可编程数据中心：对话 SF Compute 的 Evan Conrad

英伟达显然是房间里的大象。他们为什么不——为什么他们没有造一台超级计算机？

Nvidia's obviously like the 800-pound gorilla in the room. Why don't they—why didn't they make a supercomputer?

我们的印象是，英伟达造不了，但它从超级计算机的存在中获益巨大。他们造不了的原因是，那样的话他们就不够独立，人们无法在其上交易。

So our impression is that Nvidia can't make it, but benefits tremendously from it existing. The reason why they can't make it is because they would then not be independent enough for people to trade on it.

总之，Evan，欢迎回到 Latent Space。欢迎回到 AIE。其实是第一次来 AIE 吧？

Anyway, Evan, welcome back to Latent Space. Welcome back to AIE. Actually, first time at AIE?

我想是的，对。

I think so, yeah.

这非常特别，因为我觉得你已经在计算机交易、计算机市场上工作了一段时间。现在感觉正是时候。

This is a very special one because I feel like you have been working on computer trading, computer markets for a while. Now it's like feeling of the moment.

是的。

Yeah.

Jump 基本上宣布他们要进入计算机市场，这就是为什么我很兴奋能把人们聚在一起。你见到了 SemiAnalysis 的 Oliver，他基本上正在开始覆盖计算机市场。这是我的解读。所以人们会去看你的演讲。

Jump basically announced that they are getting into computer markets, which is why I'm excited to sort of cluster people. You met Oliver from SemiAnalysis, who is basically initiating coverage on computer markets. This is my read on this. So people are going to watch your talk.

我认为这很好地解释了你的起源故事、你为什么需要存在以及你解决的难题。

I think it's a very good explanation of both your origin story and why you need to exist and what the hard problems you work on.

以非常简单引人入胜的方式完成。我非常嫉妒你的写作技巧，因为你展示了你在 NeurIPS 上展示的传单。是的，它非常直接、简单，很好。

Done in a very simple engaging manner. I'm very jealous of your writing skills because you showed the flyer for that you showed at NeurIPS. Yeah. It was like it's pretty straightforward and like simple and like it's good.

是的。

Yeah.

总之，恭喜。

So anyway, congrats.

谢谢，我很感激。也恭喜 AIE。看起来你们在滚动发展。

Thank you. I appreciate that. And also congrats on AIE. It seems to—y'all are rolling.

我们在滚动发展。我刚从明年 AIE 的场地考察回来。然后对于——

We're rolling. I just came back from the site tour for next year's AIE. And then for—

给我点提示。

Give me like a hint.

在贾维茨中心。和他们举办动漫展的地方一样。

At Javits Center. It's the same place they do Comic-Con.

好的，哇。

Okay, wow.

哪个——

Which—

哇。哇。该死。

Wow. Wow. Damn.

但我认为金融焦点确实打动了人们。

But I think the finance focus is really hitting for folks.

是的。

Yeah.

所以我们将尝试拆分，就像这里我们把所有金融都归在一起。你的金融、他们的金融，保险也是金融。但它们都超级不同。

So we're going to try to split up like this is—here we just lump everything in finance. Like your finance, their finance, like insurance is finance. But like they're all super different.

是的。

Yeah.

所以未来我们可能会更细致地划分，因为人们想深入了解对冲基金、银行或个人财富或——

So we'll probably break out things more minutely in future because people want to go deep on hedge fund stuff or banking or personal wealth or—

是的，尤其是在纽约。

Yeah, especially in New York.

是的，就像 Anthropic 说，是的，我们现在专注于 RIA。RIA 就是注册投资顾问。

Yeah, like Anthropic was like, yeah, we're all about RIAs right now. RIAs like registered investment advisors.

是的。

Yes.

和你如此不同。

Like so different from you.

是的，不，不，我知道这个。

Yeah, no, no, I'm aware of this.

是的，是的，是的。嗯，好的，对于不了解的听众，我们两年前做过一个播客，你谈到了你的——你知道，SF Compute 当时的情况。

Yeah, yeah, yeah. Um okay, so for listeners who don't know, we did a podcast 2 years ago where you talked about your—you know, where SF Compute was.

嗯。

Mhm.

嗯，今天你更多地涉及这种实物结算。你谈到你的背板结构还是什么？

Um today you're much more in this sort of physical settlements. You talk about your back fabric or what do you call it?

呃，背板。

Uh the backplane.

背板，显然需要很多硬核工程。嗯，你能给个状态更新吗，比如 SF Compute 现在在哪里，相对于两年前在播客上听到你的听众？

Backplane, which obviously like takes a lot of hardcore engineering stuff. Um can you give a status update on like where is SF Compute versus listeners who last heard you on a podcast 2 years ago?

当我们开始时，我们是一家叫 Junnar 的公司，然后我们只是试图转租我们的集群。嗯，我们从技术栈的那里开始。嗯，一种思考方式是，就像你在酒店里，试图出售那个酒店房间。然后有人问你，‘嘿，你会打扫我的房间吗？’然后你说，‘哦，我想我可能也需要酒店房间，或者我需要清洁壁橱。’于是你去获得清洁壁橱的使用权。然后他们想要客房服务，所以你去获得厨房的使用权。然后，你知道，他们想要安保，所以你去获得安保壁橱的使用权，你得到这个，你做那个，你做这个。在某个时刻，你就是在经营整个酒店。嗯，当你经营整个酒店的那一刻，你说，‘哦，等等。相反，我在这个市场上的供应商更像是基金或云——在我们的情况下，是能访问数据中心的云，我们要升级。’所以我们在过去几年所做的是构建云的整个运营能力，以便提供——以便真正可靠地交付算力。嗯，我们称之为实物结算。我们借用了一个术语——

So when we started, we were this company called Junnar, and then we were just trying to sublease our cluster. Um and we started there in the technical stack. Um and a way to think about it is it's like you're in a hotel and you're trying to sell that hotel room. And then somebody asks you, 'Hey, like are you going to clean my room?' And then you say, 'Oh, I guess I probably also need the hotel room as well or like I need the um the cleaning closet.' And so you go get access to the cleaning closet. And then they want room service, and so you get access to the kitchen. And then, you know, they want um security, and so you get access to the security closet, and you get this, and you do that, and you do this. And at a certain point, you're just running the whole hotel. Um and the moment you're running the whole hotel, you say, 'Oh, wait a second. Um instead, my supplier on this market um is somebody that looks closer to a fund um or a cloud that has access—in our case, a cloud that has access to the data center, and we're going to level up.' And so what we spent the last several years doing is building the entire operating capacity of a cloud um in order to offer—in order to actually deliver um the compute at a reliable level. Um we call that physical settlement. We're taking a term from—

期货市场——期货市场，来自其他市场。嗯，在石油或大豆之类的市场，实物结算意味着你得到石油和大豆的交付。

The futures market—futures markets, from other markets. Um where in, you know, oil or soybeans or something, physical settlement means you get delivered the oil and the soybeans.

是的。

Yes.

嗯，但在一些市场，比如石油和电力，嗯，有交付网络。在石油中，你必须要有管道。如果你想让石油交付给你，必须有办法到达你那里。

Um but um in some markets, um like in oil and also in power, um there's like a delivery network. In oil, you have to have pipelines. If you would like the oil to be delivered to you, there has to be a way for it to get to you.

是的。

Yes.

嗯，对于对金融历史感兴趣的人，在新冠期间，石油价格变成负值，因为呃，就像我会付钱让你从我这里拿走石油，因为它就放在港口，我付不起这个。

Um and and and for people interested in financial history, during COVID, oil went negative because uh so like I will pay you to take oil from me because it is sitting there in the the port and like I can't pay for this.

是的。

Yes.

所以我会付钱让你拿走我的石油。

And so I will pay you to take my oil.

是的。嗯，这就是实物结算的奇怪之处。

Yes. Um that that is like the weirdness of physical settlement.

是的。其他市场也有很多这样的奇怪之处。所以在电力市场和电网中，嗯，你填写的订单中有部分，买卖双方，嗯，是因为锅炉需要一段时间冷却。嗯，你可以，你知道，停止锅炉，然后它会在一定时间内重新启动，所以这必须被纳入订单簿。嗯，这个 Jenner 的事情在计算机市场也是一样。所以我们同时共同设计了实物结算和订单簿，这意味着我们能够实际处理算力的真实约束，这可能包括像‘哦，我可以接受被中断，但前提是设备上有存储，我可以快速恢复。’嗯，非常类似的问题。

Yes. There's a bunch of weirdnesses like this in other markets as well. So in the power markets and the grid, um there's parts of the orders that you fill out, the buyers and sellers, that um are there because the boiler takes a while to cool off. Um and you can, you know, stop the boiler and then it will—you can start it back up again within some period of time and so that has to get factored into the order book. Um this Jenner thing is the same thing that happens in computer markets. So we co-designed physical settlement and the order book at the same time and that meant that we get to actually handle the real constraints of compute, which might include something like, 'Oh, I'm okay with being interrupted, but only if there's like storage on the thing that I can recover from really quickly.' Um very similar types of problems.

是的。所以呃，人们可以参考你的演讲，你已经做过几个其他播客了。嗯，所以我想成为 201 级别而不是 101 级别的简短介绍，呃，也因为——我——我曾经是衍生品交易员，所以我可以给人们更多你通常不涉及的细节。

Yeah. So uh people can refer to your talk and you've done a couple other podcasts already. Um so I want to be like the sort of 201 level instead of the 101 level short uh and also because I—I am—I have been a derivatives trader so I can give people a bit more detail that you normally don't get into.

好的。

Okay.

嗯，例如，嗯，今天的计算机市场有多成熟、多深或多流动？

Um so for example, well, how how mature or deep or liquid is the computer market today?

呃，不是很成熟、很深或很流动。嗯，所以我们一直在做的是，首先我们必须——我们必须先让大量供应上线，这需要一段时间，然后我们必须构建整个平台来实现这一点。嗯，我们刚刚让它运行起来，所以现在我们将容量转移到订单簿上，嗯，这使其具有流动性。

Uh not very mature or deep or liquid. Um so, what we've been doing is first we had to—we had to go get a bunch of supply online in the first place, which just takes a while, and then we had to build the entire platform in order to do this. Um and we just now got that working, and so now we move um capacity over onto the order book, um which makes it liquid.

嗯，所以我们总容量的 5% 在订单簿上，而且我们一直在不断增加，然后我们还会带来更多容量，这些都会放在订单簿上。嗯，好处是，也许我应该说订单簿上的实际交易量很小，但订单簿上的流动性很高。所以，对于你拥有的那一小部分容量，你有非常集中的流动性。所以，对于使用这个的人来说，一个标准指标是，一个从我们这里购买容量的客户，他们花 100 美元购买容量，平均能通过转卖给其他人赚回 25 美元。嗯，这就是我们试图追踪的指标之一，因为我们希望流动性对商业参与者有用，你可以用多种不同的方式来衡量流动性。你可以说，“哦，我只是想要大量的做市商之类的。”嗯，你可以用高尚的方式来衡量。我们试图通过“这是否解决了客户的问题？”来衡量。

Um and so, 5% of our total capacity is on the order book, and we're consistently moving more and more, and then we're bringing on more capacity that's all on the order book. Um the nice part is that um maybe I should have said the volume on the actual order book is small, but the liquidity on the order book is high. So, for that small amount of capacity that you have, you have very concentrated liquidity. So, a sort of standard metric I would say for someone who is using this is a customer who buys capacity from us, and they spend $100 to buy that capacity, will make back $25 on average from selling it back to somebody else. Um and that's the number that we or this is one of the metrics that we try to track because we want the liquidity to be useful to the commercial participant, and you can measure liquidity in a bunch of different ways. You can say like, "Oh, I just want tons of makers or something." Um and you can measure it in high-minded ways. We try to measure it by like, "Is this solving the problem for the customer?"

是的，明白了。那么，我们上次谈话时，也有很多关于标准化合同的讨论。是的。嗯，比如和 CFTC 之类的机构谈，那些事情的进展如何？

Yeah, got it. And so, last time we talked as well, there was a lot of contract conversation about standardizing the contracts. Yes. Um talking with like the CFTC or whatever, like what's the status on those things?

嗯，所以，我认为合同标准化总体上还相当遥远。嗯，我们当时试图做的是，或者基本上，我们过去几年学到的是，我们只需要运营整个事情，否则你无法标准化这个东西。所以，为了更接近可以标准化的东西，你首先需要让所有的操作都相同。对。举个例子，一个运营商可以提供两种不同类型的虚拟机。一种虚拟机可能比另一种虚拟机性能下降 40%。而 40% 是很大的。

Um so, I think standardizing the contracts in general is still pretty far away. Um and what we were trying to do at the time was or basically, what we learned over the last several years was we just needed to operate the whole thing, otherwise you cannot standardize the thing. And so, to get closer to something that can be standardized, you first need to make all the operations of the thing the same. Right. To give an example of what can be different is a single operator can offer two different types of virtual machines. One type of virtual machine might have a 40% performance degradation versus another type of virtual machine. And 40% is a lot.

当你说性能下降时，你是指 MFU 之类的东西，还是你经历过的燃烧量之类的，或者什么？

When you say performance degradation, are you talking like MFU type things or like amount of burning that you've experienced or what?

它会影响。所以，在这个特定情况下，我指的是 GPU 内核启动以及发生所需的时间，这是一个在你进行训练或推理时相当频繁发生的过程。

It will affect So, in this particular case, I'm talking about GPU kernel launches and how long it takes for that to happen, which is a process that happens quite frequently when you're doing sort of either training or inference.

好的。

Okay.

嗯，那基本上最终结果是你会得到更差的 MFU，但嗯就像

Um the um uh that will basically the end result is that you get a worse MFU, but um like

是中间级别吗

Is the intermediate level

基本上是中间级别。

intermediate level, basically.

是的，是的。所以，所以我想说的是，对于任何交易过大宗商品的人来说，有不同的商品等级，是的，对吧？就像不是石油，是轻质低硫原油对比轻质低硫原油的反面，或者就像，你知道的，嗯 A 级玉米或 B 级玉米之类的。

Yeah, yeah. So, so what I'm trying to So, for anyone who's traded commodities before, there's different grades of commodity, yes, right? There's like It's not oil, it's light sweet crude versus like whatever exactly the opposite of light sweet crude is or it's like, you know, uh corn level A or corn B or whatever.

嗯，所以就像人们试图嗯，你必须建立等价物

Uh and so like people are trying to Well, you have to establish the equivalents

那些东西的等价物。

equivalents of those things.

对，是的。我想知道它们是否已经建立。

Right, yes. And I'm wondering if they they have been established.

所以，我们取而代之的是创建了一个有多个不同约束的订单簿，你可以说，“我想要嗯，你知道的，介于这个和那个之间的东西。嗯，它从这个时候运行到那个时候。嗯，就像嗯 X 到 Y。”你可以附加一堆不同的属性到你的订单上，然后我们在一个大的匹配引擎中将其与具有那些属性的卖单匹配。嗯，我们这样做的原因是它允许我们生成流动性，而不必在你知道技术能做什么之前进行标准化。嗯，因为技术变化如此之快，嗯，我们希望能够基本上做到这一点。这样做的代价是你的流动性不那么集中，所以价格嗯会有，买卖之间的价差会更大。

So, what we did instead um was we created an order book that had multiple different constraints, and you can say, "I want to have um you know, something that's between this and that. Um it runs from this time to this time. Um it's like uh X to Y." It's a bunch of different properties that you can attach your order, and then we match that against sell orders that have those properties um in a big matching engine. Um and the reason why we did it this way is it allows us to generate liquidity without having to standardize before you know what the technology can do. Um and because the technology was shifting so quickly, um we wanted to basically be able to do that. The trade-off of this is that you have less concentrated liquidity, and so the prices are um going to have There's going to be more spreads between the buy and the sell.

是的。

Yeah.

嗯

Um

这这很糟糕。是的。

It It sucks. Yeah.

但是，另一方面，这意味着如果你能正确做到，就有很多套利机会。所以，如果你有一个可以使用多种不同类型容量的产品，嗯，那么你可能会非常喜欢 SF compute，因为你将能够使用多种不同类型的容量，并且你会赚更多的钱。嗯，你可以赚取价差。

But, on the other hand, it means there's lots of arbitrage to make um if you can do that correctly. So, if you have a product that can use lots of different types of capacity, um then you might really like SF compute because you will be able to use lots of different types of capacity and you'll make more money. Um you can take the the spread.

好的。在这个网格中，哪些维度重要？

Okay. What dimensions matter in this grid of things?

嗯，所以有很多不同的东西。嗯，一个是维度类别，我会说比如现在是否有工作负载在运行，还是未来的容量？所以是现在的容量吗？嗯，现在的容量，为了让你出售它，嗯，你必须中断你当前的工作负载。所以你必须考虑我是否关心我当前的工作负载被中断？一个月后开始的出售容量，嗯，你可以随时买卖。所以你可以很容易地获得流动性。嗯，有一些担忧，比如哦不，下个月我会有容量吗，但这与你现在中断工作负载非常不同。嗯，中断发生的方式很重要。所以一个例子是，如果你在一个集群上，而该集群没有你想要的那么好的存储。嗯，所以一个例子是，你在底层是否有 Weka 或 Vast 系统？嗯，你提供那个吗，大小是多少？嗯，这改变了你对被抢占或从集群中移除的反应方式。嗯，同时存储的位置也很重要。所以，如果你能够快速地将数据从一个集群移动到另一个集群，那么也许你可以分布在多个不同的集群上。嗯，你可以在这里被中断，然后在那边重新出现。嗯，为了实现这一点，你还需要一个集群之间的快速传输路径。通常这是用 WAN 实现的，但

Um so a bunch of different things. Um one is dimensions categories that I will say is like is there a workload running on it right now or is it capacity in the future? So is it capacity right now? Um capacity right now in order for you to sell that um you have to interrupt your current workload. So you have to think do I care if my current workload is interrupted or not? Capacity that's sold that's starting a month from now um you can buy and sell that all the time. And so you get um liquidity on that really easily. Um there is some concern like oh no will I have capacity next month but it's very different than if you're interrupting a workload right now. Um the way that the interruption happens matters. So an example is if you're on a cluster and the cluster doesn't have as amazing storage as you want it to have. Um so an example is like do you have a Weka or Vast system under the hood? Um and do you offer that and like what's the size of that? Um uh this changes how you respond to um being preempted or removed off the cluster. Um meanwhile the location of that storage matters. So um if you're able to move that data from one cluster to another um quickly then maybe you might be able to spread across multiple different clusters. Um and you can be interrupted over here and then pop back up over there. Um to make that work you also need um like a fast transit path between one cluster and another. Typically this is implemented with a WAN but

大多数这些集群不一定有

most of these clusters don't necessarily have

嗯

um

一个广域网。

a wide area network.

是的，一个广域网。所以基本上就是连接集群的电缆。私有互联网是另一种思考方式。嗯，今天正在建立的许多集群，它们不像传统的 CPU 集群那样为了延迟原因而靠近客户。在 GPU 中，它们靠近电源。嗯，无论哪里，你知道你把数据中心放在电源旁边而不是客户旁边。嗯，这意味着很多东西被放置在以前他们不期望有数据中心的地方。所以，你可能没有从一个到另一个的那么优秀的传输路径。所以，这开始变得重要。嗯，其他例子是嗯，奇怪的，我们理论上支持但实际上不支持，嗯，比如，这是我们正在考虑但基本上还没做的事情。嗯，早期的 GB200 坏了。早期的 H100 坏了。

Yeah a wide area network. So um basically cables that connect the clusters together. A private internet is another way to think about it. Um uh lot of clusters that are being spun up today are like they're not attached to um traditional CPU clusters were um next to the customer for latency reasons. In GPUs they're next to the power. Um wherever the like you know you put the data center next to the power rather than next to the customer. Um, this means that a lot of stuff is getting put in places that previously they weren't expecting data centers there. And so, you may not have um, as um, excellent of a transit path um, from one to the other. So, this this starts to matter. Um, other examples are um, weird where we theoretically support but don't actually um, uh, like, this is a thing that we're thinking about but haven't done basically. Um, the early GB200s broke. The early H100s broke.

所以，如果你要买 H100 或 GB200，你可能真的想要那些在特定时间之后部署的，因为你不会拿到第一批从机架上下来的 GD1。所以，如果你是一个非常精明的客户，你可能愿意支付更多。

And so, if you were buying H100s or you were buying GB200s, you might really want the ones that were deployed after a certain time because you're not getting the first GD1s off the rack. And so, if you're a super sophisticated customer, you may be willing to pay more.

哦，所以这是关于它们生产时的批次，而且它们会随着时间的推移变得更好？

Oh, so it's about what cohort they were produced and they get better over time?

所以，就像，第一批从机架上下来的可能有一些故障或问题。对。所以，如果你是拿到第一批的人，或者第一批发货中的第一批，那么是的，它们会坏掉或以某种方式出问题。

So, like, the first ones off the rack maybe had some faults or some problem. Right. And so, if you were the ones who got the first ones, or first ones from the first shipment of them, then yeah, they would be broken or in some way.

我假设那种过程就像，看，到处都有缺陷。就像，它就是会发生，对吧？

I assume that sort of process is like, look, there's defects everywhere. Like, it just happens, right?

完全正确。

Totally.

所以，就像，你从台积电或其他地方拿到新一批货，你就运行测试，然后你找出哪些是坏的，然后把它们扔掉。难道不是这样运作的吗？我之前不知道这种批次效应，比如说 2026 年第四季度的 B300 比 2027 年的 B300 差得多。

And so, like, you get a new shipment off from TSMC whatever and you just kind of run your tests and like, you figure out which ones are bad and you throw them out. And isn't that how it works? And I didn't know about this element of cohorting where like, let's say Q4 2026 B300s are way worse than 2027 B300s.

是的，我举个例子。你描述的是老化测试。是的。所以，当你得到一个集群时，一个好的云服务商会尝试在 GPU 到达客户手中之前杀死所有 GPU。然后，坏掉的 GPU 或节点会进行 RMA。你基本上会去更换它们。是的，你退回它，然后得到一个新的，这有一个交货时间，但不管怎样，你会得到一些回报。然后客户永远不必经历这些。我指的是，在早期的 H100 中，存在固件错误，这些错误甚至可能与特定的无关，它们可能不是 Nvidia 的。可能是 Supermicro 在你购买的机器中有固件错误。而且它适用于所有机器。解决方案是，你需要让他们更新固件，或者他们需要在每个新机器中提供新的东西。

Yeah, so I'll give an example. You're describing burn-in. Yes. So, when you get a cluster, a good cloud, what it does is it tries to kill all the GPUs before they ever get in the hands of the customer. And then the GPUs that are broken or the nodes that are broken, you get RMAd. You basically go and replace them. Yeah. You return it and you get a new one and there's a lead time to that, but whatever, you'll get some back. And then the customer never has to experience this. The thing I'm referring to is that like in early H100s, there were like firmware bugs that may not even be related to the specific, like they may not be Nvidia. They might be like Supermicro has a firmware bug in the box that you bought. And it applies to all of them. And it doesn't matter like the solution to that was you need to get them to update their firmware or they need to ship a new thing in every new box.

是的。或者 Cornelis 对我们拿到的 B200 也有这个问题。然后你必须使用 DevCloud 去修复它。

Yeah. Or Cornelis had this for like the B200s that we got. And then like you had to use DevCloud to go fix it.

是的，正确。所以像这样不同类型的问题都可能发生。

Yes, correct. So like different types of problems like this can happen.

有趣。好的。所以我觉得，与整体相比，这似乎很小。我只是想让大家理解这个市场的形态。你提到了抢占，你知道，我们谈到了现货实例以及类比如何。然后还有，你知道，提前一个月预订。这些市场能延伸到多远，实时或抢占对你来说更有趣，还是更远的未来？

Interesting. Okay. So I'm okay, that seems small compared to the overall. Like I just want to like I want people to understand that the shape of this market. You mentioned like preemption you know and and you know we talked about spot instances and how how that the analogy is. And then there's also, you know, booking a month out. How far out do does do these markets go and is is real-time or sort of preemption more interesting to you versus uh far being further further out?

回答你的第一个问题，3 到 5 年。

To answer your first question, 3 to 5 years.

哇。

Wow.

我们开始看到

We're starting to see

逐月还是逐季度，或者怎么样？

month by month or quarter by quarter or how's this?

你是在问现金流如何运作吗？

Are you asking what how the cash flow works?

合同。就像期货合同，对吧？再说一次，我是前合同衍生品交易员。

The contracts. Like the few the futures contracts, right? Again like I'm a former contracts derivatives trader.

在市场本身

In the market itself

是的。

Yeah.

就像在更广泛的市场中。

like is in the broader market.

是的。

Yeah.

集群融资 3 到 5 年。

Clusters are financed 3 to 5 years.

好的。

Okay.

所以

So the

付款

The payments

是在那段时间内。

are over that period of time.

是的。

Yeah.

还有像容量块在我们的订单簿中，可以延长任意时间。

There's also like blocks of capacity is in our order book which can extend any arbitrary time.

是的。

Yeah.

关于现金流，我们有一些事情要做。现在我们所做的就是要求人们为容量块全额预付。对于租赁、长期合同，我们不要求他们这样做。这涉及到我肯定需要白板来解释的东西。但你还有一个问题我正要回答，我忘了是什么

There's some stuff that we have to do for the cash flow on that. Right now what we do is we just ask people to pay entirely up front for a block. We don't ask them to do that for a lease, for a longer-term contract. This gets into something that I definitely need a whiteboard to explain. But you had another question I was about to answer, I forgot what it

呃，你想要实时还是你想要，比如，你想要更多，呃，什么对你更有趣，实时还是更远的未来？

Uh, do you want real-time or do you want Like, do you want more, uh, what's more interesting to you, real-time or long longer out?

所以，我们更关心更远的未来，因为我们试图建立一个具有大量流动性的远期曲线。这些有点相同，也有点不同。所以，我们所做的是，比如说你可以买一小时后的，你可以买现在到下一小时。或者你可以买现在到未来 3 年以及之间的任何时间。

So, we care more about longer out because we're trying to build a forwards curve that has a lot of liquidity on it. These are kind of the same thing and kind of not. So, what we do is we since say you can buy an hour out, you can buy right now to the next hour. Or you can buy right now to the next 3 years and anywhere in between.

太疯狂了。

That's wild.

我知道。所以，或者技术上更精细。每个客户的情况有所不同。基本上，比如，如果我有一个批处理推理的工作负载，我可以在任何时候被中断，我不太在乎。那么，我被切断的能力大概是，我需要一个小时的时间来知道我将被切断，然后我就没事了。或者我甚至可以在几分钟内被通知。但如果我有实时推理客户，他们期望我在线，那么，你可以中断我，但要在一个月后。也许两个月，也许三个月。然后我可以在那段时间内适应那个容量。但你不能在一分钟内中断我。所以，这两件事的定价不同，定价不同的原因是，你愿意支付什么价格来中断你的批处理作业？这与你中断实时作业的价格非常不同。如果你有实时作业，当你为客户中断时，你可能要支付 SLA。这合理吗？

I know. And so, or finer technically. There's a certain difference that happens that's dependent for each customer. Basically, like, if I have a workload that is batch inference, basically, where I can be interrupted at any point, I don't really care. Then, my ability to be cut off is probably like, I need an hour of time to know that I'm going to be cut off and then I'm good. Or I can even be notified in a minute or two. But if I have live inference customers who are expecting me to be online, then, you can interrupt me, but in a month. Maybe 2 months, maybe 3 months. And then I can adapt to that capacity over that period of time. But you cannot interrupt me in a minute. And so, those two things are priced differently and the reason why they're priced differently is because what is the price you would pay in order to interrupt your batch job? It's very different than the price you would pay to interrupt your live job. If you have a live job, you are probably paying an SLA when you break for your customer. Does that make sense?

是的，我明白了。所以，这实际上是打破租约的市场。

Yeah, I see. So, it's This is actually the market for breaking leases.

基本上，这些是打破租约的市场。

These are the market for breaking leases, basically.

天啊。

god.

然后还有未来租约的市场，或者像租约

And then there's also the market for future leases, or like for leases that are

哦，我以为这是像转租的情况。这是像哦，租约中断。

Oh, I thought this was like a subletting situation. It's This is like Oh, the lease break.

所以，一种思考方式是，在电力市场中，实际上有两个市场。有日前市场和实时市场。计算也有类似的情况，有即将上线的容量，或即将到达你手中的容量，然后是当前正在使用的容量。这些有不同的价格动态。

So, one way to think about it is in power markets, there's actually two markets. There's like the day-ahead market, and there's the live market. Compute kind of has the same thing, where it has capacity that is coming online, or capacity that's about to arrive to you, and then capacity that is currently in use. And these have different price dynamics.

好的，有趣。呃，现货溢价还是期货溢价？

Okay, interesting. Um, backwardation or contango?

现货溢价还是期货溢价？

Backwardation or contango?

呃，所以，更远的计算价格是高于现货还是低于现货？

Do uh, so do uh, is compute further out priced higher than spot, or lower than spot?

呃，更远的计算价格是，它是像这样，还是像那样，在，呃，你知道，在 1 年、1 年、2 年之后？

Um, is compute further out priced Does it Does it go like this, or does it go like that over, uh, you know, it over, like, 1-year 1-year out 2-year out?

是

Do

哦，这是一个例- 两者在不同

Oh, this is an ex- Both at different

我觉得你不知道这件事很有趣。

I find it fascinating that you don't know.

所以你在问：算力是现在卖，也许不是今天，也许一个月后，对比一年后、两年后。那条曲线是上升还是下降？

So you're asking: is compute sold right now, maybe not today, maybe a month out versus 1-year out versus 2-year out. Is that curve going up or going down?

我们看到一个驼峰。

We see a hump.

我的天。解释一下。

Oh my god. Explain.

我觉得这很直观。基本上就像，如果我有一批牛奶快过期了，我会降价。我们看到的基本上就是这个。

I think this will make intuitive sense. It's basically like, if I have milk, and it is about to expire, I will lower the price on it. And so, that's basically what we see.

所以基本上就是租约周期的末尾，对吧？

So it's the end of the lease cycle, basically, right?

我们更想说的是，如果有一块算力，而且它现在临近了，它不能在 past 卖出，所以它必须降价才能以某个价格出清。所以一开始你会看到这个下跌，现在卖的东西，现在市场上任何东西，都是还没卖出去的，因此价格会下跌。然后它上升到它想要的位置。然后也许未来的算力价值更低，因为人们想要现在的算力，所以他们在为更短的交货时间付费。而对于未来更远的算力，他们付得更少。我认为这就是一个正常健康的曲线。

Well, we're more saying like, if there's a block of compute, and it's approaching now, it can't be sold in the past, and so it has to drop its price in order to clear at some price. And so, initially, you see this drop, where stuff that's sold right now, anything that's on the market right now, is something that hasn't been sold, and so therefore the price will drop. And then it goes up to where it wants to be. And then maybe compute in the future is worth less because people would like the compute now, and so they're sort of paying for the shorter lead time. And they pay less for the compute that's off in the future. This is what I think a normal healthy curve looks like.

是的。

Yeah.

让我觉得“哦天哪，不是吧”的地方是，现在如果你拿一块 B300 试着跑推理，你卖出去每小时能赚 40 美元左右。

The place where I'm like, 'Oh man, no shit.' is that right now if you took a B300 and you tried to run an inference on it, you can make $40 an hour or so selling that.

任何模型？

Any model?

基本上任何模型。所以那是如果你 100% 使用它，也许如果你只有 50% 左右的利用率，那就是每小时 25 美元。

Any model, basically. And so that's if you use it at 100%, and maybe if you only have like 50% utilization or something, so like $25 an hour.

是的。

Yeah.

所以每小时 25 美元。与此同时，你可以以每小时 5 到 6 美元左右的价格买到 B300。其中一个数字必须改变。

So, $25 an hour. Meanwhile, you can buy B300s at like $5 to $6 an hour. One of these numbers has to change.

是的。这是巨大的超额收费。

Yeah. It's a huge overcharge.

这是巨大的超额收费。

It's a huge overcharge.

你在演讲中经常提到这一点。

You mentioned this in your talks a lot.

是的。所以，很多人基本上看着这个在想，如果我认为价格会上涨，我可能愿意在未来支付比之前想的更多的钱，因为我可能会想，“好吧，我现在可以以 5 美元买入，但如果我不能再买了，那么也许未来我愿意支付更多只是为了锁定它，如果它以后上线的话。”所以我们也看到了另一面。这取决于你是谁。

Yeah. So, a bunch of people are basically looking at this and thinking if I think prices are going to go up, I might be willing to pay more than what I previously thought in the future because I might think, 'Okay, I can buy at five right now, but if I can't buy any more, then maybe I will be willing to buy even more in the future just to lock it in if it's going to come online later.' So, we see the other side as well. Kind of depends on who you are.

酷。所以，是——好吧，我只想非常清楚地说明这一点。你关心不同代的芯片吗，还是只关心总的 flops，然后你把它们分配到任何你想要的芯片上？

Cool. So, is it—okay, and I just really want to make super clear on this. Do you care about the different generations of chips, or do you just care about overall flops and you spread them among whatever chips you want?

我们关心不同代的芯片，因为——

We care about different generations of chips because—

你关心。我想我只有——我猜你的客户关心吗？

You care. I think like I only—I guess do your customers care?

是的。

Yes.

对。哦等等。当你说不同代的芯片，你是指像——

Right. Oh, wait. When you say different generation chips, do you mean like—

比如 B300 代。是的，没错。

So, like B300 generation. Yeah, exactly.

是的。或者甚至 AMD 之类的，对吧？

Yeah. Or like even AMD or whatever, right?

是的，我认为人们完全关心芯片的代际。我认为部分原因是像 VRAM 这样的东西。部分原因是像更大的 NVLink 域。诸如此类。

Yes, I think people totally care about the generation of chips. I think part of this is because of things like VRAM. And part of this is because of like larger NVLink domains. And things like that.

所以，根本没有办法把这一切抽象成某种通用的算力单位，你可以随便揉捏之类的。

So, there's just no way to abstract all this to like some generalized unit of compute that you can sort of mush around and whatever you like.

我觉得就像你说石油，甚至石油也不是石油。石油实际上只是一堆不同类型的油。

I think like when you were saying oil, even oil isn't oil. Oil is actually just like a bunch of different types of oil.

是的，像喷气燃料或汽车燃料之类的。

Yeah, like jet fuel or car fuel or whatever.

我认为是一样的。你会得到一堆不同类型的——

I think it's the same thing. You're going to get a bunch of different types of—

是的，但你有炼油厂，你可以有效地把石油放进黑箱，然后得到任何东西。如果算力如此异构，我猜这个词是，你只需要把底层复杂性暴露给最终客户，这对你作为做市商是不利的。

It is, but you have oil refineries that you can effectively a black boxes you can put oil in and get it like whatever out. If compute is so heterogeneous, I guess the word is that you just have to expose the underlying complexity to your end customer, which is bad for you as a market maker.

也许。

Maybe.

就像你想确定一个瓶颈点，或者我不知道瓶颈点的更好词是什么，就像谢林点，每个人都同意，好吧，这算是一种标准，我们都会同意这个标准，你知道它会有些低效。

Like you want to identify the one like choke point or like I don't know what the better word for choke point is like shelling point where everyone agrees like okay like this is kind of the standard and we'll all agree to this standard and you know it'll be somewhat inefficient.

是的。所以我们已经能够把问题减少到一定程度。

Yes. So we've been able to reduce the problems to a certain level.

是的，就像你的工作就是降维之类的。

Yeah, it's like your job is like dimensional reduction or all that.

我们减少了很多维度，但仍然有维度。它们仍然会存在。

We reduce a lot of the dimensions but nonetheless there are still dimensions. They can still going to exist.

你知道你有多少 SKU 吗？

Do you know how many SKUs you have?

所以——

So—

就像 5 乘以 100 乘以 5,000，那等于我不知道。就像什么是——

Like 5 * 100 * like 5,000 that equals I don't know. Like what's—

我们实际上把 SKU 做成了一个概念，目标是让别人可以创建自己的 SKU。所以如果我想买 Blackwell，我可以说我愿意以每小时 6 美元或更低的价格购买 Blackwell。你可以把我匹配到 GB300 或 B300，我不在乎是哪种 Blackwell，只要低于每小时 6 美元。

We actually made SKUs like a concept and the goal is to make it so that somebody else can make their own SKU. So if I want to buy blackwells, I can say I'm willing to buy blackwells at $6 an hour or less. And you can match me against GB300s or you can match me against B300s and I don't care which blackwell as long as it's less than $6 an hour.

但你刚才说你在意内存和其他东西。

But you just said you care about memory and other stuff.

完全正确，但不是每个客户都这样。所以有些客户就是可以在 Blackwell 上工作，不管是什么。有些客户非常非常具体，他们需要确切的东西，他们愿意为此支付更多，因为市场会向他们收取更多。这就是我们做的。

Totally, but not every customer does. So some customers like just can work on blackwells whatever it is. Some customers are super hyper specific and they need this exact thing and they're going to pay more for that because the market's going to charge them more. And that's what we do.

所以人们使用你的 API 以编程方式定义他们想要什么，然后如果将其与任何卖家组合匹配，他们可能能够满足。

So people use your APIs to programmatically define what they want and then if match that with whatever composition of sellers they might be able to fulfill.

是的。但需要注意的是，命名查询是我们正在做的新事情。但查询系统作为一个概念，就是当前订单簿的工作方式。

Yes. With the caveat that having a named query is a new thing we're doing. But the query system as a concept is how the order book works right now.

好的。有趣。好吧。所以，我认为这是很多细节。你有没有——我猜，有什么你想教育人们的事情，无论是金融人士还是工程师，关于你们开发的算力交易市场的现状，因为实际上你们在定义它。

Okay. Interesting. All right. So, I think that's like a lot of nitty-gritty. Have you—is there anything about, I guess, that you want to educate people on whether it's financial people or engineers about the state of the compute trading market as you've developed it because effectively you guys are defining it.

主要是实物结算的要求。我们认为如果你——我们花了 3 年时间学习为什么需要实物结算。而我们今天整个对话就是，这些东西并不像你想象的那么可替代。我们将尝试使它们可替代。这是行业的工作。但如果没有这一点，你会面临很多风险。

Mainly just the requirement for physical settlement. We think that if you—we spent 3 years learning why you need physical settlement. And our entire conversation today is that these things are not as fungible as you think. We're going to try and make them fungible. That's the industry's job to do that. But without that you run a lot of risks.

如果你在行业还没来得及建立实物结算之前就试图过度金融化，那么我认为你会得到一个非常容易被操纵的市场，以及一个存在大量基差风险的市场，你用来对冲的价格实际上与你真正能买到的产能价格毫无关系。

If you try to build too much financialization before the industry has had time to build out the physical settlement, then I think what you have is a market that's very easy to manipulate and a market that has a lot of basis risk, where the price you may be hedging against actually has nothing to do with the price that you can actually buy the capacity for.

是的。

Yeah.

是的。

Yeah.

所以对于听众来说，我们上次简要谈过这个，但人们常常震惊地发现，比如说，信用违约互换的市场规模是其保险债券的 12 倍。类似这样，或者像抵押贷款再保险的市场等等——它比标的实物结算要大。所以这就是实物结算与现金结算的区别。

So for listeners, we talked about this briefly last time, but people are often shocked to find out that, let's say, the market for credit default swaps is 12 times the size of the bonds that they insure. Something like that, or like the markets for mortgage reinsurance or all that—it is larger than the underlying physical settlement. So that's why physical versus cash settled.

是的。基本上，现金结算正是你在《大空头》中看到的那种旁注，它们可能与标的物本身有关，也可能无关，并且可能被用作其他整体事物的代理。你不希望这种情况失控。

Yeah. And basically cash settlements are exactly the kind of side bet that you see in The Big Short, where they may or may not have anything to do with the underlying thing itself, and it may be used as a proxy for overall other stuff. And you don't want that to go out of control.

是的。是的。所以，明确地说，我们认为现金结算作为一个未来概念是个好主意。我们只是认为它应该从实物结算中衍生出来。因为，正如你所说，它是一种旁注，这有点像你在做实物结算之前就做了现金结算期货。

Yes. Yeah. So, to be clear, we think cash settlement as a future concept is a great idea. We just think it should be derived from physical settlement. Because, as you said, it's a side bet, and it would be kind of like if you made a cash settlement future before you made physical settlement.

我认为人们已经在某种程度上这样做了。

Which I think people are kind of doing already.

还没有完全。人们试图在实物结算之前就进行现金结算。

Not yet, totally. People are trying to make cash settlement before you make the physical settlement.

对各种东西都这样做，包括初创公司股票。

Do that for all sorts of things, including startup stock.

是的。

Yes.

对吧？就像你可以买卖一个虚构的 OpenAI 股份，但它与实际股份无关。

Right? Like you can buy and sell a made-up share of OpenAI, but it's nothing to do with the actual share.

是的。而且很多这样的交易结果很糟糕。就像如果你做了一个——如果你开始对一个还不存在的股票出售衍生品，那会很糟糕。

Yes. And lots of those go very poorly. Like it would be kind of bad if you made a—it would be bad if you started selling the derivative on a stock that doesn't exist yet.

是的。是的。

Yeah. Yeah.

那么，我的意思是，我们做的一件事——所以我曾经是一名货币交易员，对吧？还有利率。我们做的一件事是，有一个董事会聚集在一起说，这是伦敦银行同业拆借利率，LIBOR，我们都同意这是每月定盘价，我们都按照那个定盘价交易。这在 3 年后的某个时候还有意义吗？

Well, so, I mean, one thing that we do—so I used to be a currency trader, right? And interest rates as well. And one thing we do is you have a board that gets together and says like this is the London Interbank Offered Rate, LIBOR, and we all agree like this is the monthly fixing and we all trade to that fixing. Is that even meaningful at some point in 3 years from now?

所以，关键的是，LIBOR 被操纵了。

So, critically, LIBOR was manipulated.

是的。

Yes.

我们目前的担忧是——

Our concern at the moment is—

12 家银行。是的。

12 banks. Yeah.

是的。嗯，我们的担忧是存在一个非常相似的问题，你可以操纵一个指数。我们的印象是，为什么你可以操纵那个指数，是因为很多指数基于非常少量的数据，而这些数据与购买新储备合同的实际价格没有太大相关性。这就是我们所担心的。

Yes. Well, our concern is there's a very similar problem, where you can manipulate an index. Our impression is why you can manipulate that index is because a lot of the indexes are based on a very small amount of data that has not a lot of correlation with the actual price of buying a new reserve contract. That's what we're concerned about.

好的。有趣。好吧。我认为这就是大致的情况。我想今天的另一部分是你与 Oliver 交谈过，你正在与 Jump 举办活动，并且你与 Silicon Data 是朋友。你从所有这些其他参与者那里学到了什么？比如你想指出哪些主要参与者正在做有趣的事情？

Okay. Interesting. All right. I think that's the rough state of things. I think the other part of just today is you talked with Oliver, and you're doing an event with Jump, and you're friends with Silicon Data. What are you learning from all these other players? Like who are the major players that you would want to call out as doing interesting things?

哦。所以，我们是 Silicon Data 和 Computer Change 的忠实粉丝。我认为他们也很早进入这个领域，我认为他们在 A 路径上做出了合理的尝试，我们认为我们最终可能会成为合作伙伴。

Ooh. So, we're big fans of Silicon Data and Computer Change. I think they were also very early to this, and I think they are doing a reasonable shot at the A path that we think that we'll likely end up being partners with.

你能区分一下你们不同的路径吗？

And can you distinguish like your different paths?

所以，大体上是他们做现金结算，我们做实物结算。我们认为两者都可以存在，也应该存在，并且它们彼此互利。所以，完全没问题。所以，我认为他们是这个领域最早的之一，我认为他们显然比任何人都更早下了赌注，我认为这非常值得尊敬。

So, the general gist is they're doing cash settlement, we're doing physical settlement. We think both can exist and should exist, and that they're sort of mutually beneficial to each other. And so, totally fine. So, I think they are one of the earliest in the space, and I think they very clearly took the bet well before anybody else did, and I think that's very honorable to do.

你提到了 John Seminalysis 说 Jump。这些其他参与者是否也在涌现——就像你是我第一次接触这个，然后我看到所有这些人进来，我想，‘好吧，这里正在发生一些事情。’我不像你那样清楚地了解这个领域。

You mentioned John Seminalysis says Jump. Like are these other players emerging like as far—like you were my first introduction to this, and then I see all these other guys coming in, and I'm like, 'Okay, something's happening here.' I don't know the landscape as clearly as you do.

是的，所以我们看到的类别是，有些人试图制定更好的标准。所以，我认为 Seminalysis 正在努力创建更好的标准。有些人试图做残值保险等等。这个领域有许多不同的公司。我对它们都不是很了解，但我知道这是一个概念，我认为这很重要。有些人试图做存款贷款，他们将为云提供信用风险。有一个公司生态系统正在围绕如何解决计算的财务瓶颈形成，这很酷。

Yeah, so the categories of folks that we've seen is there's folks that are trying to set better standards. So, Seminalysis I think is working to try and create better standards. There's folks that are trying to do residual value insurance and so on. There's a number of different companies in the space. I don't know any of them super well, but I know this is a concept, and I think that's important. There are folks that are trying to do deposit lending where they'll front your credit risk for the cloud. There's an ecosystem of companies that are forming around how do we resolve the financial bottlenecks of compute, and it is cool to see.

我认为在这个生态系统中，你有效需要的是买家和卖家，你需要一个分析师，你需要——我们说过保险还是什么——

I think effectively what you want in this ecosystem is buyers and sellers, you want an analyst, you want—we said insurance or what's the—

呃，残值保险。

Uh residual value insurance.

残值。嗯，你希望有一个健康的生态系统。但显然你们正在解决实物结算的问题，这真的很有趣。好的，所以我的意思是，我们可以进入硬科技方面，因为显然你们也在招聘极其顶尖的工程师。描述一下你们的技术栈。

Residual. Um, and you want this like healthy ecosystem out there. But you obviously you guys are owning the problem of physical settlement, which is really interesting. Okay, so I mean, we can go into like the hard tech side because obviously you are also hiring for extremely cracked engineers. Describe your stack.

我们做云做的所有事情，但不是云。也就是说，我们不拥有硬件。我们不拥有硬件的原因是因为如果我们拥有，我们就能操纵远期市场。所以，我们必须保持中立，必须独立，这给公司带来了很多奇怪的事情，但为了进行实物结算，你必须进行实地考察。你必须参观数据中心。你必须设计物料清单并确保——你必须基本上做供应链咨询，与不同的 OEM 建立关系，确保按时交付。然后你必须做集群架构，所以你在设计集群本身。然后你部署机架，或者与合作伙伴一起部署机架，然后监督。

We do all the things the cloud does without being the cloud. As in we don't own the hardware. The reason why we don't own the hardware is because if we did, we would be able to manipulate the forwards market. So, we have to be Switzerland, we have to be independent, which creates a lot of sort of strange things about the company, but in order to do physical settlement, you have to do site visits. You have to visit the data center. You have to design the bill of materials and ensure that—you have to basically do like supply chain consulting, where you have relationships with different OEMs that ensure that things get delivered on time. Then you have to do cluster architecture, so you're sort of designing the cluster itself. Then you're rolling out racks or you're working with partners to roll out racks and then supervising that.

然后你们做背板管理，所以你们接管了 PDU、交换机、DPU、UFM 和 BMC，以及围绕你们所部署节点的运营框架。

Then you're doing backplane management, so you're taking over the PDUs and the switches and the DPUs and UFM and BMC and the operating harness around the nodes that you're settling.

你作为 SF Compute 的员工？

You as an SF Compute employee?

是的。

Yeah.

哇。

Wow.

好吧。所以当你从云提供商那里购买一个节点时，有裸金属，但裸金属周围还有一大堆东西。为了让那个裸金属节点工作，你必须运营其他所有东西。

All right. So when you buy a node from a cloud, there's the bare metal, but there's all this around the bare metal. And in order to make that bare metal node work, you have to operate everything else.

因为传统上数据中心的工作就是帮你把其中一些东西抽离出来。

Because I just like traditionally the job of the data center is to extract some of this away from you.

是的。我们就是做这个的。

Yeah. We're the ones that do that.

对对。好的，明白了。但你们不在数据中心业务里。你们不是像 Equinix 那样的公司。

Yeah, yeah. Okay, all right. But you're not in the data center business. You're not the Equinixes of the world.

我们不算是在数据中心业务里。我们做的是——我们提供超级计算即服务。这意味着人们来找我们说，你知道，我有 X，X 就是我有 X 兆瓦，有时他们的意思是，我有一个数据中心，我想把它变成 GPU 集群。我的选择是，我可以卖掉 GPU 集群，或者把数据中心卖给别人，或者我自己成为云。然后我们去找他们说，太好了，我们会把你变成云，我们是超级计算部分，你是数据中心云。我们合作，然后我们一起销售，一起赚钱。有时是那些拥有可以变成数据中心的东西的人，在这种情况下，我们基本上会项目管理到那个层面，然后我们会与那些人合作，通过合作伙伴将他们有电力的土地或任何他们拥有的东西变成数据中心，然后我们在超级计算部分接手。这要求我们在内部拥有某种数据中心专业知识，但我们不自行建造。所以是的。

We're not in the data center business, sort of. What we do work with is—so, we offer supercomputing as a service. And what this means is folks come to us and say, you know, I've got X, where X is I've got X megawatts here, and maybe sometimes they mean, I've got a data center that I would like to turn into a GPU cluster. Like my options are I could sell the GPU cluster or I could sell the data center to somebody or I could become the cloud. And then we go to them and say, great, we will turn you into the cloud, we're the supercomputing part, you're the data center cloud. Let's partner together and then we will sell and all make money. Sometimes it's people who have something that could turn into a data center, in which case we're going to basically project manage that up to the layer, and then we will work with those folks to turn their powered land or whatever they have into a data center by working with partners, and then we take over at the supercomputing part. This requires us to have some sort of data center expertise in-house, but we don't self-perform builds. So yeah.

是的，所以这有点模块化，因为你们需要在不同的场景下部署。

Yeah, so it's somewhat modular because you have different scenarios you need to deploy under.

是的，有人曾把我们描述为计算领域的 Palantir，因为每次都是新的挑战，但我们与双方的人都深度合作。

Yeah, someone once described us as like Palantir for compute because it's like these each time it's this new rodeo, but we deeply partner with folks on both sides.

是的。主要在美国？

Yeah. And primarily the US?

主要在美国，但我们在全球运营。

Primarily the US, but we operate around the world.

是的。到了某个时候，SF Compute 会不会变成一个误称？

Yeah. At some point does SF Compute become a misnomer?

所以，我们正在全球部署集群，当我们这样做时，我们与那些自己也想成为云的人合作。他们看起来像主权计算参与者。我们对主权计算的推销很棒。让我们帮助你在自己的国家建立云，我们将带来旧金山的承购。我们将带来旧金山的专业知识。我们将建立集群。这就是我们的全部业务。我们将使其可融资。

So, we're deploying clusters all around the world, and when we do that, we're working with people who themselves are trying to be clouds all around the world. They look like sovereign compute players. And our pitch to sovereign compute is amazing. Let's help you set up a cloud in your own country and we will bring the San Francisco off-take. We'll bring the San Francisco expertise. We'll set up the clusters. That's our whole business. We're going to make it financeable.

啊。你能说出你们宣布的主要最大合作伙伴吗？

Ah. Who are your main biggest announced partners that you can name?

基本上没有我能说的。我们预计在一两周内会有一些公告。

Nobody I can name basically. We will have some announcements in a week or two that we expect to come on.

是的。所有事情有点——当然，当然。祝好运。是的。不，我的意思是，我认为这对每个国家来说都非常有趣，比如谁在投资主权 AI 来解决这个问题，有趣的是，SF 仍然有点像那个品牌名称，而且非常有市场价值。

Yeah. All things sort of uh—Sure, sure. Good luck. Yeah. No, I mean I think this is really interesting for every country to like who's investing in sovereign AI to figure this out and it's interesting that like SF is still sort of like that brand name and that is very marketable.

所以我认为有趣的事情之一就是思考这个市场。

So I think one of the interesting things is just thinking through this market.

嗯。英伟达显然是房间里的大猩猩。他们为什么不——为什么他们没有做出 SF Compute？再说，你知道，答案会是市场操纵，但让我们真正深入探讨多个参与者的激励因素。

Mhm. Nvidia is obviously like the 800-lb gorilla in the room. Why don't they—Why didn't they make SF Compute? And again like I you know, the answer's going to be market manipulation but like let's just really get into the incentives of the multiple players.

是的，所以我们的印象是，英伟达无法做到，但从它的存在中获益巨大。他们做不到的原因是他们不够独立，人们无法在上面交易。如果你不在上面交易，那么你就没有足够的流动性让贷款人针对远期市场融资。这也是为什么 Crusoe、Coreweave 和 Lambda，为什么超大规模厂商做不到，为什么 Base10、Fireworks 和 Modal 做不到，因为他们在那一边有太多需求。这是对整个行业有益的事情，因为它降低了融资成本。但没有人能做到，除非他们放弃赚取所有大额推理收入和大额云交易的能力。你必须扼杀你的业务才能做到这一点。这就是为什么没有人做。就像结构上他们不是这样设置的。而你基本上是在减少摩擦，这就是你所有的收入。

Yeah, so our impression is that Nvidia can't make it but benefits tremendously from it existing. The reason why they can't make it is because they would then not be independent enough for people to trade on it. If you don't trade on it, then you don't have enough liquidity for lenders to finance against the forwards market. This is also why Crusoe and Coreweave and Lambda, why the hyperscalers can't make it, why Base10 and Fireworks and Modal can't make it cuz they have too much demand on that side. This is the thing that benefits the entire industry because it makes the financing costs go down. But nobody can make it unless they give up the ability to do all the big inference money and all the big cloud deals. You have to kill your business in order to make this. That's why nobody does it. It's like structurally they're not set up to do that. And you're basically reducing friction and that's all your revenue.

你们宣布了一些基准，一些里程碑。实际上，我忘了在前面说。你能重复一下它们是什么，并稍微炫耀一下吗？

You've announced some benchmark, some milestones. Actually, I forgot to say that up front. Can you repeat what they are and just take a victory lap a little bit?

我们完成了数亿美元的交易量。我们有许多新集群即将到来。然后我认为我最兴奋的事情是我们发布了一大堆新功能，使 SF Compute 对平台即服务公司非常可行。一个例子是我们允许客户细分他们的 InfiniBand 网络结构，而无需询问我们。所以，许多云当你去购买时，他们会给你一个 InfiniBand 网络结构。然后如果你想重新划分那个 InfiniBand 网络结构的线路，你需要联系他们，然后他们会为你做些什么。我们现在可以实时为客户做到这一点，我认为这真的很棒，因为它与市场很好地接口，你希望能够重新划分东西以扩大和缩小你的集群。

We closed several hundred million dollars of volume. We have a significant number of new clusters on the way. And then the things that I think I'm most excited about is we've shipped a whole bunch of new features that make SF Compute very viable for platform as a service companies. An example is we allow customers to subdivide their InfiniBand fabric without having to ask us. So, many clouds when you go to buy them, they will give you one InfiniBand fabric. And then if you want to redraw the lines of that InfiniBand fabric, you need to contact them and they will then do something for you. We now can do that on the fly for customers, which I think is really great because it interfaces with the market really well, where you want to be able to redraw things in order to scale up and down your clusters.

你会把这描述为你们正在虚拟化数据中心并使其可编程吗？

Would you describe this as like you're virtualizing the data center and making it programmable?

是的，我们正在虚拟化数据中心并使其可编程。所以，如果你拥有硬件，你会得到的所有东西，我们正试图虚拟化，我认为这对我们来说很令人兴奋。

Yeah, we're virtualizing data center and making it programmable. So, all the stuff that you would have gotten if you had owned the hardware, we're trying to virtualize that, which I think is exciting to us.

有趣的是，上次发生这种情况就像云一样。就像——是的。这真的很酷。然后最后我想你对人们的行动号召是什么，你知道，是工程师吗？是像给我你的供应？像是什么？

It's funny like the last time this happened was like the cloud. Like the—Yeah. Which is really cool. And then finally I think what is your call to action for people, you know, is it engineers? Is it like give me your supply? Like what is it?

如果你想拥有一个 GPU 集群，并希望从中获得剩余价值，你想把它纳入你的投资组合，我们很乐意与你合作构建那个 GPU 集群。我们将接管并完成其中的超级计算部分，确保事情顺利进行。我们拥有 SLA。

If you would like to own a GPU cluster and you want the residual value from it, you want to have it for your portfolio, we would love to work with you in order to build that GPU cluster. We will take over and we will do the supercomputing portions of that and ensure the things go right. We own the SLA.

然后如果事实证明那是个坏主意、你需要退出那个集群，我们可以立刻把那个集群卖给别人，这样就不会冲击你的资产负债表。

And then if it turns out that that was a bad idea, that you need out of that cluster, we can immediately sell that cluster to someone else so it doesn't hit your balance sheet.

非常简单的推介。好了，非常感谢你抽出时间。你在为这个行业做很大的贡献，我很高兴能成为你的朋友。

Very simple pitch. Well, thank you so much for your time. You're doing a great service to the industry, and I just am happy to be your friend.

谢谢。

I appreciate it.

好的，就到这里。

All right. That's it.