was no AI, I would be very sad. But because of the advances that we made in computing, we democratized deep learning. We made it possible for any researcher, any scientist anywhere, any student to be able to access a PC or a GeForce graphics card and do amazing science. And that fundamental promise hasn't changed, not even a little bit.
**Jensen Huang:** 所以如果你看 GTC 大会,开头的一大段内容——完全不涉及 AI。那些关于计算光刻(computational lithography)、量子化学(quantum chemistry)、数据处理的工作——所有这些都跟 AI 无关,但仍然非常重要。我知道 AI 非常有趣、也相当令人兴奋。但有很多人在做很多非常重要的工作,那些工作跟 AI 无关,而且张量(tensor)也不是唯一的计算方式。
**Jensen Huang:** And so if you watch GTC, there's the whole beginning part of it — none of it's AI. That whole part with computational lithography or our quantum chemistry work or data processing work — all of that stuff is unrelated to AI and it's still very important. I mean, I know that AI is very interesting and quite exciting. But there's a lot of people doing a lot of very important work that's not AI related, and tensors is not the only way that you compute.
**Jensen Huang:** 我们想帮助每一个人。
**Jensen Huang:** And we want to help everybody.
**Dwarkesh Patel:** 确实如此。非常感谢你。
**Dwarkesh Patel:** It doesn't. Thank you so much.
**Jensen Huang:** 不客气。我很享受这次对话。我也是。很棒。
**Jensen Huang:** You're welcome. I enjoyed it. Me too. Sweet.