机器时代基金与新一代硬件基础设施投资

机器时代基金与新一代基础设施

我们迎来了一项全新的技术。这是有史以来最重要的技术。而你需要一整套全新的基础设施。

英文原文

We have a whole new technology. That's the most important technology ever. And you need a whole new infrastructure.

通常当我们谈论基础设施领域时，谈论的是服务器、存储和网络。在这里，它一直延伸到底层的铜矿。它的波及面将会有这么广。

英文原文

Normally when we talk about the infrastructure world, we're talking about the servers on the storage and the network. Here it goes all the way down to the mines of copper min. That's how widespread this thing is going to be.

过去当你构建某种东西时，它是一个工程问题。而在这里，感觉它确实是一个资源限制。所以不管它是不是 token，我们都在向系统投入大量的资金，然后这些系统正在产出结果。而目前我们的瓶颈在于，系统实际上无法匹配我们投入其中的资源的能力。

英文原文

It used to be when you built something, it was an engineering problem. And here it feels like it really is a resource limitation. So whether it's tokens or not, we're pouring a ton of money into systems and then those systems are producing a result. And right now we're bottlenecked on the systems ability to actually match the resources we're pouring into them.

领先的存储芯片供应商表示，他们目前的积压需求需要他们 3 年的产能来供给。

英文原文

The leading memory vendor said the demand they have today will take them 3 years of capacity to supply.

如果这只基金实现了我们认为它能做到的事情，在 5 到 10 年内我们会看到一个怎样的世界？

英文原文

If this fund does what we think it will do, how do we see the world in 5 to 10 years?

美国在基础设施方面赢得胜利 [音乐]，那将是极好的。

英文原文

America wins in the infrastructure [music] and that would be awesome.

Ben、Martin、Ragu，欢迎你们。

英文原文

Ben Martin Ragu, welcome.

谢谢。

英文原文

Thank you.

好的。谢谢。

英文原文

All right. Thank you.

我想先引用 Mark 的一段话来介绍这只新基金：“这是我一生中最大的技术革命。这显然比互联网还要大。能与之相提并论的是微处理器、蒸汽机和电力，或者可能还有轮子。”各位，“机器时代基金”。请介绍一下。Ben，给我们开个头吧。

英文原文

I want to start with a Mark quote to introduce this new fund. This is the biggest technological revolution of my lifetime. This is clearly bigger than the internet. The comps on this are the microprocessor, the steam engine, and electricity, or maybe the wheel. Guys, the machine age fund. Please introduce it. Ben, start us off.

嗯，基本上所发生的事情是，我们迎来了一项全新的技术，这是有史以来最重要的技术。每当有一种剧烈的新方式来使用我们热爱的基础设施的方方面面时，你都需要一套全新的基础设施，而且从来没有哪一次像这一次那样具有如此高的影响力。所以我们不仅需要新芯片、新系统软件，我们还需要全新的供电方式，我们需要替代铜。[笑声] 我是说，几乎需要改变一切。所以这是一个非常令人兴奋的时刻。因此，特别是对于这个新时代的硬件方面，我们需要一种新方法。

英文原文

Well, um, basically what's uh happened is we have a whole new technology that's the most important technology ever. And what happens every time um there's a dramatic new way of using all of the things that we love infrastructure um you need a whole new infrastructure and never has it been more high impact as it is on this one. So not only do we need new chips, new system software, we need new ways of doing power, we need to replace copper. I [laughter] mean like it's absolutely everything. So it's a very exciting time. So you know particularly for the kind of hardware aspects of of this new era um we needed a new approach.

是的，我同意。我是说通常至少在计算领域，当我们谈论基础设施世界时，我们谈论的是服务器、存储和网络，而在这里它一直延伸到底层的矿井——铜矿。它的影响面将会有这么广。

嗯，那是第一点。第二点，我认为过去 3 年我们看到的是模型能力的稳步提升，此时模型本身不再是瓶颈，事实上借助 AI，这些模型正在变得越来越好、越来越快。现在所有的瓶颈都在我所谓的“模型以南”，所以这就是为什么我们需要致力于这方面的工作。另外，我只想非常快速地补充一点，我们往往是追随创始人的，在过去几年里我们一直在观察的是，攻坚复杂硬件问题的顶级团队数量增加了。我不知道具体的确切数字，但我周末试着估算了一下。之前来自顶尖创始人的交易中，或许有 5% 会是硬件方向。现在我认为会在 20% 或 30% 以上。所以就像创始人社区——他们往往比 VC 社区要敏锐得多——已经将这确定为一个非常活跃的创新领域，并且他们正在做出响应。我觉得 5% 可能还是说多了。

英文原文

Yeah I would agree. I mean normally when we at least in the computing when we talk about the infrastructure world we talking about the servers and the storage in the network here it goes all the way down to the mines copper mines. That's how widespread this thing is going to be.

Um that's number one and number two I think what we have seen over the last 3 years is the steady increase of the capabilities of the models where the model is no longer the bottleneck and in fact using AI these models are getting better faster and faster and faster. Now the bottleneck is all what I call south of the model and so that's why we need to work on that. You know, the only thing I'd add very quickly is like we tend to follow founders and we've been watching over the last couple of years is the number of very strong teams going after complex hardware problems has increased. I don't know the actual numbers but I was trying to estimate it over the weekend. So I think we'd get you maybe 5% of the deals from top founders would come in would be hardwood before. Now I would say north of 20% or 30% right now. So like the founder community which tends to be much smarter than the VC community has identified this as a very active area for innovation and they're responding. I think 5% is probably generous.

是的，非常低。是非常低。是的。3%。是的。

英文原文

Yeah, it's very low. Is very low. Yeah. 3%. Yeah.

需求无限与硬件物理限制

能否解释一下导致这种转变的一些宏观条件，也就是追求这些想法的创始人过剩的现象？比如他们看到了什么促成了这一点？

英文原文

And explain some of the macro conditions that have led to this this trans this change in terms of the surplus of founders pursuing these idea. Like what are they seeing that's that's enabled?

嗯，显而易见的是，对 AI 的需求基本上是无限的，其结果就是供应链的每一个环节都承受着压力。我是说每一样东西，包括用来制造像存储芯片等产品的原材料。另外，AI 是否存在某种独特之处也非常耐人寻味。由于需求是无限的，增长是无限的，你往往担心的是公司的利润率，也就是它的效率如何。通常你会担心增长，比如我能不能让人们购买这些东西。在这里你不用担心这个。问题在于你能不能以盈利的方式来做这件事？而许多效率实际上纯粹受到硬件物理条件的限制。[哼鼻声] 因此，即使是 AI 浪潮的商业模式，也确实给现有系统带来了很大的压力，因为它们不是为 AI 构建的，不是为那些工作负载构建的。而且我认为大家普遍观察到，我们实际上需要改变核心组件来获得那种效率，以帮助推动增长并推动这些业务的价值。

英文原文

Well, I mean the obvious is like you know the the demand for AI is basically infinite and as a result of that every part of the supply chain is under under duress. I mean everything including like materials used to make things like like memory. Um it's also very interesting whether there's something unique about AI. Um which because the demand is infinite and growth is infinite. Um uh what you tend to worry about is the margin of companies which is how efficient it is. Like normally you worry about growth like can I just you know can I just get people to buy this stuff. You don't have to worry about that here. The question is is can you do this in a way that's profitable? And a lot of the um efficiencies are actually strictly [snorts] a physical limitation of hardware. And so even the business model of the AI wave is really putting a lot of stress on the existing systems because they weren't built for AI. They weren't built for those workloads. And I think there's just this you know this global observation that we actually need to change the core components to get that efficiency to help drive the growth and to drive the value of the businesses.

是的。那我们怎么知道这里的需求实际上是在超越供应，而不是又一次的炒作周期呢？嗯，[笑声] 我是说，

英文原文

Yeah. And how did [clears throat] we know that demand is actually outpacing supply here rather than this being you know another hype cycle? Well, [laughter] I mean,

是的，今天有无数个迹象。首先，一些对需求最具判断力的人正在开出巨额采购订单。我是说，如果你看看超大规模云厂商，对吧？

英文原文

yeah, there are any number of cities today. Um, firstly, it is that some of the smartest judges of demand are cutting huge purchase orders. I mean, if you look at the hyperscalers, right?

是的。

英文原文

Yeah.

他们的资本支出一直在暴增。明年据称几大超大规模云厂商的总和将达到 1 万亿美元。今年大约是 7000 亿美元，对吧？如果你想想超大规模云厂商在行业中的地位，他们看到来自各个方向的需求，对吧？他们显然看到了前沿实验室希望获得他们的算力，他们看到了 AI 原生公司，他们看到了企业客户，他们看到了美国本土以及国际市场。所以如果说有人具备能见度，那就是他们，而且他们一直在以前所未有的力度推高资本支出，对吧？所以这是一个非常非常清晰的信号。其次，如果你看看我们日常接触的公司，它们都在飞速增长。所有应用类公司，其增长都是疯狂的。前沿实验室的增长也是疯狂的。这都是有记录可查的。所以我想说在需求端，信号从未如此明确过——这不是炒作。更重要的是，这一切都在发生，并且价格还在上涨。

英文原文

Their capex spend has been exploding. Next year supposedly, it's going to reach a trillion dollars collectively across the big hyperscalers. This year, it's about 700 billion dollars, right? And if you think about the hyperscalers position in the industry, they see demand from everywhere, right? They see obviously the frontier labs um wanting their compute, they see the AI native companies, they see the enterprise, they see the US geography, the international geography. So if anybody has visibility, it is down and they've been jacking up their capex like it's never been seen before, right? So that's a clear clear sign. And secondly, if you look at the companies that we see on a day-to-day basis, they are all ripping. All the application companies, the growth is insane. The frontier labs, the growth is insane. It's been documented. So I would say on the demand side, the signals have never been clearer that this is not a hype. It's to top it all, all of it is just and prices are going up.

就像我们从未见过芯片价格上涨一样。

英文原文

Like we've never seen prices go like up on chips.

你一直看到价格下降。它们总是下降的。是的。它总是在下降。如果你看一下价格曲线，它之前是这样的，然后又变成了这样。

英文原文

You keep your prices went down. They always go down. Yeah. It always goes down. If you look at the price curve, it it went like this and then went back this.

是的。

英文原文

Yeah.

而且我们知道目前只满足了目标潜在市场的 5% 到 10%。

英文原文

And we know only like 5 10% of the address for the market to stop today.

我是说，供应方面，如果你全面看一下供应情况，基本上到 2028 年已经全部被订满了。严重到我们实际上已经看到为几千张 GPU 举行的持续数天的拍卖。当然，另一面就是需求。正如 Ragu 所说，我们看到了行业历史上增长最快的公司。

英文原文

I mean, the the the supply, if you look, if you look at the supply across the board, it's basically all booked out to 2028. I mean, it's so bad we've actually seen multi-day auctions for a few thousand GPUs. Um, you know, the other side of that, of course, is demand. And as Ragu said, we've seen the fastest growing companies we've seen in the history of the industry.

而且 AI 能够完成的工作单元，其工作单元的价值在持续增加。但在表象之下，所消耗的 token 数量正在呈数量级增长，对吧？如果对话是 1 个 token——我是说 100 个 token，那么智能体就是几千个 token，对吧？所以你在需求的两端都看到了扩张。一是工作单元对 token 的消耗变得越来越大，

英文原文

But also the unit of work that AI can do, the value of that unit of work keeps increasing. But underneath the covers, the number of tokens that are consumed is going by orders of magnitude, right? If it's one token for I mean 100 tokens for chat or an agent, it's thousands of tokens, right? So you got expansion of both sides of demand. One is the unit of work is becoming more and more consumptive

消耗 token，其次，因此将要受益的人数不仅是开发者，还将是所有知识工作者以及更广泛的群体，所以这就是我们所看到的。

英文原文

of tokens and then secondly the number of people therefore that are going to be benefit it's not just the developers it's going to be all knowledge workers and then all of beyond that so that's that's what we see

与互联网时代的对比

你提到关键供应组件已经售罄到 2027 年，甚至到 2028 年。整个行业被预订到那么久之后，这意味着什么？

英文原文

you said the key components in supply are sold out to 2027 maybe in 2028 what does it mean for an entire industry to be sold out that that far like

我不知道这在以前是否发生过。你们还记得吗？我是说，还记得在互联网时代，当我们进行大规模建设时，实际上铺设在地里的大部分都是投机性的，而且是闲置的。还记得暗光纤吗？而在这里，基本上生产出来的每一颗 GPU 都已经预售出去了。所以——

英文原文

I don't know this has ever happened before do you guys recall I mean remember in the Um, in in in the internet days when we were doing massive buildout, the majority that was actually being put in the ground was speculative and was dark. Remember the dark fiber? And here basically every GPU that's being created is already pre-sold. So

是的，我们当时并没有真正达到那种程度。我是说当时存在带宽不足的情况，比如在 98、99 年那个时间段，但是并没有那么多对带宽的真实需求，因为当时互联网上根本没有那么多人。所以那是一个双向的情况，公司都涌向那里，从理论上需要更多带宽，但在另一端却没有那么多的用户

英文原文

yeah, we weren't quite there. I mean there was there was a lack of bandwidth like in the 989 time frame, but there wasn't there there wasn't that much real demand for it because there just weren't that many people on the internet. So like it was a two-sided thing and the companies were all rushing there and needed more bandwidth theoretically, but there weren't the users on the other side

必然要去消耗它。而且，要真正消耗大量的带宽，你就得做高带宽的事情，比如视频，而当时视频因为许多与数据中心内部带宽完全无关的原因，根本不可行。所以表面看起来有些相似，但当时并不是现在这样。现在的情况是，我们全力以赴、满负荷运转，人们转手卖出 GPU 的价格还是他们买入时的四倍，类似这样的情况。完全不是一回事。然后，

英文原文

to consume it necessarily. And then to really consume a lot of bandwidth, you have to do high bandwidth things like video, which weren't really viable for, you know, a number of reasons that that had nothing to do with how much bandwidth was in the data center. So it smelled similar, but it wasn't this. This is like we're flat out and people are reselling GPUs for four times what they bought them for and this kind of thing. Like it's just not. And then,

你知道，我们电力和散热也已经耗尽了。而且在此之上，建设起来非常困难，因为面临着巨大的政治阻力。所以，在我整个职业生涯中，我们从未遇到过任何类似的事情，这真的是前所未有的。不，

英文原文

>> you know, we're we're also out of power and cooling. Uh and then on top of that, it's really hard to build because there's these incredible political headwinds going into it. So, it's it's really unprecedented in my career. uh that we've had anything anything like this. No,

我想给你讲个小趣事。我当时正在和一家大公司的首席财务官（CFO）交谈，一家大型上市公司，他们历史上一直对上云非常抗拒。所以，他们有很多服务器，在做库存盘点时，他们发现自己服务器里的内存价值涨了这么多，甚至足以资助整个上云迁移项目。所以我只是觉得，我们处在一个非常不寻常的境况中。

英文原文

>> I want to give you a quick anecdote. So, this I was talking to a CFO of a large company, large public company who had historically been very resistant about going into the cloud. So, they had a lot of servers and they were doing an inventory uh check and they realized that the memory in their servers had increased so much it could uh fund the entire migration to the cloud. So, I just feel like we're in a very unusual situation.

是的，[笑] 没错。我们在很多东西上都短缺了。电力、散热、内存、GPU，凡是你能叫得出名字的，我们都短缺。

英文原文

>> Yeah, [laughter] that's right. We're out of many things. power, cooling, memory, GPUs, like you name it, we're out of it.

供应链瓶颈与基础设施投资

是的。该行业的旗舰级会议——在斯坦福大学举办的 Hot Chips 正在进行中，领先的内存供应商表示，他们目前面临的需求，需要消耗他们 3 年的产能才能供应。这仅仅是今天的需求，甚至都还没算未来的需求。

那么，既然一切都在同时短缺，是因为人们低估了模型会变得多么出色、多么有用，完全无法预见这股需求吗？

英文原文

>> Yeah. So, the flagship uh conference for the industry is the one called hot chips is going on in Stanford and the leading memory vendor said the demand they have today it will take them 3 years of capacity to supply it. It's just today it's not even future demand.

So in terms of being about everything simultaneously, is it because people just underestimated how good the models would be, how how useful they would be, they just couldn't have foreseen the demand?

嗯，我认为甚至都不能归结于此。我的意思是，这东西完全是凭空冒出来的，对吧？我们进入这一领域才只有四年。所以即使我们有完美的先知，一旦它开始起作用，我认为

英文原文

>> Well, I don't even think it's that. I mean, this stuff came out of nowhere, right? We we're only four years into this. So even if we had a perfect oracle, once it started working, I don't think

我们也没法把产能建设起来。

英文原文

>> we could have built the capacity.

我们不可能建得出来那些产能。绝无可能。我们讨论的是芯片周期，通常要三到四年。我们讨论的是破土动工并建造数据中心，那需要四到五年，我们讨论的是

英文原文

>> We could have built the capacity. There's no way. And we're talking about like chip cycles, which tend to be three to four years. We're talking about breaking ground and building data centers which is you know four to five years we're talking

以及接入、破土动工、建造它们，还要有电力来源。所以你必须要么自己建电厂，要么通常两者兼备：你必须自己建设电力设施，并且拥有电力来源，这并不容易。是的。机器学习行业习惯了如果能增长 20% 到 30%，那就是很棒的增长率了，对吧？而它现在被连接到了一个三位数增长的 AI 软件行业，底数就是……你知道的，

英文原文

>> and connecting breaking down and building them and having uh power source. So like you either have to build your own power or like usually both you've got to build your own power and have a power source which is not easy. Yeah. the ML industry that's used if it's growing at 20 30% it's a great growth rate right and it's being connected to an AI software industry that's like triple digits is the base you know

所以你能看出其中的脱节，对吧，所以这完全是个巨大差距

英文原文

>> so you can see the disconnect right so it's it's just wide bank

那么，为什么这支基金没有在五年前或七年前成立呢？或者说，为什么在过去它不是一个适合以同样方式投资的优质品类呢，布拉德（Brad）？

英文原文

>> and so why didn't this fund exist you know five years ago or seven years ago or why was it not a great category to to invest in in in the same way Brad

嗯，我想说，我们可能——我倾向于认为恰逢其时，但你知道，我们可能在至少几年前成立会更合适。

英文原文

>> well I would say we're probably I'd like to think for just in time, but you know, we probably would have been uh well suited to have it at least a couple years ago.

我想说，实际上你几乎可以在每个时代都指出一家崛起的新兴独立公司，对吧？显然，从大型机向客户端-服务器架构的演进中，我们看到了一批公司涌现出来。向互联网的演进中，我们迎来了 Cisco 和 Juniper。甚至在超大型数据中心时代——顺便说一句，这主要是由现有的云服务提供商垂直整合推动的——你也看到了 Arista 的崛起。因此，一直以来都有投资芯片和硬件的机会，但规模相对较小，因为变革的幅度相对较小，比如出现了一家芯片公司、一家交换机公司。而现在，一切都在发生变化。所以我和本（Ben）看法一致，我们可能，你知道，我们或许本可以稍早一点起步，但现在的变革规模如此之大，以至于它已经成为一件显而易见的事情。

而且另一件事是，对智能的需求是如此垂直向上，并且完全看不到尽头。我的意思是，因为每一家采纳了它的公司，其使用量都在飞速增长；而绝大多数公司还没有高程度地采纳它；消费者端也才刚刚起步。

因此，对 token 的需求可能每年会增长接近 1000%，而你的供给不可能增长得那么快。我们根本不可能……

英文原文

>> I will say you could actually point on on on basically every epoch to uh an independent company that came up, right? Clearly the move from the client from uh in mainframe to the client server, we saw a bunch of companies come up. Um the move to the internet, this we got Cisco and Juniper. uh even in the mega data centers which by the way was largely driven by the incumbent cloud providers verticalizing you saw the uh arising of Arista so there has been the ability to invest in you know silicon and hardware but it's been relatively minor because the change has been relatively minor like one chip company one switch company where here everything is and so I I agree with Ben we're probably you know we probably could have started a little bit earlier um but the amount of change is so high now that it's just an obvious thing

And the other thing is the demand for intelligence is so vertical um with really no end in sight. I mean cuz every company that's adopted it is growing very fast in its usage and then most companies haven't adopted it to a high degree and then consumers are just getting started.

And so it's going to probably the demand for tokens is probably going to grow close to a,000% a year, which you cannot grow supply that fast. Like like we're not

就像我们要全方位付出的工作量，才能达到能够让基础设施以那种速率增长的地步，这是相当庞大的。所以我认为在这一过程中有大量的投资机会。顺便提一下，另一件事是，所有硬件系统的架构都是为了一个完全不同的计算时代而构建的。因此，不仅仅是我们单纯需要更多产能，我们需要建设……这里面有大量的机会来构建不同种类的基础设施。是的，它们都在触及它们当初设计时的物理极限，对吧？就像本刚刚谈到的铜线等等。你可以去逐一查看这些门类中的每一项，你都能发现，好的，这就是这种技术的极限。所以现在你必须取得一些技术突破才能跨入下一个阶段。

英文原文

>> like the amount of just work we're going to have to do across the board to get to the point where we can grow like infr at that kind of rate is is pretty vast. So I think there's a lot of investing opportunity on the way. And by the way, the other thing is like all the architectures of the hardware systems were built for a whole different era of computing. And so more than just we need more capacity, we need capacity to build there there's lots of opportunities to build different kinds of infrastructure. Yeah, they're they're all reaching their the physics limits for what they were designed, right? Like what Ben was talking about copper and so on and so forth. And you could go across every one of these categories and you could find, okay, this is the limit of this type of technology. So now you got to get some technical breakthroughs to get to the next one.

推理、Token 增长与自我催化

是的。我想在需求端深入探讨片刻。随着我们从聊天机器人演进到推理，再到智能体（Agent），再到多智能体，每一步都使单个任务所消耗的 token 数量呈数量级甚至更高的数量级增长。

英文原文

>> Yeah. I want to dive deeper on the demand side for a second. As we've moved from chat bots to reasoning to agents to to multi- aents, each step has multiplied the number of tokens a single task takes up by orders of increasing orders.

是的，没有任何人比 AI 更喜欢使用 AI 了。[笑]

英文原文

>> Yeah, nobody likes to use AI more than AI. [laughter]

那么，为什么这种情况会持续发生，而不是趋于平稳？你认为这种情况会……你知道，会无限期地继续下去，就这么持续……

英文原文

>> So, um why does that keep happening instead of a leveling off? Do you just see that happening you know indefinitely just continuing to

嗯，这包含几个方面。首先，毫无疑问，目前如果你看一下我们实现扩展（scaling）的方式，我们是通过大量的推理来实现的，也就是通过大量的 token。对吧，如果你思考什么是强化学习（RL），你知道，它就是大量的推理；如果你思考思维链（Chain of Thought），它也是大量的推理；你思考长程运行的智能体，那当然更是大量的推理。因此，这基本上就是我们一直用来实现扩展的方法之一。

我认为，如果你退一步来看，思考这里的宏观趋势是什么：过去当你构建某个东西时，它是一个工程问题，你投入一堆工程师来解决，但这无法扩展，而且会遇到工程物理的自然法则——这也正是《人月神话》（The Mythical Man-Month）的由来。而在这里，感觉它确实是一个资源限制问题。因此，不论是不是 token，我们都在向系统中倾注大量的资金，然后这些系统产生结果。而现在，我们的瓶颈就在于这些系统实际匹配我们所倾注资源的能力。所以我认为，目前 token 大概就是我们在扩展曲线上所处的位置，但我们没有像以前那样拥有工程方面的自然调节机制。因此我认为我们应该预料到这种情况会持续下去，并且我们必须构建供应来支持它。

英文原文

>> uh well so there's a couple as that the first one is is for sure right now if you look at like the the way we're achieving scaling the way we're doing it is through a lot of inference so through a lot of token right if you think about what RL is you know it's it's a lot of inference if you think about chain of thought it's a lot of inference u think longunning agents of course it's a lot of inference and so that's just basically been one of the approaches that we've been using to um uh to scaling um I think if you want to step back and say kind of what is the macro trend here it used to be when you built something it was an engineering problem and you throw a bunch of engineers at it and that doesn't scale and that would have a natural law of engineering physics uh which is what the mythical manmouth came from and here it feels like it really is a resource limitation so whether it's tokens or not we're pouring a ton of money into systems and then those systems are producing a result and right now we're bottlenecked on those position those systems ability to actually uh match the resources we're pouring into them. And so I think like tokens right now is probably where we are on the scaling curve, but we don't have a natural regulator like engineering like we did before. So I think we should expect this to continue and we have to build a supply to support it.

是的。简单的理解方式就是：你拥有的任何问题，只要有足够的基础设施都可以解决，[笑]

英文原文

>> Yeah. Like the simple way to think about it is any problem that you have can be solved with enough infrastructure [laughter]

基本上就是

英文原文

>> basically

GPU、电力和资金。因此，在我们的问题耗尽之前，我们的需求不会耗尽。这就是……这就是挑战所在。我

英文原文

>> GPUs and power and money. Uh and so until we run out of problems, we're not going to run out of demand. And that's the that's the uh challenge. I

意思是，AI 变得越来越强大的应对方式就是使用更多的 AI，对吧？推理是它一遍又一遍不断使用的一个基础构建模块，所以这也是为什么这些 token 在每一步都会成倍增加。

英文原文

>> mean AI's answer to getting better and better is to use more AI, right? inference is one basic building block that it keeps using over and over and over again and so that's why these tokens multiply at each step.

是的，甚至是自我催化效应。因此，哪怕是用 AI 来创造更多 AI 的想法——比如编写 GPU 内核，当然也只是在流程中使用更多的 AI。所以再一次地，我们的一种思考方式是：在过去，资金投进来，你面对的是一个工程问题，我们知道这需要两年时间，通常会失败，你知道，这是一个天然的调节阀，然后你在另一端得到产品。而这一次，在投入资金与硬件创造智能之间，没有任何阻隔。因此现在我们仅仅受限于我们创造供给的能力。这是一个非常非常不同的

英文原文

>> Yeah, even the autoc catalytic effect so even the idea of using AI to create more AI like creating a GPU kernel of course is just using more AI um as part of the process. So again, one way that we think about it is in the past money would come in, you have an engineering problem, we know that it takes two years, normally fails, you know, it's a national governor and then you get the product on the other end. This there's there's there's nothing between the money going in and then the hardware, you know, creating intelligence. And so now we're just limited by our ability to create supply. It's a very very different

更多的 GPU 和我们……

英文原文

>> be more GPUs and our sol.

是的，没错。

英文原文

>> Yeah, that's right.

所以这就是循环。只要

英文原文

>> So that's the cycle. as long

只要你有资金、GPU 和数据，在可预见的未来，你就能不断扩展这些模型。

英文原文

as you have the money, the GPUs, and the data, you know, for the foreseeable future, you'll be able to scale these things.

这太令人着迷了，因为过去十年里，很多人普遍抱有的看法是，流入初创公司的资金太多了。我们过度资助了这些初创企业，风险投资领域的资金太多了。Ben，请多谈谈这意味着什么，因为过去曾存在这样一种怀疑，即你向这个行业投入的资金越多，产出的成果是否就越大；而现在，我们在战略研讨会上提到，在某种程度上，这个市场的大小取决于我们共同投入的规模。

英文原文

>> And it's it's fascinating because, you know, over the last decade, it feels like there so many, you know, people and the pervasive sentiment was there's too much money going to startups. We're overfunding these startups. There's too much money in in in in venture capital. Say more Ben about what that means because there there used to be this um sort of skepticism that the more money you put into into the industry that you know there would be bigger outcomes and now you know we were saying at the offsite that there's to some degree the the market is as big as we collectively contribute to it.

资本密集型扩张与算力需求

是的。听着，在初创公司领域，我们大家都明白一件事：如果我领先你两年，而你想通过招聘一千名工程师来追赶我，你就会毁掉你的公司。这种做法从来行不通。这就是人月神话，九个女人不可能在一个月内生出一个孩子。就是这样，那从来都没用。

英文原文

>> Yeah. So, this is this look, the one thing we all knew with in startup world is that if I have a two-year lead on you and you try and catch me by hiring a thousand engineers, you're going to wreck your company. Like, that never works. It's a mythical man month. Nine women can't have a baby in a month. That like that's it. Like, that never works.

好吧，现在这种做法行得通了。[笑声] 但这不是雇佣十万名工程师，而是拿出 30 亿美元，点亮一个极其宏大的集群，然后突然之间，无论是 Grok 还是 Kimi 等等，都可以不知从哪冒出来，一下子就成了气候。

英文原文

Okay, now that works. [laughter] But it's not hiring a hundred thousand engineers. It's taking $3 billion and like lighting up a magnificent cluster and then all of a sudden, you know, whatever Grock can come out of nowhere and like, oh, all of a sudden it's real or or a Kimmy or or what have you.

这些领先优势……你完全可以砸钱解决问题，你几乎可以把钱砸向任何问题，而且这招奏效。这与我们经历过的任何事情都截然不同。顺便说一句，我们所有人都在心理上适应这一点。ChatGPT 应用拥有十亿周活跃用户。大约有 3000 万开发者，消耗了相当大比例的算力需求。考虑到人们实际使用 AI 的方式，我们现在以及未来应该如何看待算力需求？

英文原文

It's just like these leads um you can throw money at the problem and you can throw money at almost any problem and that works. And so that is just completely different than anything we've ever lived through. So we're, by the way, we're all psychologically adjusting to this. The Chad GBT app has a billion weekly activives. There's about 30 million uh developers um who are using, you know, relatively a big portion of compute demands. How do we think about compute compute demands needs now and in the future in light of what people are actually doing with AI?

这就是它的演进路径，对吧？ChatGPT 最初是一个休闲应用，而现在它是专业人士的代码工具。通过编程，人们为知识工作者构建了惊艳的工具，那是下一个前沿；现在排着队等待的知识工作者超过十亿人，对吧？随之而来的这种需求还有很长的路要走。顺便说一句，包括他们从事的工作，所有的工作流和自动化等等，然后进入后勤部门，那就是各种 Agent。因此，按部就班地，其中每一项都会释放出——我会说是几个数量级的更大需求。我是说，这仅仅是个开始。

英文原文

>> That's the progression, right? So Chadic was a casual app and it's coding for professionals right now using coding you have now built amazing tools for knowledge workers so that's the next frontier and now there are over a billion knowledge workers in the line right and with that it's a long ways to go for that demand and by the way the work that they do all this workflow and automation and so on and then you get to the back office which is all the agents So progressively each of these things unlocks uh I would say order of magnitude more demand. I mean just at the start of this

而且现在有了 Grokbot，就像编程领域发生的事情一样，通过电脑操作，各种计算机使用也正在发生同样的情况，因此我们正迎来另一波全新的需求浪潮，而且毫无疑问未来还会更多。目前来看这似乎确实是没有上限的，我们甚至还没有谈到具身智能或机器人，它们将成为另一个需求来源。

英文原文

>> well and now you have Grockbot which is kind of uh you know what uh happened with coding is kind of happening with all use of computer viaot and so we're in a whole another wave of demand and most certainly there's going to be more to come. Uh so it does seem quite unlimited at the moment and we haven't even gotten into embodied AI or robots uh which are going to be another source of demand.

Martinez 是专家，但我的理解是，那部分使用的是计算机操控（computer use），就像是……

英文原文

>> Martinez is expert but my understanding is that part uses computer use which is just like

就像一个人坐在电脑里敲键盘。

英文原文

>> human being sitting inside the computer typing away.

我周末确实用它在一堆我一直懒得处理的服务上更新了信用卡信息，并取消了一堆订阅。我的意思是，这不是写代码之类的，这是真正的操控电脑。

英文原文

>> I literally used it over the weekend to to update my credit card with a bunch of services that I'd been like lazy to do and cancel a bunch of subscriptions. I mean, this is not coding or whatever. This is true computer use.

突然之间，你就像创造了五亿名知识工作者，只不过他们全都坐在电脑里面[笑声]在干活？

英文原文

>> All of a sudden, you're creating like half a billion knowledge workers except they're all sitting inside of the computer [laughter] doing what?

类比蒸汽机：长期算力增长与新型数字员工

我确实认为 Marc Andreessen 是对的。恰当的类比其实就像蒸汽机或电力，原因在于：我们引入了这种可以转化为工作的新事物，现在有一些非常显而易见的应用，但在未来三四十年里，可能会持续将计算能力投入到各种问题中。任何具有明确奖励信号的事物——而我们现在才刚刚起步，我们掌握了语言和代码，仅此而已，而且才刚刚开始涉足计算机操控。除此以外我们还在关注什么？从科学、材料到生物学，当然创意也是一个巨大的应用方向。听着，我们在一条漫长旅程的非常早期的阶段，而且我们移除了关键瓶颈，也就是传统软件工程。当然，瓶颈会转移，其他地方会出现更多复杂性，但我认为在用计算解决问题的长期历程中，我们仍处于非常早期的阶段。因此，可以预见这种算力需求将持续数十年。

英文原文

>> I I do I do think that that that Mark Mark Andre is right. It's like the right analog here is like the steam engine or electricity in the following way. Like we've we've introduced this new thing that you can turn to work and there are some very obvious applications now, but there's probably 30 40 years of throwing computer at problems. anything with a clear reward signal and we we're just starting like we've got language and code that's it and just starting computer use but like what else are we looking at we're looking at uh in terms of science materials biology I mean of course creativity is a massive use and so listen we're at the very very early part of a very long journey and we've removed this key bottleneck which is you know traditional software engineering now of course you know bottlenecks will move and there'll be kind of more complexity elsewhere but I think we're at a very early in a very long run of throwing computer problem. So let's expect you know this compute need to persist for decades.

不过既然提到了，Marty，谈谈 Grokbot 吧。我们在战略研讨会上聊过让你感到触动的地方。显然我们以各种可能的方式参与其中，但你觉得它最有意思的地方是什么？

英文原文

>> But because we mentioned it u Marty talk about grabbot um because we um we're talking at the offsite about how you know what what struck you about it. Obviously we're involved in every possible way you could you could be involved but what um yeah what what did you find so interesting about it?

我认为整个行业对于 AI 如何融入我们的生活，经历过几次认知的转变，对吧？最初，大家觉得，你在产品里加入 AI，就像加了一个搜索栏，然后你跟它聊天，它回复你，因为那是传统的做法。接着 OpenClaw 之类的出现了，那是今年早些时候的事。面对这类事物，大家会说：“好吧，也许它不仅仅是像更好用的谷歌。也许那并不是它的完整形态。不如让它成为一个独立的存在，但作为你的延伸，它共享你的密钥，知道你的密码，替你做你会做的事情，对吧？所以它算是你的延伸，但更像是一个人，一个属于你的延伸。”

英文原文

So I think we've I think we've as an industry gone through kind of multiple realizations for how AI enters our lives, right? And and uh very early on we're like okay well you add AI to a product and it's like whatever it's like a search bar and then you kind of you know you you do chat with it and it chats back because that's kind of the traditional way to do it. Um uh and then Open Clock kind of showed up and that was it earlier in the year. And with Open I say, "Okay, well maybe like it just being like Google but better. Maybe that's not the full embodiment of it. How about we'll have it be a standalone thing but it'll be an extension of you and it'll share your keys and it'll know your passwords and it'll just kind of do stuff that you would do, right? So it's kind of an extension of you but it's more like a human, an extension of you."

然后，[哧哧笑] 我认为 Grokbot 真正做对的是：不，它其实是一个员工。现在你拥有了这个实体，它不需要对你的密钥等具有特殊访问权限。它有自己的电脑，有自己的浏览器，而且因为这些是世界上最聪明的模型，它能做任何员工能做的事。这很有意思，因为现在如果我想做某件事，我的第一反应是：“它能帮我做吗？”而且答案往往是肯定的，即使是你意料之外的事情。显而易见的用例比如管理日程、预订会议，但也有不那么显而易见的。比如，我会让它通读我的电子邮件并进行分拣分类。我并没有教它具体怎么做，但它在真正执行分拣前会知道先向我核实。这些东西已经足够精细，你可以给出一个相对高层级的任务，它就会因此去执行复杂的事务。

英文原文

And then [snorts] what I think Rockbot got really right is no, how about it's actually an employee. So now you have this thing that's an entity and it doesn't have like special access to your keys or whatever. It has its own computer and it has its own browser and because these are the smartest models in the world, it can do whatever an employee can do. And it's kind of interesting because now actually if I if I want something done my first thing I I think is like well can do it for me and and often the answer is yes even if it's something you wouldn't you know expect it. So, the obvious ones are like whatever. It'll like manage my calendar. It'll like book a meeting, but there's also non-obvious ones as well. Like, so, for example, I'll have it um uh read through my email and do triage. And and I did I don't tell it how to do that, but it will know to check with me before actually doing the triage. So, like these things are sophisticated enough that you can give a relatively high level task and it'll do kind of, you know, like sophisticated things as a result.

Ben，我知道你对这方面思考了很多，包括它在组织中如何运作。当然，你对文化思考很多。你对此有什么想法？

英文原文

>> Ben, I know you're thinking a lot a lot about this and how this you know, works in the organization. you think a lot about culture of course. What are your thoughts here?

好吧，如果看看我们自己，这就像是拥有了一种新型员工，而且会有很多这样的员工。我们花了非常非常多的年头来摸索如何与普通的人类员工共事，而现在我们有了这些其他类型的员工。它们也存在学习曲线。它们可能会消耗大量 token、花掉很多钱，却没做出任何有成效的事情。它们会遗忘东西，会胡编乱造。它们可以表现良好，也可能会表现不佳，它们……

英文原文

Well, I mean, I think if you just look at us, um, you know, it's like having a new kind of employee and there's going to be a lot of them and we have to just like, you know, with our we spent many, many, many years figuring out how to work with our kind of regular human employees and now we've got these other kinds of employees and, you know, there is a learning curve with them. So um they can burn a lot of tokens and spend a lot of money and get nothing productive done. Um they can forget stuff. They can make stuff up. Um you know they can have good behavior. They can have bad behavior. They can

它们会带来安全问题。

英文原文

>> they can create security problems.

所以方方面面的问题都存在，但它们也可以拥有超级极高的生产力。因此，如何将它们整合进来，让它们与正在共事的人员、与真实的人类良好协作，这些都是我们正在学习如何去做的事情。我的意思是，我不想坐在这里说我已经破解了密码，我们已经完全搞定了，[笑声] 整个公司现在完全自动化了，而且因为我有能力，我打算慢慢淘汰所有人类。我们完全不是处于那种状态。我们现在的状态更像是：好吧，我们如何让人类都成为超人，同时又不会因为机器人失控而把整个地方搞得一团糟

英文原文

So like there there there's all those aspects to it, but like they can also be like super duper productive. And so I think figuring out how to integrate them in, have them work nicely with the people that they're working with, um, the actual humans, uh, is all something that we're learning how to do. I mean, I I don't want to sit up here and say I've cracked the code. We've got this more of a sloop and [laughter] the whole firm is just completely automated now, and I'm going to slowly get rid of all the humans because I can. Like, that's not at all where we are. were much more going like, okay, how do we make all our humans superhuman um without like wrecking the place um because the bots got out

失控？

英文原文

of control?

对。而且这很有意思，因为我们尝试过几种不同的方式来让智能体以最佳方式进入系统，如果你愿意这么说的话；最终是马丁的洞见：就直接把它们当成人来对待，把事情做成就行。而这就是我们正在做的事情。这已经被证明是在组织内部把这件事推行下去最持久有效的方式。

英文原文

>> Yeah. And it's interesting because we've done we've tried a couple of different ways as how best to get agents into the system if you will and eventually it was Martin's insights just treat them as people and get it done. And that's what we're doing. that's turned out to be the most durable way of getting this thing going inside of an organization.

AI 时代基础设施与供应链瓶颈

我想回到供应端，更深入地探讨一下瓶颈问题。我们之前谈到，在数据中心、芯片架构、系统软件、设施本身等方面，没有一个是专门针对 AI 进行设计的。如果它们是针对 AI 设计的，会是什么样子？思考这可能意味着什么的思维模型是什么？

英文原文

>> I want to go back to the supply side and go deeper into the the bottlenecks. You we were talking about how you know in terms of the data centers, the chip architecture, system software, facilities themselves that none of them were designed with with AI in mind. What would it look like for them to be designed with AI? Like what is sort of the mental model for thinking about what that could mean?

好的。所以，我的意思是，如果你从这样一个论断开始：即任何这些模型上的原有基础设施模式都必须改变，你可以逐列逐项去看它是如何卡住的，对吧，然后你开始去解开其中每一项的瓶颈。所以最终你必须达到这样一个系统：如果你看看推理引擎是做什么的，对吧？它占用大量内存，它在计算的同时生成新的 token。因此你可以想想，我该如何优化所有这一切？内存需要做成什么样？算力需要做成什么样？它们之间需要如何进行通信？它们各自需要多少电力？如果它们需要所有这些电力，你又该如何为它们各自进行冷却？对吧？然后你该如何把这些东西的集合体组合在一起？

这就是行业中许多创始人目前正在进行的工作。所以他们把问题分解为其基本组成部分，并询问正在完成的计算的确切本质是什么。好的，它将如何变成矩阵乘法。我该如何围绕这种场景优化我的算力，然后它们都需要逐步去生成这些 token？分层排列这些内存的最佳方式是什么，对吧，然后电力消耗如何，我的意思是，接下来你必须将它连接在一起，在同一芯片上、跨芯片以及跨数据中心有哪些连接方式，每一次这些数据传输需要消耗多少电力？所以你必须逐步将它彻底拆解开来，并从这些基本构件重新构建它，这就是我们看到正在进行的事情，也是我们看到机会的地方。

英文原文

>> Yeah. So um I mean if you start with a statement that you just say hey original model um of infrastructure on any of these models has to change you can go se category by column and see how it bricks right and then you start uh unlocking the bottlenecks in each one of these things. So eventually you have to get to a system where if you look at what an inference engine does, right? It takes up a lot of memory, it generates new tokens along with the compute. And so you can just think about how do I optimize all of this? What does the memory need to be? What does the compute need to be? How do they need to talk to each other? How much power does each of them need? And if they need all of these power, how do you cool each of these? Right? And then how do you put the collections of these things together? That is the exercise that's underway in the industry right now with a lot of the founders. So they're breaking down the problem into its fundamental components and saying what is the exact nature of the compute that's getting done. Okay, it's how it's going to be matrix multiplications. How do I optimize my compute around that kind of a scenario and then they all need very progressively to generate these tokens? What is the best way of hierarchically arranging this memory right and then how does the power consume I mean then you got to connect it together what are the ways of connecting it on the same chip but across chips and across data centers how much power does each of these data transmission take so you have to progressively break it all down and rebuild it from these fundamental building blocks and that's what we see underway and that's where we see the opportunity

让我给你一个有趣的思维模型，来思考行业格局发生了怎样的变化。今天构建一个前沿模型，成本假设是 30 亿到 50 亿美元，对吧，这还只是训练成本。因此推理必须至少收回那个成本，当然，为了让这一切具有可行性，比方说两倍于该成本，那么假设现在推理必须赚取 100 亿美元。如果能在其中节省 20% 的效率，那就是 20 亿美元，而用 20 亿美元你完全可以轻松制造一款 ASIC。对。所以，我们实际上已经到了行业中这样一个有趣的节点：为每个模型单独制造一款 ASIC 实际上是合乎逻辑的，仅仅是因为投入到该模型中的资本投资规模。再者，与传统软件不同——传统软件包含大量状态，非常动态——这些模型是固定不变的，模型权重是固定的。因此，我们不知道世界是否会走向每模型一款 ASIC 的方向。但这提供了一个极佳的思维模型，展示了你会如何演进该架构，使其针对我们正在进行的这些庞大资本投资变得更加定制化。我认为在行业历史上，我们从未创造过任何耗费了约 50 亿美元直接投入其中的数字产物。因此，我认为这将对硬件提出前所未有的最高要求。顺便说一句，机架功率需求正从大约 5 到 10 千瓦走向 100 到 150 千瓦。计算密度正在攀升大约 70 倍。

英文原文

>> let me give you an interesting mental model to think about how the landscape's changed so um uh so today to build a frontier model costs let's say $3 to5 billion right so and let's you know and and that's to train it and so the inference has to pay back at least that of course right you know in order for any of this stuff to be viable let's say two times that so let's say that now inferred has to to make $10 billion so if you can save 20% of efficiency on that that's $2 billion and you can easily build an ASIC for $2 billion Right. So, so we've actually gotten to this interesting point in the industry where it actually makes sense to build an ASIC per model just because the amount of capital investment in that model and then unlike traditional software, traditional software has a lot of state and a lot of you know is very dynamic. These models are fixed. The model weights are fixed. And so we don't know if the world goes to per model A6. But it gives you a great mental model of how you would evolve the architecture to be far more bespoke for these massive capital investments we're doing. Like I don't think in the history of the industry we've ever created a digital artifact with something like $5 billion that went directly into that artifact. And so you know like this I think is going to put the greatest demands on hardware that we've ever seen. Well, to that end, rack power requirements are moving from roughly 5 to 10 kilowatts to 100 to 50 kilowatts. Compute density is climbing something like 70x.

冷却正从风冷转向液冷作为硬性要求。由此产生的投资机会有哪些？

英文原文

>> Cooling is moving from air to liquid as a requirement. What are the investment opportunities as a result of this?

电力、冷却与数据中心物理重构

嗯，首先，当每个机架达到这种功率水平时，交流电（AC）已经不管用了。[笑声] 所以，这可以说是一件相当疯狂的事情。现在你就进入了直流电（DC），顺便提一下，直流电本身也需要专门的冷却。而且，顺便说一句，超级危险。这有点讽刺，因为当年爱迪生推广直流电时，曾声称交流电有多危险，并通过电击动物等方式来进行演示 [笑声]。

英文原文

>> Well, first of all, when you get to that level of power per rack, AC power doesn't work anymore. [laughter] So, like that's a pretty wild thing. Um, so now now you're into DC power, which by the way also requires its own cooling. Um, and is like, uh, by the way, super dangerous. Uh, which is kind of ironic because this was, um, Edison promoted DC power by claiming how dangerous AC power was in demonstrating it [laughter] by like electrocuting animals and things.

电击马，对。

英文原文

>> The horse. Yeah.

电击马，对。但他其实是对的，只不过是对他自己那种电力而言，好消息是它极其强大。所以，仅从电力开始，那将变得非常非常不同。

我认为在冷却方面，这涉及……是的，我们正在从风冷走向液冷。我认为对于任何最先进的数据中心，我们已经处于液冷阶段了，那基本上已经成定局了。

英文原文

>> The horse. Yeah. So but he was right but around his own kind of power which is extremely powerful is the good news. Um so you know just starting with power uh yeah that's going to be like very very different I think with cooling. So and this gets into so yes we're going air cooling to liquid cooling. I think we're already at liquid cooling for any state-of-the-art data center. Like that's already kind of a done thing.

但随之而来的问题是，鉴于政治环境等等，仅有液冷是不够的，它必须是环保型液冷。直流电也是不够的，它必须是能够为社会的电力做出贡献、而不是掠夺社会电力的电力。因此，现在有一些表现恶劣的数据中心，比例很小，大概只有 10%，浪费大量的水。虽然不像人们在互联网上论证的那样耗水量赶上开心果或杏仁等等，但它们本可以在这方面大幅提高效率。另外还有一些就像电力的寄生虫，不向电网回馈电力。我认为所有这些都将结束，它必须结束，仅仅是因为我们在这方面已经走过了一扇单向门。所以这需要一种许多人尚未投入的工程水平，因此这即将到来。

再者，如果机架密度达到那种程度，地面设计的方式等其他方面也必须能够承受那样的重量。这种事情实际上是实实在在的。另外，我认为就是几乎每样东西都需要大量的量，而且设备运行声音非常大，所以必须用更厚的墙来建造数据中心，否则就会扰乱周边的安宁，这是不可接受的，我认为没有任何一个州会允许那样。

英文原文

>> But it gets into okay, you know, given the political environment and so forth, you like liquid cooling isn't enough. It's got to be eco-friendly liquid cooling. Um, and you know, kind of DC power is not enough. It's got to be um power that contributes uh to the power of society, not takes away from it. And so you have data centers who have been behaving badly. Um small percentage actually probably 10% wasting a lot of water. Um you know not as much as pistachios or almonds and so forth as people demonstrate on the internet but like they could be a lot more efficient with that. Uh and then there are ones that you know kind of are uh parasites of power and don't contribute power back. I think all that's going to end. it's going to have to end just because like like that we we've kind of gone through a one-way door on that. Uh so that requires like a level of engineering um that you know many haven't invested in yet. So that's coming. Uh and then you know like if racks are that dense um there are other things that like the way the floors are designed have to support that kind of weight. uh you know that kind of thing is is actually for real. Um and then you know I I think that there's you know you just need a lot of everything and so there's going to be you know kind also by the things are really loud so you have to build the data center with thicker walls or you're going to disturb the peace in the neighborhood which is not going to be acceptable like I don't think any state's going to allow that.

因此，人们过去构建和设计建筑本身的许多方式，现在已经完全过时了。一旦我们到了 Fineman 时代，我们今天拥有的数据中心中能够继续适用的比例要小得多。事实上，大家都在谈论内存价格，但涨价最快的领域之一正是他们所需的钢筋混凝土。

英文原文

>> And so a lot of the ways people have architected and designed the buildings themselves are already completely obsolete. Like once we get to Fineman, um a much smaller percentage of the data centers that we have today work. In fact, I mean everybody talks about memory prices, but one of the fastest areas where prices is increasing is reinforced concrete from for their sake.

当这些数据中心向机架输送 800 伏电压时，发生的另一件事是：第一，它变得极度危险；但第二，我们没有足够的电气承包商具备在数据中心内部应对 800 伏的专业知识，因为这是高压电。在美国，只有 2% 的电气工程师或电工获得了直流电操作认证，这可以让你对此有个概念。现在 Meta 推出了一个完整的项目来培训人员等等，这很好，就像是一个新的职业培训基地，免费培训人员从事这项工作。但你知道，有趣的是人们总说 AI 正在夺走所有工作，而 AI 实际上将创造大量的电工新岗位。

英文原文

>> The other thing that happens when these data centers are uh sending 800 volts to the whack is it's become so dangerous number one, but secondly, we don't have enough electrical contractors that have the expertise to deal with the 800 rules inside the data center. because this is high voltage. Only 2% of electrical engineer or uh electricians in the US have been certified on DC power. So like that gives you an idea. Now Meta's got a whole program to train people up and so forth which is great. It's like a new job court where they train people for free to do this job. But you know it's funny AI is taking all the jobs. AI is going to create a lot of new electricians.

对。我认为在这个领域我们也正在做一些事情。那些拥有大型云数据中心的大公司，他们全都在疯狂地试验机器人，对吧？用来从事在数据中心内部组装或装入服务器等工作。因此，随着 AI 的演进，你会看到这一现象越来越多。

英文原文

>> Yeah. I think we're just doing something in space too. the big guys that own the big cloud data big data centers, they all are furiously experimenting with the robots, right? To do the work of assembling or putting servers into the data center, etc. And so you will see that increasing as a result of the evolution in AI.

顺便说明一下，为了对我们正在募集的实际基金明确起见，我们的重点是在计算机科学基础设施上。因此是模型运行其上的任何属于计算机科学范畴的东西，对吧，比如芯片、网络互连、存储，一直往下可能延伸到电力……

英文原文

>> By the way, to be to be clear on the um on the actual fund that we're raising, our focus is on computer science infrastructure. So anything a model runs on that's computer science right so think you know chips network interconnect storage all the way down probably to the electricity

对，再多说

英文原文

>> yeah and and saying more

关于机器人领域，就我们要做的事情而言，与 American Dynamism 相比，或者该如何看待这一点？

英文原文

about

the

robotics

arm

terms

what

we'll

doing

versus

maybe

American

dynamism

how

think

about

that

对，对，当然。所以，嗯，你知道，我们再次认为，任何 AI 将在其上运行的平台——比如 AI 带来的巨大突破之一就是它允许计算机与物理世界互动，对吧？它能看、能听、能说，对吧？这意味着新的平台，对吧？人们最简单的说法是边缘设备，但这并没有什么实质意义，对吧？我是说，它可以是移动设备，可以是 CDN，可以是笔记本电脑，但也可以是一个四处活动的具身设备。所以再次强调，作为专注于基础设施的投资者，我们不做受严格监管的行业或更垂直化的行业，但任何将推动 AI 走向更边缘的计算机科学平台，我们都非常感兴趣。

英文原文

>> yeah

yeah

for

sure

you

know

again

think

that

any

platform

that

will

run

one

the

great

breakthroughs

that

does

allows

computers

interact

with

the

physical

world,

right?

can

see,

can

hear,

can

talk,

right?

And

this

means

new

platforms,

right?

And

the

simplest

way

people

say

edge

device,

but

that

doesn't

really

mean

anything,

right?

mean,

could

mobile

device,

could

CDN,

could

laptop,

but

also

could

embodied,

you

know,

device

that

goes

around.

And

again,

you

know,

infrastructure

focused

investors

don't

heavy

regulated

industries

verticalized

industries,

but

any

sort

computer

science

platform

that's

going

push

further

out,

we're

quite

interested

in.

数据中心与吉瓦级电力瓶颈

对。回到数据中心，到 2028 年，新建数据中心将需要大约 44 吉瓦（gawatts）的额外电力，而预计电网新增电力可能只有 25 吉瓦（gawatt）。

英文原文

>> Yeah.

Going

back

the

data

centers,

2028,

new

data

centers

are

going

need

something

gawatts

additional

power

against

maybe

gawatt

expected

grid

additions.

等等，等等。我们用吉瓦这个词。

英文原文

>> Hold

on,

hold

on.

use

that

word

gigawatt.

不，就像是说我们会有 100 吉瓦，[笑声]

英文原文

>> No,

it's

we'll

have

100

gigawatt,

[laughter]

Martine，1 吉瓦是什么概念？

英文原文

>> Martine.

What's

gigawatt?

我是说，我是说它有多大？相当于好几个足球场。我是说它太庞大了。是 5 万、5 万人。

英文原文

>> I

mean

how

big

it?

It's

multiple

football

fields.

mean

it's

massive.

It's

50,000

people.

你是什么意思？它到底是什么？

英文原文

>> What

you

mean?

What

it?

就像，就像

英文原文

>> Like

相当于大约 5 万户家庭。

英文原文

>> the

equivalent

50,000

houses.

城市

英文原文

>> City

5 万，5 万户家庭。5 万户家庭。我是在亚利桑那州的弗拉格斯塔夫长大的，[笑声] 这是一个有 4 万到 6 万人的城镇，取决于大学的情况。我们的耗电量还不到 1 吉瓦。我是说，这——

英文原文

>> 50,000

50,000

homes.

50,000

homes.

grew

Flagstaff,

Arizona,

[laughter]

which

town

60,000

people,

depending

the

universities.

have

less

than

gigawatt

power

consumption.

mean,

this

所以 1 吉瓦基本上可以为你整个城镇提供照明和空调。

英文原文

>> so

you

can

basically

light

and

air

condition

your

entire

town

for

gigawatt.

是的。我是说，这——

英文原文

>> Yeah.

mean,

this

只是随口到处乱提这个词。[笑声]

英文原文

>> just

throwing

them

around.

[laughter]

不，但顺便说一句，每个人都在谈论吉瓦。实际上建成的吉瓦级数据中心极少。我是说，我们还有很长的路要走，对吧？但那为什么公用事业公司和超大规模计算企业不能建得更快一些呢？

哦，这里面有太多事情了。首先，现在你需要人来建造它们。所以，这里面就存在常规的施工建设，但远不止这些，你需要许可证。

嗯，你需要接入可以插电的电源。所以你要么是两者结合——你必须获得电力接入，这是一种巨大的监管竞标争夺。就天然气电网等等而言，你能接入的量和途径都非常有限。嗯，但接下来你还得自己发电，猜猜怎么着？我们的变压器、涡轮机以及所有相关的设备都面临短缺。所以，就是，你知道，你必须把所有这些东西都弄到手。这不——这不是一个软件问题。这不仅仅是一群工程师能不能周末加班之类的搞定——这并不是那样就能奏效的，而且

这里面存在真正的瓶颈，这些交付周期不是那么容易压缩的。听着，我们有世界上最聪明的人试图弄清楚如何压缩它们。嗯，但这并不容易。这并不容易，而且需求并没有放缓。所以我们已经落后了。需求现在正以每年 10 倍的速度增长。而供应根本不可能增长得那么快。

英文原文

>> No,

but

the

way,

everybody

talks

about

the

gigawatt.

There's

very

few

gigawatt

data

centers

that

are

actually

up.

mean,

we've

got

long

way

go,

right?

But

then

why

can't

utilities

and

hyperscalers

just

build

faster?

Oh,

there's

many

things

there.

Well,

there's

first

all,

right

now,

you

need

humans

build

them.

So,

there

there's

just

the

regular

construction,

but

much

than

that,

you

need

permits.

Um,

you

need

access

power

that

you

can

plug

in.

you're

either

you're

doing

combination

you've

got

get

access

power

which

massive

kind

regulatory

bidding

struggle.

There's

very

limited

kind

amounts

and

things

you

can

tap

into

terms

natural

gas

power

grids,

what

have

you.

but

then

you

also

have

build

your

own

power

and

guess

what?

We've

got

shortages

transformers

and

turbines

and

everything

that

goes

into

that.

So,

it's

just,

you

know,

you've

got

get

all

that

stuff.

It's

not

this

not

software

problem.

It's

not

just

bunch

engineers

can't

you

work

weekends

and

that

type

stuff

this

not

that

works

anyway,

but

there

are

real

bottlenecks

this

and

these

lead

times

are

not

that

easy

compress.

And

look,

have

the

best

minds

the

world

trying

figure

out

how

compress

them.

but

it's

not

easy.

It's

not

easy

and

the

demand

not

slowing

down.

we're

already

behind.

The

demand

growing,

you

know,

10x

year

right

now.

And

you

the

supply

just

can't

grow

that

fast.

顺便说一句，情况糟糕到目前如果有新公司去寻找 GPU，往往是在墨西哥、澳大利亚或其他国家，就因为在美国太困难了。是的，由于这里禁止数据中心，我们正在其他国家创造巨大的就业机会和长期经济机遇。我认为，正确的答案应该是设定一个标准，让数据中心回馈社区——比如供电得到改善，没有噪音，没有水资源问题，嗯，并且它能增加就业，这应该成为标准。是的。然后每个人都应该按照那个标准来要求。顺便说一句，现在确实有数据中心做到了这一点。这并不是什么未来的梦想之类的。只要它们在那里，能源费率每年都会下降。原因是它们自己供电。它们在白天向州电网供电，然后在夜间州里不需要电时再借电，因为按照发电厂的运作方式，总是在按峰值容量发电，而由于数据中心日夜具有稳定的容量，城市则在白天大幅上升、在夜间大幅下降，嗯，那是一种共生关系。

放大来看。呃，我们为什么认为，你知道，我们有一阵子一直在反复探讨这个名字。呃，我们为什么认为“机器时代”（machine age）对于我们在这里所做的事情是一个令人信服的术语？

英文原文

>> By

the

way,

bad

that

right

now

have

new

companies

going

for

GPUs,

it's

often

Mexico

Australia

another

country

just

because

difficult

the

United

States.

Yeah,

we're

creating

huge

both

job

and

long-term

economic

opportunity

other

countries

banning

data

centers

here.

think

look

the

right

answer

would

set

standard

where

data

center

contributes

back

the

community

that

power

gets

better,

there's

noise,

there's

water

issue,

and

it's

adding

jobs

that

ought

the

standard.

Yeah.

And

then

everybody

ought

just

held

that

center.

And

the

way,

there

are

there

are

data

centers

that

now.

that's

not

you

know

futuristic

dream

something.

Rates

energy

rates

have

gone

down

every

year

they're

there.

And

the

reason

they

provide

their

own

power.

They

give

power

the

state

during

the

day

and

then

night

they

borrow

power

from

the

state

when

the

state

doesn't

need

because

you're

always

the

way

power

plants

work

you're

always

generating

peak

capacity

and

since

data

center

has

steady

capacity

during

day

and

night

and

city

goes

way

the

day

and

way

down

night

that's

symbiotic

relationship.

zooming

out.

why

think

the

you

know

we've

been

batting

around

the

name

for

little

bit.

why

think

machine

age

compelling

term

for

what

we're

doing

here?

“机器时代”与算力基础设施的创新空间

听我说，让我来回答。首先，我认为 Ben 说的完全正确。人工智能（artificial intelligence）用词不当。我们本不应该这么称呼它，它是机器智能（machine intelligence）。嗯——

英文原文

>> Well,

listen,

let let

let

take

it.

So,

the

first

one

think

Ben's

absolutely

right.

Artificial

intelligence

was

the

wrong

word.

shouldn't

have

called

it's

machine

intelligence.

多说一点这个。为什么会这样？

英文原文

>> say

about

that.

Why

that?

呃，因为它不一定像人类那样思考，对吧？我是说，它是人类思考方式的缓存，是人类思想的集合。但迄今为止，我们还不知道如何把一个毫无知识的 AI 放到世界中让它重建语言，对吧？那并不是我们所做的事情，对吧？我们制造了一种东西，它能从我们已经学到的一切中学习，然后以一种富有成效的方式利用它。听着，AI 在计算机科学中是一个可以正式追溯到 70 年前的通用术语，适用于许多不同的事物。当然，它带有许多包袱，要么来自科幻小说，要么来自写过这方面内容的 Nick Bostrom 等人。所以，第一点就是承认这确实是机器智能。然后你想要强调其中的“机器”部分。我的意思是，有一种深深的讽刺——这来自于“软件正在吞噬世界”那帮人——你真的来到了这样一个境地：你把钱投入到某件事物中，然后你被其下方的实际机器所限制。因此，我认为这算是一种致敬，即硬件构件在这一波浪潮中如此重要，我们希望承认这一点。

英文原文

>> Uh

because

not

how

humans

think

necessarily,

right?

mean,

cache

how

humans

thought

collection

humans

thoughts.

But

date,

don't

know

how

take

with

knowledge

and

put

out

the

world

and

have

reconstruct

language,

right?

that's

not

what

we've

done,

right?

We've

we've

built

something

that

can

learn

off

everything

we've

already

learned

and

then

use

that

productive

way.

And

listen,

general

term

that

goes

back

years

computer

science

formally

that

applies

many

different

things.

And

course

it's

got

lot

baggage

either

from

science

fiction

from

you

know

Nick

Bstrom

who

wrote

about

whatever.

And

the

first

one

just

acknowledgement

this

really

machine

intelligence.

And

then

you

want

emphasize

the

machine

part

it.

mean

there's

kind

this

deep

irony

and

this

from

you

know

the

software's

eating

the

world

people

that

you

know

you've

really

come

place

where

you

pour

money

into

something

and

then

you're

limited

the

actual

machines

below

it.

And

think

kind

nod

the

hardware

component

significant

this

wave

and

want

acknowledge

that.

是的。我认为在智能领域创造下一次突破的，将是底层机器的质量。所以这就是这个名字的基本缘由。

英文原文

>> Yeah.

think

that's

what's

going

create

the

breakthroughs

intelligence

the

quality

the

machines

underneath.

And

that's

basically

reason

for

the

name.

这也是个很酷的名字。[笑声]

英文原文

>> It's

also

cool

name.

[laughter]

听起来不错。

英文原文

>> That

sounds

good.

具有未来感。

英文原文

>> Futuristic.

是的。鉴于迄今为止在 AI 基础设施上的投入规模，以及这些业务所需的资本支出密度，我们是否已经过了新公司能够切入并达到实质性规模的时点？为什么像 Nvidia、Core 等现有巨头不能直接占据这些市场的大部分份额？

毫无疑问，他们都做得很好，对吧？但正如我们之前讨论的那样，当你需要根本性的新创新来维持持续增长或保持改进速度时——无论是以每美元每秒生成的 token 数、每瓦特 token 数，还是每机架 token 数来衡量，对吧？或者电力，任取一个指标——如果你想在这些指标上实现 10 倍的提升，就必须有新的创新。而传统上，新的创新来自于杰出的创始人从第一性原理出发、以不同的方式思考解决问题，对吧？而这正是为了实现下一次创新飞跃所需要的。

英文原文

>> Yeah.

the

given

how

much

has

been

spent

infrastructure

date

and

how

capex

intensive

these

businesses

can

are

past

the

point

where

new

companies

can

break

sort

material

levels

you

know

you

know

why

not

incumbents

Nvidia

Core

etc

just

take

the

line

these

markets

they're

all

doing

well

there's

question

about

right

but

our

discussion

earlier

when

you're

get

you

need

funly

new

innovations

keep

the

growth

both

continuing

the

pace

improvement

continuing

whether

it's

tokens

per

second

per

dollar

tokens

per

watt

tokens

per

rack

right

power

you

take

any

metric

you

want

have

10x

those

metrics

you

got

have

new

innovation

and

new

innovation

traditionally

comes

from

brilliant

founders

thinking

about

solving

the

problem

from

first

principles

different

way

right

and

that's

what's

needed

here

for

the

jump

innovation.

我是说，这就是市场规律，对吧？假设现有的芯片在位巨头拥有数万亿美元的市值——事实绝对如此。

英文原文

>> I

mean

this

the

law

markets,

right?

mean

let's

assume

that

the

existing

silicon

incumbents

are

multi-

trillion

dollars

market

cap

which

absolutely

the

case.

哪怕

英文原文

>> Even

其中的 5% 也是一家庞大的非上市公司，极其庞大的非上市公司，对吧？我们谈论的是年度规模。你可能会说，好吧，但 Nvidia 也能做那个。他们确实能做，但如果他们专注于占 90% 且能带动同等增长的业务，他们为什么要去做呢？你总会问这些问题。我们在云计算时代就问过这些问题，对吧？比如，亚马逊为什么不做这个？你在微软时代也会问这些问题：微软为什么不做这个？市场中存在着一种非常自然的规律，那就是一旦你达到了一定的规模，就会有巨大的创新机遇在在在

英文原文

that

massive

private

company,

massive

private

company,

right?

We're

talking,

you

know,

annual.

and

you

could

say,

well,

but

Nvidia

could

that.

They

could,

but

why

would

they

they're

focused

things

that

are

the

90%

which

also

driving

the

same

amount

growth?

You

always

ask

these

questions.

ask

these

questions

during

the

cloud

days,

right?

why

wouldn't

Amazon

did

this?

You

would

ask

these

questions

during

the

Microsoft

days.

Why

wouldn't

Microsoft

this?

There's

very

natural

law

markets

once

you

get

certain

scale,

there's

tremendous

opportunity

for

innovation

the

margins.

英文原文

margins.

对，我们合伙人 Alex Rmpelle 讲过一句挺有意思的话。他之前做过一家叫 Trial Pay 的创业公司，当时想把它卖给，或者说想把这项业务卖给 Meta，也就是后来的 Facebook。当时负责企业发展的 Dan Rose 对他说：“Alex，这挺棒的，听起来你能捡到很多银砖，但我手里的金砖多到我都拿不过来，所以我最不可能做的事就是去看一块银砖。”我觉得英伟达现在就是这种处境。

英文原文

>> Yeah, there's a funny uh quote from our partner Alex Rmpelle. He had this startup called Trial Pay and he was trying to sell it to or sell the it his services to to Meta and uh then Facebook and Dan Rose who was the head of corp dev at the time said Alex that's great you're it sounds like you can collect a lot of silver bricks but I'm like I have so many gold bricks I can't even pick them all up so the last thing I'm doing is looking at a silver brick and I think Nvidia is in that position

市场扩张与分散

100% 对。我们之前谈到模型提供商时也说过，如果你正好处于 open airropic 能做的事情的核心领域，也就是他们主要的几个关注方向之一，那日子可能会很难过；但这三到五个领域之外的任何东西，可能……大家知道，随着市场扩大，它们就会走向分散，对吧？这种情况一直在发生。想想早期的福特，1913 年建了胭脂河工厂。说白了，这就是造车的地方，一头运进水、煤炭和橡胶树，另一头一辆辆汽车就开出来了。

英文原文

>> 100% yeah we were talking about as relates to the model providers that if you're, you know, in the sweet spot of of what open airropic can can do, you know, one of their main sort of interest areas, that might be a tough place to be, but anything outside of those maybe, you know, three to five areas might uh, you know, as as markets expand, they fragment, right? And it happens all the time. And remember in the early days of Ford, there was the 1913, there was the Rouge River plant. Literally this, you know, this was like made cars like in went like water, coal, and rubber trees and out came cars.

顺便说一句，他在亚马逊雨林买下了整整一片……

英文原文

>> By the way, he bought a whole

橡胶种植园，

英文原文

>> rubber

对吧？有一本很棒的书叫《福特利亚》，讲的就是因为他想彻底实现垂直一体化，所以在亚马逊雨林建了一座叫福特利亚的城市，里面完全美国化，有演奏台、冰淇淋等等这类东西。这套东西还真运转了一阵子，直到他要求人们准时出勤，大家就受不了了，觉得“去你的，赶紧滚出这里”。[笑声] 所以现在你看汽车行业，显然存在多层级的供应商，有大批的公司，这种情况向来如此。随着市场扩大，就会走向分散。等增速放缓后，往往又会走向整合。整合要么是通过收购，要么是崛起新的挑战者，这就是一级市场永恒的循环。

英文原文

plantation in the Amazon jungle,

right? And and there's a great book called Ford Landia where so because he wanted to like own like the complete vertical thing where he created this city called Ford Landia in the Amazon jungle um which had like was all Americanized band stands and ice cream and all this kind of stuff. And it actually worked for a while until uh he made people like show up to things on time and then they were like screw this get the out of here. [laughter] So, so now if you look at the car industry, of course, there's multiple levels of supplier and there's a bunch of companies and this always happens. So, you know, as markets expand, they fragment. There's a lot of and then and then once that growth slows down, they tend to consolidate. The consolidation can either be acquisition or it can be like new challengers rise up and that that is the, you know, everlasting cycle of private markets.

是的。应用场景正在成倍增加，要是你是规模最大的公司，固然可以拿下最大的那些应用场景，但应用场景实在太多了，而且正如 Martin 所说，全都是极有价值的场景，单凭一家很难全部完美覆盖。

英文原文

>> Yeah. the use cases are are multiplying and there's no way like if you're the biggest company you can get to the biggest use cases but there's so many use cases and all as Martin was saying very valuable use cases that it's just very hard to get to in a great way

对，就连推理曾经也是一种简单的架构，对吧，现在已经不是了，现在变得非常复杂，所以必然可以通过不同的方式来进行优化。

英文原文

>> yeah even inference used to be one simple architecture right and it no longer is like it's so complex now so it's inevitable that you can optimize things in a different way

硬件与系统级创业

顺便说一句，这里有个很有意思的现象：大家往往不太明白，以前用软件做技术的时候，利润率其实是随着标准做法自然而然产生的，对吧？它并不是一个真正的技术问题。一旦把业务做起来，通常就会有相当不错的利润率，因为软件的运作机制大致就是那样。当交付软件产品时肯定是这样，哪怕作为服务来做也是如此。但在 AI 领域情况不一定这样。我们可能正在迈入这样一个时代：硬件层面的优化，对业务的上升空间具有绝对重大的意义，其程度是以往未曾见过的。所以这里面存在大量的机会。我们来深入探讨一下我们会投资的公司类型。或许我们可以先列举一下各个细分领域，或者聊一聊我们已经投过的几个项目——我知道有些项目还没对外公布，不过 Robert，你想带我们梳理一下吗？

英文原文

>> by the way here's a very interesting thing like people don't um uh often don't understand that like the margins kind of fell out of the standard way of doing the technology with software, right? Like it wasn't really a technology problem. Like once you got the business working, you tended to have pretty good margins because that's just kind of how software works. certainly when you shipped it but even as a service uh and that's not necessarily the case with AIS we may actually be entering an era where the optimization in the hardware is absolutely meaningful to the upside of the business in a way that we haven't seen in the past so there's a lot of opportunity here let's get deeper in talking about the types of companies we'll be investing in um maybe we could start by either illustrating the the the subsectors or if we can talk about a few um or a couple investments that we've I know there's some that haven't been announced yet, but Robert, do you want to take us down?

好的。正如我们一直在聊的，细分领域涵盖这里的每一个类别，对吧？显而易见的类别是计算机芯片，但如今仅仅做出一颗芯片已经不够了，你必须构建一整套系统，对吧？那么，进入这套系统的都有什么？可能有存储方面的创新，可能有网络方面的创新，可能有电源芯片等等。所以这里面的每一个类别，你都能看到上市公司级别的企业正在涌现，这些全都是我们正在调研的方向。当你把这一切整合在一起后，外围还需要一层软件来实现所有这些环节的自动化、管理这些集群等等。所以那也是一个重要领域。这些东西在不断层层累加，但其中的每一个类别都至关重要。

英文原文

>> Yeah, I mean the sub sectors as we've been talking all is every one of these categories, right? The obvious ones are um uh computer chips, but these days it's not enough to build a chip. You need to build a full system, right? And then therefore, what goes into the system? There's potentially memory innovation. There is potentially networking innovation. There's potentially power chips and so on and so forth. So each one of these categories are categories where you can see public company style companies emerging and those are all things that we are looking into. Um and then once you put it all together there's a layer of software around it to automate all of these things to manage these fleets and so on and so forth. So that is another important area. So these things keep building on each other but every one of these categories is important.

能不能聊聊这类公司跟通常的公司相比有什么不同？从我们已经公布的公司就能看出，它们的首轮融资规模都非常庞大，动辄数以亿计。这是因为创始人的类型不一样，还是说在我们思考围绕这类业务的创办和投资实践时，相对于我们传统的软件投资，还有哪些不一样的地方？

英文原文

>> Talk about what's different about these kinds of companies from from the usual company. I mean one thing you can tell from the companies we announced is their first rounds have been massive. You know hundreds of millions. Is it a different kind of founder or what else is different as we think about just the practice of you know building and investing these kinds of businesses relative to our traditional software?

我觉得最核心的一点是，你刚才切中了一个关键：在拿出产品之前，就已经投入了大量的资金。这本身就是其特性决定的。大模型其实也是这样，但我认为大模型走的是一条相对已知的路径；而这类业务比我们做过的其他一些事情风险稍大一些，需要的资金也稍多一些。不过……

英文原文

>> Well I I think the big thing is you hit on one of the big things which is a lot of money goes in um before they get to a product. Um and that's just kind of the nature of it. Now that's true on big models too, but I I would say that's a little more of a known path. Uh whereas this has got a little more risk and a little more money than uh than some of the other things that we've done. But

而且你知道，许多芯片领域的创始人都是业内老兵。

英文原文

>> um and you know, look, a lot of a lot of the chip founders um are here from the past.

比如那些懂得如何制造存储器的人，[嗤笑] 年纪都不小了。对吧，[笑声] 所以这一点也是不同的，但这挺令人兴奋的。

英文原文

>> You know, like the guys who know how to make memory, [snorts] they're not young. Yeah, [laughter] you know, so it's uh you know that part is different too, but it's kind of exciting, you know.

对。关于这些创始人的另一件事是，他们都必须是系统级创始人。我的意思是，你不能仅仅是个研究员或者厉害的计算机科学家，对吧？你必须能够构架和设计芯片或系统，不管具体是什么；然后你必须思考这套东西到底要怎么制造出来，对吧？谁来提供供货，以及一大堆下游的事情。通常做软件的时候，根本不需要考虑所有这些事情。因此，真正最顶尖的创始人——当然 Jensen 是这方面的迈克尔·乔丹，对吧——他们从一开始、在着手设计链条之前，就已经在通盘考虑整个生态系统了，这是因为瓶颈的性质以及所有必须捏合在一起的环节。所以这是一个非常不同的大特征。

英文原文

>> Yeah. The other thing about these founders, they have all got to be systems founders. So what I mean by that is you can't just be a researcher or a great computer scientist, right? You got to be able to architect and design the chip or the system, whatever it is. Then you got to think about how is this thing going to actually get manufactured, right? who's going to be supplying this and a whole bunch of these downstream things which normally if you're building software you don't have to think about all of these things. So the to really the best founders um and of course Jensen is the Michael Jordan of this right they think the entire ecosystem right from the get-go before they start designing the chain right because of the nature of the bottlenecks and all these things that have to come together. So that's that's a big characteristic that is different.

环境变化与创始人画像

还有两个大环境因素也很重要。第一个是，那些实验室现在迫切到了愿意同创业公司合作的地步。因此，我们在极早期其实就能获得相当多的信号，因为这些实验室在创业公司硬件实际做出来之前，就已经在跟他们敲定合作协议了。这相比五年前是一个非常巨大的转变，对吧？以前你根本不可能拿着某种半成品的破烂硬件跑去卖给谷歌之类的公司。所以这是一个转变。第二个是资金充裕度大幅放宽了。我认为大家普遍达成了一种共识，即现在确实是重塑这些东西的时候了，所以在后续轮次中有大量的资金可用。显然，你总会希望投资于资金充裕的领域，所以外部大环境也是完全不同的。Patrick Carlson 几年前曾经点评过，说：“感觉现在的年轻创始人变少了，不像扎克当年在大学里创办下一代 Facebook，或者盖茨对微软那样，当然还有全世界的迈克尔·戴尔们。如今虽然仍有一些年轻创始人在打造标志性公司，但正如你所指出的，似乎有更多年长的创始人在打造这些公司，或者说二十几岁的年轻人变少了。”我很好奇这是否引起了你的共鸣，以及为什么会这样。

英文原文

>> There there's two environmental factors that are important too. The first one is the labs are so desperate that they will engage with startups. And so like we actually have quite a bit of signal early on because they're you labs are inking deals with companies before they actually have hardware available. And that's a big big shift than you know 5 years ago, right? Like you just didn't go and you know sell your kind of janky hardware thing to Google or whatever. So that that's a shift. The second one is the the capital availability has loosened up a lot. I think there's general consensus that that you know it is the time that to reshape this stuff and so follow on rounds there's a lot of capital available which you know of course you want to be investing into areas where there's capital available and so the atmospherics are also just different. Patrick Carlson, you know, remarked a few years ago said, "Hey, it feels like there's less younger founders today, you know, in the way that you know, Zuck, you know, in college building next Facebook or or Gates, um, you know, in the same way with Microsoft and of course, you know, the Michael Trolls of the world, there there's still, you know, some young founders building iconic companies, but it does seem, you know, to your point that there's more older founders building these these these companies or or less 20-year-olds. I'm I'm curious if you resonates and why.

嗯，我认为正如 Ragu 所说，如果你做的事情有着非常复杂的供应链，必须制造实体产品，而且在技术上极其复杂，那么经验肯定是有帮助的。你看埃隆或者特拉维斯·卡兰尼克，他们年轻时创办的公司都是软件公司。甚至连他们这些最顶尖的人，也是积累了创办公司、研发技术等等各方面的经验之后，才逐步晋级到

英文原文

Well, I think it's Ragu's point that if if you're building something that has like a very complicated supply chain, has to manufacture things um and is technically complicated that you know some experience helps uh and you know if you look at Elon or Travis Kalanick their companies when they were young were software companies. It wasn't until they got like a lot even those guys the best guys um needed some experience in building a companies building technology and so forth to kind of graduate to

在那些更加复杂或者说更精密的领域，其中牵涉的东西要多得多，有更多的活动部件。所以你看，当你学习如何创立一家公司时，即便你对产品了如指掌，这本身就已经够难了。要是你对产品并非完全理解，还得一边创办公司一边去摸索学习，嗯，那对一个初次创业者来说学习曲线实在太陡峭了。所以，我认为我们现在看到的情况是：一方面你看到了迈克尔（Michael），嗯，他非常年轻，才华横溢，但他做的是一种纯软件、纯 AI 的东西。

英文原文

the much more kind of complicated or I would say elaborate domains you know there's just much more there many more moving parts in these things and so look when you're learning how to build a company it's hard enough if you completely understand the product. If you don't completely understand the product and have to learn it while you build the company, um that's just such a steep learning curve for a brand new entrepreneur. So, I think that what we're seeing is you see Michael on the one hand, um who is a very young guy, brilliant, but what he built was kind of a pure software AI thing.

是的。而在另一端，你会有像埃隆（Elon）或特拉维斯（Travis）这样拥有足够经验的人。我觉得迈克尔从现在算起的 10 年后大概也能做到，但在今天，那会很难。

英文原文

>> Yeah. And then on the other end, you have like an Elon or a Travis who can who's got enough experience. I think Michael could probably do that, you know, 10 years from now, but today that would have been hard.

并且必须记住，过去 20 年来，整个行业和学术界其实都不太关注这个方向了，对吧？就是没有同样的机会。它一直存在，但从不是一个增长领域。那些增长领域一直是，你知道的，软件、网络诸如此类的方向。所以我也认为，我们从大学毕业或在大公司有这方面经验的人才很匮乏。这样的人真的不多。

英文原文

>> And it's important to remember like it's been defocused by the entire industry and academia for the last 20 years, right? It just there just hasn't been the same opportunity. Like it's been there, but like it's never been a growth area. The growth areas have been, you know, software, um, networking, things like that. And so I also think we just have a posity of people coming out of the universities or having experience at large companies that have have done this. I there just not that many.

就像你不会去实习然后做出一款芯片一样。不过现在很多情况正在改变。你看，我们将会培养出整整一代来自这些新公司的创始人，他们会懂得如何做这些事情。而且他们会以更年轻、更初级的资历被招揽进来。其实我觉得，埃隆在这方面留下的最伟大遗产之一，当然在于他创立了这些伟大的公司，但从 SpaceX 走出来、正在改变整个工业复合体的创业者数量，或许是比这些公司本身更伟大的遗产。我认为在计算机科学和硬件领域，我们也将看到同样的景象。

英文原文

>> Like you don't go intern and like build a chip. So but a lot of that's changing now. like listen we're going to create a whole generation of of you know founders that come from these new companies that will know how to do this and you know they'll be hired in much more junior and like I would say actually one of the greatest legacies of Elon towards this is is of course he's created these great companies but the amount of entrepreneurs that have come out of SpaceX that that have are changing the entire industrial complex maybe even greater um legacy than than the companies themselves and I think we're going to see the same thing uh for for computer science and hardware

事实上，我们的一项投资就是由两位 20 多岁的创始人发起的。但如果你去他们的其中一个办公室看看，你也会看到经验丰富的人。所以这是很理想的结合。

英文原文

>> y as a matter of All of our investments was started by uh two founders in their 20ies. But if you go watch one of their offices, you see the experienced people as well. So it's ideal combination here.

是的。

英文原文

>> Yeah.

是的。没错。不一定非得是创始人拥有经验，但那个创始人最好能够调动并利用那份经验。

英文原文

>> Yeah. Yeah. It doesn't necessarily have to be the founder with experience, but that founder better be able to tap into that experience with it.

新基金与硬件底色

没错。对，而且还要能够与他们共事，他们也必须很优秀，以及诸如此类的事。这挺复杂的。说到经验，这是我们正在发起的一只大型新基金，并且没有引入新的普通合伙人（GPs）。嗯，是的，我们某种程度上是在收拢资源。这是因为你们以及团队的其他人，在这个领域拥有很多经验，这些经验之前一直处于相对潜伏和休眠的状态。是的。其实挺有意思的。我认为我们甚至几乎需要被提醒警惕这一点，正是因为我们的背景来自……硬件很艰难。我觉得我们之所以需要这个提醒，是因为我们所有人在各自的职业生涯中都在硬件领域投入了太多心力。我们天然地会被它吸引。所以你看，这些年来我们显然一直有在硬件领域投资，对吧？我们投了 SpaceX，投了 Anduril。这些都是非常早期的支票。我们还投了 Astranis，投了 Waymo，等等，所以即便在早期我们也做过许多这类的投资。嗯，但你看，这是因为这已经深深植根于我们的基因中，因此我认为并不一定需要增加团队的能力储备，这只是额外优势。

英文原文

>> Yeah. Yeah. Well, and then be able to work with them and and and they have to be good and and all these kinds of things. It's complicated. Speaking of experience, this is a a big new fund we're we're launching and there's no new GPS. Um yeah, we're sort of collecting. It's because you guys have a lot of experience and the rest of the group, you know, in in this field that it's been kind of latent and uh dormant. Yeah. Well, it's kind of funny. I think we almost had to be warned against it almost just because like our backgrounds are from is hard. And I think the reason that we needed a reminder is because all of us have spent so much in our careers existence and hardware. We're kind of drawn to that. And so listen, we've been clearly invested um in harbor over other years, right? We're in SpaceX, we're in Andrew and all. These are very early checks. We're in astronomics, we're in Whimo, you know, so we even even early on we did a number of those investments. Um but like you know this is because it's so much in our DNA and so I don't think this is necessarily need to increase the team competencies just bonus.

如果这只基金达成了我们预期的目标，

英文原文

>> If this fund does what we think it will do,

我们认为 5 到 10 年后世界会发生怎样的改变，或者会是什么样子？嗯，你知道的，但愿美国在基础设施竞赛中胜出。并且我们能在各处拥有大量超级环保、高效的数据中心，拥有大量的芯片、充沛的内存和充沛的电力。那样的话就太棒了。而且我认为，你看，这又回到了那个问题——我们真的认为美国是一个特别的地方，我们不仅对身处这里的每一个人很重要，对世界上任何想做出贡献、做一番超越自我之事的人来说也同样重要；这里可以说是一无所有地来到此地并成就深远事业的最佳之所。所以我们希望能把这种势头保持下去。我认为，如果我们丢掉了在技术上的领先地位，这一切就无法维系了。我想届时我们将进入另一个时代，换作另一个国家，而他们围绕这些可能有着截然不同的一套价值观。

英文原文

>> how do we see the world changing or looking like in 5 to 10 years? Well, you know, hopefully uh America wins in the infrastructure game. Um and we have lots of like super eco-friendly efficient data centers out there and lots and lots an abundance of chips and abundance of memory and abundance of power. Uh and you know that would be awesome. Uh, and I think we look we we, you know, it goes back to like we really think uh America is a special place and um we're important not only to everybody here but anybody in the world who wants to kind of make a contribution and do something bigger than themselves that it's kind of the best place to come with nothing and do something profound. So we'd like to keep that going and I I think that doesn't continue to go if we lose our lead in technology. Think I think we'll be in another era and it'll be another country and maybe they have a different set of values around that.

谢谢。录制结束。

英文原文

>> Thanks. That's a wrap.

很高兴见到我，很有趣。马丁（Martine），本尼（Beni）。谢谢你们。谢谢。

英文原文

>> Great for seeing me fun. Martine Beni. Thank you. Thank you.

谢谢。谢谢。好的。

英文原文

>> Thank you. Thanks. Yeah.