AI 实验室的计算主导地位与经济

好的，我再次请到了 SemiAnalysis 的创始人 Dylan Patel。我们每年一次的播客，就像家庭感恩节晚餐一样。但我们其实没有血缘关系。别告诉别人，这会打破神话。基本上，世界经济的走向越来越取决于实验室经济的走向、算力市场的走向等等。我想了解几年内这个疯狂未来会走向何方。但让我们从今天开始。请带我了解一下目前实验室的算力和收入情况，并预测未来一两年。

Okay, I’m back with Dylan Patel, founder of SemiAnalysis. Our version of a family Thanksgiving dinner is a regular yearly podcast. But we're not actually related. Don’t tell the people this. It will destroy the myth. Basically where the world economy is headed is more and more becoming a function of where lab economics are headed, where the compute market is headed, et cetera. I want to understand where the crazy future ends up within a few years. But let’s start with where we are today. Walk me through lab compute and lab revenue right now, and maybe project out a year or two.

回顾去年，即使到年底，美国 GDP 增长的大部分都只是 AI 基础设施。展望今年，新增算力中约有三分之一流向了实验室，即 OpenAI 和 Anthropic。这些算力可能是由其他公司建造然后租给它们的，但最终客户是它们。展望未来，算力数字正在急剧膨胀。今年我们的资本支出略超 1 万亿美元。到 2028 年，将超过 2 万亿美元。实验室所占的比例也在不断增加。所以最终，你会看到一个非常有趣的情况：实验室从每年花费数百亿美元的公司，发展到每年花费数千亿美元，再到预测到本十年末每年花费数万亿美元。这至少是它们已经开始与合作伙伴签订的部分合同。这需要对其经济模式进行重大重塑。到目前为止，它们大多是亏损的公司。Anthropic 在第二季度开始盈利。据信在第三季度的某个时候，OpenAI 甚至可能开始盈利，随着 Codex 和 5.6 等的更大规模崛起。但如果我们回到一年前，它们所有的钱都是风险投资资助的亏损。即使回到今年年初，也是风险投资资助的亏损。现在它们已经扭亏为盈，实际上开始盈利了。这并不意味着它们不再接受新资本。新资本仍在涌入，以进一步加速增长。但最终，它们越来越多的业务是由自己的收入资助的，而不是资本注入。在过去一年半里，它们的利润率确实飙升了。算力的基础成本通常约为每兆瓦 1000 万、1300 万或 1500 万美元。现在最有趣的一点是：以前，如果它们提供模型——比如在 Nvidia Hopper GPU 上提供 GPT-4——OpenAI 的毛利率是负的。但现在，当 OpenAI 提供 GPT-5.6 或 Anthropic 提供 Opus 5 或 Fable 5 时，它们的收入生成已经远远超过了每兆瓦 1000 万到 1500 万美元的增量成本。以 Anthropic 为例，收入已高达每兆瓦 5000 万美元。这使它们能够做到的是：“嘿，如果我在推理能力上花 10 美元，我实际上能产生 50 美元的收入，然后我可以转身把所有这些利润增量地花在训练上。”

When we go back to last year, even at the end of the year, most of GDP growth in America was just AI infrastructure. As we look towards this year, about a third of the compute coming online is for the labs, for OpenAI and Anthropic. It may be built by others and then rented to them, but at the end customer, it’s them. As we go forward into the future, the numbers for compute are ballooning. We’re at a little bit over a trillion dollars of CapEx this year. As we go out into ’28, it’s going to be more than $2 trillion. The labs are also taking an increasing percentage of this. So ultimately, you’ve got a very interesting situation where the labs are going from companies that spend tens of billions of dollars a year to hundreds of billions of dollars a year, to forecasting to spend trillions of dollars a year even towards the end of the decade. This is at least some of the contracts they’ve begun signing with their partners. This requires a big reshaping of what happens with their economics. Up until now, they have been companies that mostly lost money. Anthropic started turning a profit in Q2. It’s believed at some point in Q3, OpenAI could start turning a profit even, with the bigger rise of Codex and 5.6 and all this. But if we go back a year ago, all the money they had was venture-funded losses. If we go back to even the beginning of this year, it was venture-funded losses. They’ve now turned the corner and are actually starting to profit. That doesn’t mean they’re not taking in new capital. The new capital is still coming in to accelerate the growth further. But ultimately, more and more of their business is being funded off of their own revenue rather than capital injections into them. Over the last year and a half, their margins have really skyrocketed. The base cost of compute tends to be around $10 or $13 or $15 million per megawatt. The most interesting aspect about what’s happening now is this: Before, if they served a model — GPT-4 being served on Nvidia Hopper GPUs — it was generating negative gross margin for OpenAI. But now, when OpenAI serves GPT-5.6 or Anthropic serves Opus 5 or Fable 5, their revenue generation has passed well beyond the incremental $10-15 million per megawatt. In the case of Anthropic, the revenue has gone as high as $50 million per megawatt. What that now enables them to do is: "Hey, if I spend 10 bucks on inference capacity, I actually generate 50 bucks of revenue, and then I can turn around and incrementally spend all of that profit on training."

我非常想了解的一件事是，你如何看待实验室的算力集中化，或者流向世界与流向实验室的算力相对比例。如果你说现在三分之一的边际算力流向了实验室，那么到什么时候世界上一半以上的增量算力会流向实验室？到什么时候实验室基本上拥有世界上绝大多数的算力？

One thing I’m very interested in understanding is how you see the centralization of compute happening at the labs, or the relative ratio of compute that goes to the world versus the labs. If you say right now a third of marginal compute is going to the labs, by when is over half of the incremental compute in the world going to the labs? By what point do the labs have basically a vast majority of the world’s compute?

今年年初，OpenAI 从 2 吉瓦起步，Anthropic 不到 2 吉瓦。到今年年底，它们都超过了 5 吉瓦。所以它们的算力整体增长了 3 到 4 倍。当你看到新增的算力时，这大约占今年新增算力的 30%。展望明年，鉴于已经签署和敲定的合同，情况会更加戏剧化。Anthropic 和 OpenAI 明年将占据多达 40% 到 50% 的算力。这种集中化似乎没有放缓或停止的迹象。事实上，它看起来只是在加速。为它们建造算力的公司将会改变。明年，一个大的新进入者，例如 SpaceX，正在建造大量算力。它们很可能会积极地将其中相当一部分租给 Anthropic 和 OpenAI，因为它们是最有能力支付最高价格的。此外，OpenAI 和 Anthropic 也开始建造自己的算力——OpenAI 用自己的芯片，Anthropic 用从 Google 购买并通过 Fluidstack 部署的 TPU。所以你问：“嘿，什么时候世界上新增算力的一半会只流向 OpenAI 和 Anthropic？”实际上，到明年年底，一半的增量算力就已经流向 Anthropic 和 OpenAI 了。因为算力增长如此之快，增量算力基本上将占算力的大部分。

At the beginning of this year, OpenAI started at 2 gigawatts and Anthropic at less than 2. End of this year, they’re both above 5. So they’ve 3-4x’d compute as a whole. When you look at the incremental compute added, that’s about 30% of the compute added this year. As we step forward to next year, given what’s already been signed and penned and inked, you’ve got something even more dramatic. Anthropic and OpenAI are taking as much as 40% to 50% of compute next year. This centralization doesn’t look like it’s slowing down or stopping. In fact, it looks like it’s only accelerating. Who’s building that compute for them will change. Next year, a big new entrant is, for example, SpaceX, which is building a ton of compute. They’re actively going to lease quite a bit of it to Anthropic and OpenAI, most likely, because they’re the ones who have the marginal capability to pay the highest price. In addition, OpenAI and Anthropic are also starting to build their own compute — OpenAI with their own chips, Anthropic with TPUs that they’re purchasing from Google and deploying with Fluidstack. So you ask, "Hey, when does half of the world’s incremental new compute go to just OpenAI and Anthropic?" It’s really by the end of next year when half of the incremental compute is already going to Anthropic and OpenAI. Because compute is growing so fast, incremental compute is going to be basically most of compute.

所以很快——你是说大概一年半到两年内——世界上大部分算力将由两家实验室拥有，或者至少是为两家实验室的需求服务。有一个趋势是，世界算力（以吉瓦计）每年翻一番，但前沿实验室的算力每年翻三倍。如果保持当前趋势，从今年年初的 2 吉瓦到今年年底接近 6 吉瓦。只需乘以 3。到 2027 年底是 18 吉瓦，到 2028 年底是 54 吉瓦。你会不会觉得：“好吧，到那时，鉴于世界算力的总量，它们根本不可能继续翻三倍”？你如何看待未来几年的世界算力情况？

So it’s very soon — you’re saying maybe within a year and a half or two years — that most of the world’s compute is owned by two labs, or at least is serving the demand from two labs. There’s this trend where maybe world compute in gigawatts doubles every year, but the compute at the frontier labs triples every single year. If you keep the current trend going, it goes from 2 at the beginning of this year to close to 6 at the end of this year. Just multiplying out by 3. It’s 18 by the end of 2027, 54 by the end of 2028. Are you like, "Okay, at that point, they simply can’t continue tripling given the amount of world compute"? How do you see the world compute situation over the next few years?

如果今年新增算力 30 吉瓦，明年 50 吉瓦，后年大约 70 吉瓦，你最终会看到这个非常有趣的现象。

If the incremental compute this year adds 30 gigawatts, next year 50 gigawatts, and the year after that roughly 70, you end up with this really interesting phenomenon.

今年新增的每瓦算力，效率远高于两年前部署的每瓦算力。全球算力中有很大一部分是今年部署的。虽然部署的总瓦数没有翻倍，但我部署的是 GB300、TPUv7 和 Trainium3，这些芯片的效率要高得多得多。它们的每瓦性能是上一代芯片的 3 到 5 倍。所以最终你有一个巨大的阶梯。如果明年 Anthropic 和 OpenAI 拿下 45% 的算力，那么到 2027 年 12 月，它们就已经拿下了全球新增算力的一半。但那一半全球新增算力，其性能实际上比之前所有算力都高。所以这又是一个乘数效应。到 2028 年底——如果这个趋势继续下去，而我看不到任何阻止它的因素——它们就会独自控制全球大部分可用算力。

A new watt deployed this year is significantly more efficient than the watts deployed two years ago. A humongous percentage of the world's compute was deployed this year. Even though it didn't double the number of watts deployed, I'm deploying GB300s, TPUv7s, and Trainium3s, which are way, way, way more efficient. They're 3-5x more performance per watt than the prior-generation chips. So ultimately you've got a huge ladder here. If Anthropic and OpenAI take on 45% of compute next year, you've got them in, let's say, December '27 having taken on half of the world's incremental new compute. But that half of the world's new incremental compute is actually at a higher performance than everything else before it. So you've got another multiplier on that. By the time you're towards the end of 2028 — if this trend continues, and I see nothing that's stopping it — you've got them just controlling most of the usable flops in the world on their own.

我困惑的是，如果我们进入一个算力价值如此之高的世界，你为什么认为 2028 年我们只增加 80 吉瓦？

The thing I'm confused about is why you think we only add 80 gigawatts in 2028 if we enter a world in which the value of compute increases so much.

顺便说一句，那是上限。那就像是在说“我他妈的极度看多”。

That's the upper bound, by the way. That's the like, "I'm so fucking bullish."

好，我们来做个思维链。几个月前我采访你时，你说要制造一吉瓦的 Vera Rubins，需要 5.5 万片 N3 晶圆、6000 片 N5 晶圆和 17 万片 DRAM 晶圆。我知道这些数字可能已经变了。

Okay, let's do some chain of thought here. When I interviewed you a few months ago, you said that in order to make a gigawatt of, I think, Vera Rubins, you need 55,000 N3 wafers, 6K N5 wafers, and 170K DRAM wafers. I know if those numbers might have changed.

我要逗你一下，但你说“晶圆”的方式太印度了。“晶圆”。

I'm going to troll you, but the way you said wafers was so fucking Indian. Vafers.

顺便说一句，我们刚搬到美国时，我的 v/w 发音问题很严重，而且我还是个素食者。我记得你告诉过我这件事。在北达科他州，我上小学时，我会说——能给我来个“wedgie”吗？能给我几个“wedgies”吗？

By the way, when we first moved to the US, I had the v/w thing pretty bad, and I was a vegetarian. I remember you told me about this. In North Dakota, I was in elementary school, and I'd be like— Can I get a "wedgie"? Can I get some "wedgies"?

总之，这是一吉瓦的情况。我让一个 LLM 运行了你的晶圆厂设备模型，算出了每年生产一吉瓦算力所需的工具成本。它说是 30 到 40 亿美元。现在假设你加上洁净室、厂房外壳以及晶圆厂的其他一切。那么 60 亿美元的晶圆厂资本支出每年能产出一吉瓦。一吉瓦现在能产生 1000 亿美元的收入。而且这 60 亿美元的资本支出每年都在产出一吉瓦，而这一吉瓦每年都在产生 1000 亿美元。所以在五年时间里，第一吉瓦已经产生了五年的利润，晶圆厂产出的第二吉瓦已经产生了四年的利润，依此类推。晶圆厂层面的 60 亿美元资本支出将产生超过一万亿美元的最终 AI 收入。

Anyways, so that's for one gigawatt. I had an LLM run your wafer fab equipment model and figure out how much the tooling costs to produce a gigawatt of compute basically every single year. It said $3-4 billion. Now suppose you add in cleanrooms and shell and everything else at the fab. So $6 billion of fab CapEx produces a gigawatt every single year. A gigawatt produces right now $100 billion of revenue. But also that $6 billion in CapEx is producing a gigawatt every single year, and that gigawatt is producing $100 billion every single year. So over the course of five years, the first gigawatt has generated five years of profits, the second gigawatt the fab has produced has generated four years of profits, and so on. $6 billion of CapEx at the fab level will have generated over a trillion dollars of end AI revenue.

是的。沿途有很多运营支出。还有很多其他资本支出，比如数据中心、电力。而且你还要为 OpenAI 的研发付费。安装。这里有很多不同的人需要钱。

Yeah. There's a lot of OpEx along the way. There's a lot of other CapEx, like the data center, the power. And you had to pay OpenAI for the R&D. Installation. There's a lot of different people who need money here.

为所有这些中间商拿走一半。这仍然意味着晶圆厂资本支出与最终产生的收入之间存在 100 倍的差距。实际上还不止，但我们非常保守。因此……这就是资本主义。你有这么大的差距，可以把 1 美元变成 100 美元。

Take away half of it for all these middlemen. That still means there's a 100x discrepancy between fab CapEx and end revenue generated. More than that, actually, but we're just being very conservative. As a result… This is capitalism. You have this huge discrepancy where you can turn $1 into $100.

他们不会想办法制造更多的镜子吗？

They're not going to figure out a way to make more mirrors?

他们会。只是这些镜子需要一些时间才能制造出来。但紧迫性如此之大，Anthropic 和 OpenAI 会说：“我们现在就能赚一万亿美元，但我们只是被 ASML 机器里的镜子卡住了。”如果我们在这上面花 1000 亿美元，怎么能制造更多的镜子？我们很快就会面临这种情况。

They are. It's just that these mirrors take some time to make. But the emergency is so big where Anthropic and OpenAI are like, "We could make a trillion dollars right now, but we're just bottlenecked on the mirrors that go into the ASML machines." How can we make more mirrors if we spend $100 billion on this? That's the situation we're going to be in pretty soon.

我们无法解决那个供应限制吗？这似乎很难想象。

We're not going to be able to solve that supply constraint? That just seems quite hard to imagine.

你见过人们在这里做有趣的套利，他们买下涡轮机然后试图转售，因为涡轮机的价值要高得多，因为它是你数据中心的瓶颈。我认为如果有人有 4 亿美元，并且有能力说服 ASML 卖给他们一台 EUV 工具，他们就应该直接去买一台，等着，然后以超过 10 亿美元的价格卖掉。但最终，是的，资本主义会导致这些事物扩张。但这是一个鞭子。鞭子信号传到末端需要很长时间。供应链不会立即反应。事实上，你去和 Carl Zeiss 的人谈谈，他们会说：“是的，是的，是的，我们需要在十年末之前制造 100 台 EUV 工具。”今年早些时候我们做那期节目时，他们甚至认为不需要制造那么多，足够制造每年 100 台 EUV 工具的镜子。现在他们说：“好吧，我们需要做到。”但实际上，由于正在发生的所有经济因素，应该更多。这需要很长时间才能实现。

You've seen people do funny arbitrages here where they buy turbines and then try and resell them, because the value of a turbine is way more since it's the thing bottlenecking your data center. I think if anyone had $400 million and the ability to convince ASML to sell them an EUV tool, they should totally just go buy one, wait, and then sell it for north of a billion dollars. But ultimately, yes, capitalism will cause these things to expand. But it's a whip. It takes a long time for the whip signal to get to the tail end of that. The supply chain doesn't react immediately. In fact, you go talk to someone at Carl Zeiss, they're like, "Yeah, yeah, yeah, we need to make 100 EUV tools by the end of the decade." When we had our episode earlier this year, they didn't even think they needed to make that many, enough mirrors to make 100 EUV tools a year. Now they're like, "Okay, we need to do that." But in reality, because of all the economics of what's going on, it should be even more. It takes so long to pill.

假设堆栈中的每一家公司都被私有化。某个超级 AGI 信徒进来，说：“我们要最大化生产。”你认为制造更多东西的物理限制会是什么？我问的原因是，我们很快就会进入一个世界，实验室收入，或者仅仅是 AI 现金流——因为显然加速器也有这些巨大的现金流——将如此之大，以至于你可以直接用现金流来资助所有这些生产的极端扩张。

Suppose that every single company in the stack got private equitied. Somebody came in who was super AGI-pilled and was like, "We're going to maximize production." What do you think the physical constraints on making more things would be? The reason I ask is we're pretty soon going to be in a world where the lab revenue, or just AI cash flows — because obviously the accelerators also have these huge cash flows — will be so big that you can just fund extreme expansion of all this production from cash flows themselves.

我大体上同意。显然有一些物理限制。按照目前供应链的扩张方式，100 大致仍然是正确的数字。

I do agree generally. There's obviously some physical constraints. The way the supply chain is expanding currently, 100 is roughly still the right number.

到 2030 年？2030 年 100 台 ASML 工具。

For 2030? 100 ASML tools for 2030.

但如果你说：“Carl Zeiss，这是 100 亿美元。请他妈的扩大生产吧。”那就会改变局面。你必须对供应链中的每家公司都这样做。

But if you said, "Carl Zeiss, here's $10 billion. Please fucking just expand production," that would change things. You would have to do this with every company in the supply chain.

但你不认为那会在明年发生？

But you don't think that's gonna happen next year?

我不认为那会在今年发生。

I don't think it'll happen this year.

我认为这不会在明年发生，也不会在后年发生，因为世界受到资本约束。但在一个世界里，比如说，顶级实验室明年合计产生一万亿美元的收入，他们也无法从中拿出 100 亿美元——我不认为他们会这么做，但是……或者至少数千亿？看起来他们确实意识到了世界的走向。

I don’t think it’ll happen next year. I don’t think it’ll happen the year after, because the world is capital constrained. But in a world where, say, the top labs are generating, even combined, a trillion dollars in revenue next year, they’re not able to take $10B of that—I don’t think they’re going to do that, but… Or hundreds of billions at least? It just seems like they realize where the world is headed.

我觉得他们完全可以……问题是，实验室明年将产生数千亿美元的收入。但最终，明年的资本支出大约是 2 万亿美元。所以存在巨大的不匹配。晶圆制造设备供应链的规模大约为 2000 亿美元。数据中心市场供应链会更多。加速器供应链会更多。能源供应链也会有一个数字。把这些加起来，资本支出将远超 2 万亿美元。所以实验室还没有达到现金流能够支撑这些的程度。

I feel like they could just make… The thing is, the labs are going to generate hundreds of billions of revenue next year. But ultimately, CapEx next year is like $2 trillion. So you’ve got this big mismatch. The wafer fabrication equipment supply chain will do something on the order of $200 billion. The data center market supply chain will do even more. The accelerator supply chain will do even more. The energy supply chain will do a number. You sum all this up, it’s going to be well north of $2 trillion of CapEx. So the labs have not yet gotten to the point where their cash flows can fund this stuff.

显然他们永远不会达到那个点，因为你希望资本支出高于回报。

Obviously they will never get to that point, because you want to keep your CapEx higher than your returns.

是的，你会再投资。

Yeah, you reinvest.

我真正想理解的关键问题是：如果当前趋势持续，到 2028 年底每个实验室将超过 50 吉瓦。所以他们合计将有 100 吉瓦。正如你所说，这些吉瓦到 2028 年将带来比现在多很多倍的吞吐量或性能，因为硬件变得更好了。不仅每瓦特浮点运算次数增加了，而且硬件在处理 AI 工作负载方面也变得更好了。好的，那么到 2028 年底实验室有 100 吉瓦。世界算力是多少？

The key question I really want to understand is: if the current trend continues, it’d be north of 50 gigawatts per lab by the end of 2028. So between them they’d have 100 gigawatts. Those gigawatts, as you’re saying, drive many-fold more throughput or performance by 2028 than they do now, because the hardware’s gotten better. Not only have flops per watt increased, but also the hardware gets better at working with AI workloads. Okay, so 100 gigawatts for the labs by the end of 2028. How much is world compute?

我认为这可能有点困难，因为到 2028 年他们已经占据了增量算力的 70-80%。而且我不确定那时市场会发生什么。

I think that may be a little difficult, given that by 2028 they’ve taken 70-80% of incremental compute. And I’m not sure what happens to markets then.

算力价格要飙升多少，他们才能真正买到 70-80% 的算力？谷歌、Meta 或亚马逊愿意卖出那么多吗？

How much does the price of compute skyrocket for them to actually be able to buy 70-80% of compute? Is Google or Meta or Amazon willing to sell even that much?

另外，当我们谈论这些吉瓦数字时，有一个注意事项。当亚马逊提供 Bedrock Anthropic 模型时，在我们的世界观里，这算作 Anthropic 的算力，因为最终它实际上被算作 Anthropic 的收入，尽管有收入分成和回扣等等。但最终在 2028 年，如果他们合计达到 100 吉瓦，他们确实对市场做了非常颠覆性的事情。

Also, there’s one caveat when we’re talking about these gigawatt numbers. When Amazon is serving Bedrock Anthropic models, that counts as Anthropic compute in our worldview, because it is effectively, at the end of the day, counted as revenue for Anthropic even though there’s a revenue share and credit back all that. But ultimately in 2028, if they get to 100 gigawatts combined, they have done really disruptive things to the market.

因为今天任何人都可以从每兆瓦 1000 万到 1500 万美元的算力中赚钱。我没开玩笑，这并不难。去弄一个 GB300 机架，去下载 Kimi 权重，去下载 vLLM 或 SGLang，设置好。Codex 和 Fable 实际上可以帮助你做到这一点。这很简单。这不是小事，但也不是火箭科学。把它放到 OpenRouter 上。非常简单。你会开始产生比你支付的算力费用更多的收入。这已经导致算力定价，每兆瓦 1000 万到 1500 万美元，开始向上弯曲。

Because anyone can make money off of $10-15 million per megawatt compute today. I kid you not, it’s not that hard. Go get a GB300 rack, go download the Kimi weights, go download vLLM or SGLang, set it up. Codex and Fable can actually help you do this. It’s pretty simple. It’s not trivial, but it’s not rocket science. Go put it on OpenRouter. It’s very simple. You’ll start generating more revenue than you’re paying for the compute. This has already led to this compute pricing, $10-15 million per megawatt, starting to inflect up.

要在 2028 年达到 100 吉瓦，你必须相信实验室能够支付更高的算力价格，因为任何人都能在 1000 万到 1500 万美元的价格下赚钱。算力现在会涨到每兆瓦 2500 万美元吗？会涨到每兆瓦 4000 万美元吗？

To get to that 100 gigawatts in 2028, you have to believe that the labs can outpay for compute, because anyone can make money at $10 to $15. Does compute now get to $25 million a megawatt? Does it get to $40 million a megawatt?

正如你所说，实验室每兆瓦产生的收入已经远超其他所有人。如果他们保持目前这样的领先优势，你会预期这种情况会持续下去。如果存在某种递归自我改进，AI 实验室相对提升——或者他们内部有未对外发布的模型，帮助他们改进下一个模型——你会预期这种情况会更加明显。

As you’re saying, it’s already the case that the labs are generating way more revenue per megawatt than everybody else. If they stay as far ahead as they are currently, you would expect that to continue being the case. If there’s some kind of recursive self-improvement where the AI labs are relatively uplifted — or they have models internally they’re not releasing externally that are helping them make their next model better — you’d expect that to be even more the case.

你不是已经看到这种情况了吗？SpaceX 或者其他稍微落后一点的玩家，如果他们无法像实验室那样在内部变现算力，就会把算力卖给出价最高的人。你会预期他们会继续竞标越来越大的算力市场份额。

Aren’t you already seeing this, where SpaceX, or whoever is slightly further behind, will just sell compute to the highest bidder if they can’t internally monetize it as well as the labs? You’d expect them to keep bidding for larger and larger shares of the compute market.

我认为这就是我的世界观。他们会继续吞噬更多的算力。但最终他们无法以当前价格或接近当前的价格做到这一点。他们确实必须开始支付每兆瓦 2500 万、3000 万、5000 万美元，才能在 2028 年真正吞噬全球 70% 的算力，到 2028 年达到 100 吉瓦，这是一个非常激进的目标。

I think that is my worldview. They will continue to gobble up more of the compute. But ultimately they can’t do it at current pricing or anywhere close to it. They do have to start paying $25, $30, $50 million a megawatt to really gobble up 70% of the world’s compute in 2028, to get to 100 gigawatts by 2028, which is a very aggressive goal.

另一个非常具有挑战性的方面是，我们已经看到 AI 实验室出现了巨大的放缓。他们倡导的这项监管实际上对实验室的减缓作用远大于对开源中文语言模型的减缓作用。OpenAI 没有发布 Astra。OpenAI 停止训练两周。Anthropic 没有发布其安全评估所称的 Model 2，人们普遍认为这是 Mythos 的下一个版本。他们显然没有发布他们最好的模型，在这种情况下，他们的每兆瓦收入会停滞甚至再次开始下降，因为其他模型再次具有竞争力。这不是他们落后了。只是他们没有发布最好的东西。

The other aspect of this that’s really challenging is that we’ve already seen a huge slowdown for the AI labs. This regulation that they advocate for is actually slowing down the labs a lot more than it slows down the open-source Chinese language models. OpenAI not releasing Astra. OpenAI stopping training for two weeks. Anthropic not releasing what their safety assessment says is Model 2, which is widely believed to be the next version of Mythos. They’re clearly not releasing their best models, in which case their revenue per megawatt stalls or can even start to decline again because other models are competitive again. It’s not that they’re falling behind. It’s just that they’re not releasing their best stuff.

如果存在某种监管影响，阻止他们发布最好的模型呢？那么他们的每兆瓦收入就不会增长得那么快。他们以高于其他人的价格购买增量算力的能力开始减弱，然后也许他们无法达到那 100 吉瓦。

What if there is some regulatory impact that prevents them from releasing their best models? Now their revenue per megawatt does not climb as fast. Their ability to buy that incremental compute for a higher price than everyone else starts to diminish, and then maybe they can’t get to that 100 gigawatts.

但在一个安全不重要的世界里，我确实相信这正是会发生的事情。他们可以开始产生每兆瓦 1 亿美元或更多的收入，并且他们可以支付每兆瓦 5000 万美元。

But in a world where safety doesn’t matter, I do believe that’s exactly what happens. They can start generating $100 million per megawatt or more, and they can pay $50 million a megawatt.

除了说“拜托，Dario，把我手里的算力全拿走吧”，其他人在算力上没有任何合乎逻辑的用途。但有一些我们无法描述的力量在起作用，可能会拖慢这一进程。我觉得一个好的直觉泵是：如果 AI 模型真的能像全自动软件工程师一样厉害呢？它们目前还没到那一步。我认为它们离完全自动化白领工作的目标还差得远。但白领工人年薪六位数或更高。如果你有一个吉瓦的算力，能支撑大约一百万白领工人的规模。那么在此基础上……那就是 1000 亿美元。这其实低得惊人。

No one else has any logical reason to do anything with their compute besides say, 'Please, Dario, take everything off of my hands.' But there are forces at play, which we cannot describe, that would potentially slow this down. I think a good intuition pump is: what if the AI models were literally as good as a fully automated software engineer? They're not currently there yet. I think they're far from being able to fully automate the job of a full white-collar worker. But white-collar workers earn six figures or north of that a year. If you have a gigawatt that can sustain a population of, say, roughly a million white-collar workers. Then off the back of that… That would be $100 billion. That's actually surprisingly low.

是啊，每人 10 万美元，一百万人。我也不确定。

Yeah, $100K per person, million population. I don't know.

但如果实现完全 AGI，每吉瓦的产值会是数千亿美元。另一方面——我们一直看到这一点——大部分价值并没有被捕获。这些模型产生的大部分价值并没有给到 OpenAI 和 Anthropic。谢天谢地，到目前为止，大部分价值都给了用户。Jane Street 与 OpenAI 签订了 GPT-5.6 超速模式的独家合同，或者 Jane Street 作为 Anthropic 最大的客户之一，他们从支付的 token 中产生的价值远远超过 Anthropic 的利润，因为他们能从市场中赚钱。再比如 Meta，一度传闻占 Anthropic 业务的 10%。他们通过优化广告算法或诸如此类，让用户参与时间延长 5%，这些带来的效率提升远超 Anthropic。他们从使用这些模型中赚到的钱比 Anthropic 多得多。这才是关键。

But it would be many hundreds of billions of dollars per gigawatt if you get full AGI. The other aspect of this — and we've continued to see this — is that most of the value capture is not happening. Most of the value that these models generate does not get given to OpenAI and Anthropic. Thankfully, so far it is mostly just being given to the users. Jane Street, with their exclusive contract with OpenAI for GPT-5.6 Ultrafast mode, or Jane Street where they're one of Anthropic's biggest customers, is generating way, way, way more value out of the tokens they're paying for than Anthropic is generating in terms of profit, because they get to make money off of the market. Or take Meta, who at one point was rumored to be as much as 10% of Anthropic's business. They're generating way more efficiencies by optimizing their ad algorithms or what have you, getting engagement time 5% longer, all these things. They're making way more money off of using these models than Anthropic is. That's what's required.

当然，如果你有一百万个新软件工程师，软件工程师的成本也会下降。我困惑的一点是，市场会达到均衡吗？如果达到均衡，你会不会预期算力的价格等于 Anthropic 和 OpenAI 能从中产生的价值，或者非常接近，只给 Anthropic 和 OpenAI 留一点加价？现在很奇怪的是，算力的售价和 Anthropic 能从中赚到的钱之间有 4 倍甚至更多的差距。

Sure, if you had a million new software engineers, the cost for a software engineer would also fall. One thing I'm confused about is, does the market come into equilibrium? If it comes into equilibrium, would you just expect the price of compute to equal whatever Anthropic and OpenAI can generate from it, or be very close to it with a small amount of markup for Anthropic and OpenAI? Right now it's really weird that there is a 4x or more difference between what compute sells for and how much money Anthropic can make from it.

在一个每吉瓦收入持续增长的世界里，如果 Anthropic 将吉瓦变现的能力翻倍或三倍，差距继续扩大就很奇怪了。Anthropic 仅仅凭借一些权重，就能把成本 10 美元的东西变成 100 美元。

In a world where the revenue per gigawatt continues to increase, if Anthropic's ability to monetize a gigawatt doubles or triples, it'd be weird if the gap continued to increase. Anthropic, just by having some weights, can take something that cost them $10 and turn it into $100.

这总是一个有趣的问题。AI 的价值去哪了？AI 产生了所有这些价值。有最终用户，我想我们都同意他们产生的价值比任何人都多，所以他们为这些模型支付了很多。然后是应用层。到目前为止，应用层产生的价值很少。然后是模型层，直到一年前还在产生负毛利，现在产生了巨大的正毛利。看起来它正朝着每兆瓦产生 1 亿美元的方向发展。所以就像你说的，把 10-15 美元变成 100 美元。

This is always a fun question. Where does the value go in AI? AI's generating all this value. You've got the end user, which I think we all agree is generating more value than anyone else, hence they're paying a lot for these models. Then you have the app layer. So far the app layer's generated very little value. Then you've got the model layer, which up until a year ago was generating negative gross margins and is now generating massive positive gross margins. It looks like it's on the path to generating $100 million per megawatt. So turning $10-15 into $100, as you said.

但如果我们回到一年前，硬件供应链在产生所有这些毛利，而其他所有人都在亏钱。OpenAI 和 Anthropic 只是在烧 VC 的钱，许多其他初创公司也是如此。许多超大规模厂商在不知道是否有回报的情况下建设基础设施。所以最终，模型层在创造负价值，如果你愿意这么说的话，因为他们卖 token 的价格低于基础设施成本。所有价值都被芯片、晶圆厂捕获了。最初在 2023 年，内存厂商从 HBM 或 AI 内存中没赚到钱，尽管理论上他们提供的价值巨大。现在你有了……嗯，实际上台积电捕获的价值远低于内存厂商。所以价值捕获发生了很大变化，这对跟踪市场或参与市场的人来说很有趣，比如 Jane Street 就是一个例子。

But if we go back a year ago, the hardware supply chain was generating all this gross margin while literally everyone else was losing money on it. OpenAI and Anthropic were just plowing VC money in, as were many other startups. Many of these hyperscalers were building infrastructure without knowing if there was going to be a payoff. So ultimately you had this negative value being created on the model layer, if you will, because they were selling the tokens for less than it cost them on the infra side. All the value was being captured at the chip, the fab. Initially in 2023, the memory guys were making no money off of HBM or memory for AI, even though theoretically the value they were delivering was humongous. Now you've got… Well, actually TSMC captures way less value than the memory guys. So the value capture's shifted around a lot, which is very fun for people tracking the market or participating in the market, like Jane Street as an example.

这不是广告。这不是广告。这不是广告。他们是赞助商，但你不用这么卖力地推销他们。

This is not an ad. This is not an ad. This is not an ad. They're a sponsor but you don't have to plug them that hard.

那么接下来会发生什么？Anthropic 和 OpenAI 的价值捕获开始慢慢膨胀。他们会膨胀并拿走所有价值吗？

So what happens going forward? Anthropic and OpenAI have slowly started to balloon in value capture. Do they balloon and take all the value capture?

嗯，那曾是一个想法，然后 Elon 展示了，“实际上，不。我可以把我的算力以每兆瓦 2500 万美元或 4000 万美元的价格卖给 Anthropic 和 Google。即使是短期的事情，我也以这个价格卖掉了，我将在一年内收回全部资本支出。”

Well, that was a thought, and then Elon showed, 'Actually, no. I can sell my compute for $25 million a megawatt or $40 million a megawatt to Anthropic and Google. Even if it's a short-term thing, I've sold it for this price, and I'll recoup my entire CapEx in a year.'

你预测相关的算力部分——B300 或 SpaceX 以每吉瓦 400 亿美元卖给 Google 的那些——明年年底会卖多少钱？

What's your prediction of how much the relevant tranche of compute — B300s or whatever that SpaceX sold for $40B a gigawatt to Google — what does that sell for at the end of next year?

我认为大多数算力仍将以低于每吉瓦 200 亿美元的价格交易。

I think most compute will still continue to transact at sub-$20 billion a gigawatt.

即使到明年年底？因为所有这些都必须融资。如果 Meta、微软、亚马逊、SpaceX 可以在没有找到客户的情况下建造算力，只是说“去他的，我要建这个算力”，然后转身等到建好，他们现在就控制了局面。大多数算力在建造之前就已经签约了。这就是 Elon 在市场上利用的一点。他实际上拥有所有这些算力。他说，“嘿，Anthropic，我知道你们每吉瓦赚 600 多亿美元。你们为什么不以疯狂的价格买我的东西呢？”显然，这不是 Elon 决定的，也不是 Anthropic 决定的。市场自己解决了。其他人，你去随便一家云厂商，他们说，“好吧，我要建一吉瓦或 100 兆瓦的算力。我要花资本支出。我需要转身找到客户。如果我想找到客户，我需要找到资本。谁给我资本和客户？客户必须签合同。”

Even at the end of next year? Because all of it has to be financed. If Meta, Microsoft, Amazon, SpaceX can build compute without finding a customer, just saying, 'Fuck it, I'm going to build this compute,' and then turn around and wait till it's already built, they now control what's going on. Most compute is contracted well before it's built. This is what Elon took advantage of in the market. He actually had all this compute. He was like, 'Hey, Anthropic, I know you're making $60-plus billion per gigawatt. Why don't you just buy my stuff for a crazy amount of money?' Obviously it's not like Elon decided this or Anthropic decided this. The market figured itself out. Other people, you go to a random cloud, they're like, 'Okay, I'm going to build a gigawatt of compute or 100 megawatts of compute. I'm going to spend the CapEx. I need to turn around and find a customer. If I want to find a customer, I need to find the capital. Who's going to give me the capital and the customer? The customer has to sign a deal.'

然后我把客户对信贷市场的承诺拿去融资。所以这里有一种完全不同的权力结构，Meta 实际上在囤积算力。他们和 SpaceX 是唯一可能排第三的玩家，因为他们囤积了所有这些算力。他们利用自己的资产负债表和能力来建设算力，而没有一个大规模变现的终端客户。他们有真正的资产负债表，所以可以去信贷市场。你建一个吉瓦，就能赚到利润，不是疯狂的利润，但也是不错的利润。现在我有了所有这些算力。现在 Meta 和 SpaceX 有了这种选择权，环顾四周，心想：‘我的内部用例能让我赚更多钱，还是我应该出去以疯狂的高价卖给 Anthropic 或 OpenAI？’所以现在我们进入了一个 SpaceX 和 Meta 都在说‘实际上，我要建算力，我可以以不是 13 美元的价格出租。我可以卖 25 美元、50 美元甚至更高’的体制。

Then I take the customer’s commitment to the credit markets and I raise the capital. So there’s this completely different power structure where Meta is effectively hoarding compute. Them and SpaceX are the only plausible #3, because they’re hoarding all this compute. They’re using their balance sheets and capabilities to build compute without an end customer that’s monetizing at a huge degree. They have an actual balance sheet, so they can go to the credit market. You build a gigawatt, you can make your margin, not a crazy margin, but a good margin. Now I have all this compute. Now Meta and SpaceX have this optionality of looking around and being like, 'Is my internal use case going to make me more money, or should I go out there and sell it to Anthropic or OpenAI at crazy margins?' So now we’ve entered a regime where SpaceX and Meta are saying, 'Actually, I’m going to build the compute, and I can rent it out for not $13. I can sell it for $25, $50, and more.'

所以在我做视频论文和其他格式的内容时，我一直在为播客寻找新的编辑。但主动寻找编辑非常耗时，因为绝大多数候选人不符合我要求的画像。所以我在 Grok Bot 里创建了一个招聘机器人，看看是否有帮助。我给了它一个巨大的上下文转储，基本上把我想要的一切都独白了一遍，然后它启动了另外四个机器人来分别聚焦搜索的不同部分。一个浏览了我过去一年的邮件，寻找相关的主动联系。一个搜索了我的 X 信息流和私信。一个浏览了我喜欢的各种纪录片的片尾字幕。最后一个寻找为我关注的 YouTuber 工作的编辑。然后 Grok Bot 把子代理找到的所有这些不同候选人，根据我的标准进行筛选，并给出了一个最终候选名单供我审阅。说实话，第一批里有几个我想看的好候选人，但主要是些废柴。但在我给 Grok Bot 更多关于它遗漏之处的反馈后，它带着一份我实际上非常兴奋的新候选人名单回来了。我最终把整个工作流保存为一个例行程序。所以现在每周 Grok Bot 都会检查我的入站邮件和 X 私信，寻找有潜力的新候选人进行可能的面试。如果你想自己试试 Grok Bot，请访问 x.ai/bot。

So as I've been doing video essays and other formats, I've been looking to hire a new editor for the podcast. But actively searching for editors has been quite time-consuming because the vast majority of candidates don't fit the profile that I'm looking for. So I created a recruiter in Grok Bot to see if it would help. I gave it a huge context dump where I monologued basically everything that I wanted, and then it spun up four other bots to narrow in on different parts of the search. One went through the last year of my email for relevant inbound. One searched my X feed and DMs. One went through the end credits on various documentaries I like. And the last one looked for editors who work for some of the YouTubers that I follow. Grok Bot then took all these different candidates that the subagents had found, filtered them against my criteria, and delivered for a final shortlist to review. To be honest, the first batch had a few good candidates I wanted to see, but mainly a bunch of duds. But after I gave Grok Bot some more feedback about what it was missing, it came back with a new list of candidates that I'm actually extremely excited about. I ended up saving this whole workflow as a routine. So every week now Grok Bot checks my inbound email and X DMs for promising new candidates to potentially interview. If you want to try Grok Bot yourself, go to x.ai/bot.

你觉得到 2027 年底，他们的每吉瓦营收会是多少？对于 Anthropic 或 OpenAI，到 27 年底。

What do you think their revenue per gigawatt is by the end of 2027? For Anthropic or OpenAI, by the end of ’27.

我认为这在很大程度上取决于谁拥有最好的模型，以及他们是否被允许继续发布最好的模型。但我不明白为什么不会达到每兆瓦 5000 万美元以上。到 27 年底。

I think it’s highly dependent on who has the best model, if they’re allowed to keep releasing their best models. But I don’t see why it wouldn’t be $50-plus million a megawatt. By the end of ’27.

哦，到 27 年底？

Oh, by the end of '27?

那就更有挑战性了，但我认为可能会更高，达到每兆瓦 7000 万到 8000 万美元，全公司平均，甚至更高。

That’s where it gets more challenging, but I think it could get higher than that, to like $70, $80 million a megawatt, blended across the company, if not higher.

看起来低了。如果是这样，那么算力的价格会怎样？

Seems low. So if that’s the case, then what happens to the price of compute?

嗯，如果我是 Anthropic，增量算力是值得的。也许我会在 SpaceX 的算力上每兆瓦花 4000 万美元。如果我是 SpaceX，我会看向供应链，心想：‘嗯，我和 Jensen 达成了这笔交易（他现在突然开始用 Twitter 了）。’而 Elon 说他们独家使用 Nvidia，但为什么 Jensen 不提高价格？然后 SK 海力士、美光和三星看到这个，他们会想：‘嗯，我们为什么不提高价格？’所以在价值捕获方面，我认为这里有牛鞭效应。仅仅因为有人提高了价格，并不意味着整个供应链会立即重新平衡。但随着时间的推移，供应链会重新平衡，东西会越来越贵。要获得那部分增量产能，你几乎不得不这样做。所以台积电提价很慢，但内存公司提价很快。基板公司提价很快。如果价格是 15 美元，Elon 不会卖，但他卖是因为价格是 25 美元以上。所以显然他很快提高了价格。

Well, if I’m Anthropic, incremental compute is worth it. Maybe I spend $40 million a megawatt on SpaceX compute. If I’m SpaceX, I look to the supply chain and I’m like, 'Well, I’ve struck this deal with Jensen (where he’s now all of a sudden using Twitter).' And Elon’s saying they’re exclusive to Nvidia, but why doesn’t Jensen raise his prices? Then SK Hynix and Micron and Samsung look at it and they’re like, 'Well, why don’t we raise our prices?' So with the value capture, I think there’s a bullwhip effect here. Just because someone has raised prices doesn’t mean the entire supply chain rebalances immediately. But over time, the supply chain will rebalance and things will cost more and more. To get that incremental capacity, you sort of have to. So TSMC raising prices very slowly, but memory companies raising prices very quickly. Substrate companies raising prices very quickly. Elon wouldn’t have sold if it was $15, but he’s selling because it’s $25+. So obviously he raised his prices really quickly.

我很惊讶你认为每吉瓦营收到明年年底不会增加超过 100 甚至更多。RSI 什么时候发生？起飞什么时候发生？或者即使 RSI 不发生，就说当前进展速度继续。看看我们在过去一年半里取得了多少进展。一年半前的模型是什么？Claude 3.5 之类的？

I’m surprised you think that revenue per gigawatt doesn’t increase way more than even 100 per gigawatt by the end of next year. When does RSI happen? When does takeoff happen? Or even if RSI doesn’t happen, just say the current rate of progress continues. Just look at how much progress we’ve made in, let’s say, the last year and a half. What was the model from a year and a half ago? Claude 3.5 or something?

我对这个的问题是，世界上存在的最好的模型是在二月份训练的。

My problem with this is that the best model that exists in the world was trained in February.

所以你是说也许我们就是不被允许发布实验室最好的模型。OpenAI 说他们两周不训练模型，天哪。搞什么？

So you’re saying maybe we just won’t be allowed to release the labs’ best models. OpenAI says they’re not training models for two weeks, man. What the hell?

有一件事是，在内部，他们是否得到了足够的使用，以至于他们会抬高算力的价格？另一件事是，AI 的整体进展是否会因为监管而放缓？

There’s one thing where internally, are they getting enough use for it that they’ll bid up the price of compute? Another is, does AI progress as a whole slow down because of regulation?

是的，但他们甚至不被允许在内部使用这个新模型。Astra 甚至没有在内部广泛部署。但尽管如此，如果你有一个模型是……去年年初发布的模型是什么？GPT……4o？那是 4o 吗？

Yeah, but they’re not even allowed to use this new model internally. Astra’s not even widely deployed internally. But still, if you have a model that is… What was the model released at the beginning of last year? GPT… 4o? Was that 4o?

是的。

Yeah.

你说的是到这个时候，GPT-4o 到 Mythos 2 级别的飞跃，再说一次，到 2027 年底。

You’re talking about a GPT-4o to Mythos 2-size leap by this point, again, by the end of 2027.

是的，但 Mythos 2 还没出来。甚至 Mythos 也是。又是那个飞跃。甚至 Mythos 都不被允许发布。他们把它阉割了。我们不能用它来优化推理性能。我们不能用它来优化各种东西。

Yeah, but Mythos 2’s not out. Or even Mythos. That leap again. Even Mythos is not allowed to be out. They’ve neutered it. We can’t use it to optimize inference performance. We can’t use it to optimize all sorts of things.

是的，也许 AI 进展或 AI 部署会有一些放缓，这意味着每吉瓦营收可能会更低。但这是我能看到的唯一方式，到明年年底每兆瓦只有 1 亿美元。只要模型变得更好，它产生的价值就会更好。显然，谁捕获价值仍有争议，但最终每个人都会提高价格。因为他们可以，而且这是超级通胀的。尤其是如果监管方法是……现在，到目前为止，只是‘不要发布模型。’但越来越多地，监管方法是纽约禁止数据中心。德克萨斯州在暂停。

Yeah, maybe there’s some slowdown in AI progress or the deployment of AI that means the revenue per gigawatt can be lower. But that’s the only way I could see it being only $100 million per megawatt by the end of next year. As long as the model gets better, the value generated out of it gets better. Obviously, who captures the value is still up for debate, but ultimately everyone’s going to raise their prices. Because they can, and it’s super inflationary. Especially if the method of regulation is… Right now, so far, it’s just 'don’t release the models.' But more and more, the method of regulation is New York’s banning data centers. Texas is holding moratoriums.

俄亥俄州正在说，或者至少试图说，你必须支付一定半径内所有人的房产税。这类事情会减少供应并增加成本。这也会被转嫁出去。你最终会陷入一种境地，进步确实会放缓，至少在外在意义上，即使模型内部越来越好。

Ohio is saying, or at least trying to say, you have to pay everyone's property tax in a certain radius. These sorts of things are going to decrease supply and increase cost. That's going to get passed on as well. You start to end up in a spot where progress does slow, at least in the external sense, even if the models internally keep getting better and better.

在起飞场景中，为什么 Anthropic 不把最好的模型提前六个月对外发布？因为安全和监管，还有竞争优势？如果进步加速，那六个月的差距实际上会更大。所以，这将使每兆瓦收入增长的上限远低于今年上半年我们看到的增长。

In a takeoff scenario, why would Anthropic not have their best model six months ahead of what is externally available? Because of safety and regulation, but also the competitive advantage? That six-month difference, if progress accelerates, is actually a bigger differential. So that's the thing that would cap revenue-per-megawatt gains to much lower growth than we've seen in the first half of this year.

我对此非常感兴趣。随着这些公司上市并对投资者负责，假设到明年年底他们拥有接近 20 吉瓦的算力。那么 10% 的算力就是 2 吉瓦。假设他们想把训练算力的比例从 60% 提高到 70%。他们的投资者会说：“好吧，如果你每吉瓦能产生 1000 亿美元收入，你基本上是在为了增加训练算力而拒绝 2000 亿美元的收入。”所以投资者会问：“搞什么鬼？你已经在训练上花这么多钱了。为什么还要花更多？”作为一家上市公司，如果他们只是说：“不，我们会继续增加训练算力的比例，以抵消每吉瓦算力带来的收入增长”，你认为会发生什么？

Here's something I'm very interested in. As these companies go public and they're accountable to investors, let's say by the end of next year they have close to 20 gigawatts. So 10% of compute is 2 gigawatts. Let's say they want to go from 60% of compute to training to 70% of compute to training. And their investors are like, "Well, if you're going to be able to generate $100 billion per gigawatt, you're basically saying no to $200 billion of revenue in order to increase your training compute." So investors are like, "What the fuck? You're already spending so much on training. Why are you spending even more on training?" As a public company, what do you think would happen if they're just like, "No, we will keep increasing the share of compute we spend on training to offset the increase in revenue that each gigawatt of compute is giving us"?

这是我个人的看法。实验室会随着时间的推移，将越来越少的算力分配给推理。我认为这非常非共识。大多数人的标准信念是：“哦，大部分算力会用于推理。”但大部分将用于训练的前向传播，而不一定是产生收入的推理。最终，如果他们今天每兆瓦能产生 3000-4000 万美元收入，你分配 40% 给推理。如果你现在每兆瓦能产生 6000-7000 万美元，你还会分配 40% 给推理，产生所有这些利润，然后进行分红和股票回购吗？还是去构建 AGI？我认为 Anthropic 和 OpenAI 的明显答案，不仅是在高管层面，也包括董事会，是去构建 AGI，因为这更有利可图。所以最终你会看到他们不断提高训练算力的比例——

This is what I personally believe. The labs are going to allocate less and less compute to inference over time. I think that's very non-consensus. The standard belief of most people is, "Oh, most compute will go to inference." Most of it will go to forward passes for training, not necessarily revenue-generating inference. Ultimately, if they're generating $30-40 million per megawatt today, you allocate 40% to inference. If you now get to generating $60-70 million per megawatt, do you still allocate 40% to inference and generate all this profit and then do dividends and share buybacks? Or do you go build AGI? I think the obvious answer from Anthropic and OpenAI, not just at the executive level but also their board, is to go build AGI, because it's way more profitable. So ultimately you're going to see them ratchet up their percentage of compute dedicated to training—

而如果他们把每一份新增算力都用于推理，这些算力会越来越能产生利润。

While each increment of compute is getting more and more profit-generating if they had dedicated it to inference.

对。关键是，如果我在卖 token……

Right. The whole point is, if I'm selling tokens…

OpenAI 发布超快模式是只对外，还是内部也在用？

Is OpenAI releasing Ultrafast mode for just external, or are they doing it internally too?

结果发现，不是。实际上，我会把它分配给内部和外部，因为我从超快 AI 或最佳 AI 模型中获得的内部价值远高于外部用户。所以最终，当然，我每兆瓦能产生 1 亿美元，但如果我把这些算力转向 AI 研究，我能获得多少增量进展？这对我未来的盈利潜力，无论我做了什么，折现现金流有什么影响？他们不会进行这种计算，但最终，将越来越多的算力用于内部更有意义。推理算力如此庞大的唯一原因是为了发展你的训练集群。

It turns out, no. Actually, I'm going to allocate it to internal and external, because the internal value I'm generating from super-fast AI or the best AI model is way more than what someone external is. So ultimately, sure, I could generate $100 million per megawatt, but if I turn that towards AI research, what is the incremental progress that I get? What does that do towards my future earnings potential, the discounted cash flows of whatever the hell I've done? They're not going through that calculation, but ultimately it makes more sense to dedicate more and more compute internally. The only reason to have inference compute be so large is so you can grow your training fleet.

我认为这是一个有趣的经济学问题，我觉得我们可以让模型来消化。在一个他们减少推理算力比例的世界里，什么必须成立？

I think this is an interesting economics question that I feel we can have the models digest. What would have to be true about a world where they reduce the fraction of compute spent on inference?

我认为他们在过去三个月里已经在这样做了。我认为在今年的一些时候，他们增加了算力的比例……让我们逐月来看。你会同意，每个月 Anthropic 增加的算力都比上个月多。可能在他们签 SpaceX 协议或其他什么的时候会有一些噪音，但总的来说，算力是呈上升曲线的。所以在一月份，他们增加的算力比十二月份少，但他们的收入增长却飙升了。然后他们有点趋于平稳。他们现在不是每月增加 250 亿美元的年度经常性收入。这意味着他们获得的边际兆瓦，用于研发的比例高于推理。所以事实上，他们今天正在增加用于研发的算力。我认为如果你足够关注他们的所作所为，这是不言而喻的。

I think they have been over the last three months already. I think at parts of this year, they were increasing the fraction of compute… Let's just take it month by month. You would agree that every month, Anthropic has added more compute than the prior month. There might be some noise when they sign a SpaceX deal or whatever, but in general, the amount of compute is a curve up. So in January, they added less compute than December, and yet their revenue adds skyrocketed. Then they've sort of plateaued. They're not adding $25 billion of ARR every month now. That means the marginal megawatt they're getting is going as a higher percentage to R&D than it is to inference. So they are factually increasing their compute towards R&D today. I think this is self-evident if you look enough at what they're doing.

如果我看你所说的全球算力增长速度，我想理解一些事情。似乎如果我把你刚才说的数字加起来，到 2028 年底全球算力将超过 200 吉瓦，对吗？

If I look at the numbers you said for how fast world compute grows, here are some things I want to understand. It seems like if I add up the numbers you just said, it would be over 200 gigawatts of world compute by the end of 2028, right?

是的，全球范围内。

Yeah, globally.

好的。2028 年之后，全球 AI 算力还能以多快的速度继续增长？

Okay. How fast can that continue growing, global AI compute after 2028?

今年 30，明年 50，28 年 70。29 年应该在 90-100 左右。然后每年再增加 100 或什么的？我认为斜率可以继续向上。很难预测四年以后的事情。谁知道我们是否处于 RSI 状态，或者世界经济何时以每年 10% 的速度增长？因为如果你每年增加 100 多吉瓦，你就处于荒谬的 GDP 增长中。

30 this year, 50 next year, 70 in '28. '29 should be on the order of 90-100. Then just 100 more every single year or something? I think the slope can continue to go upwards. It's hard to predict anything more than four years out. Who knows whether we're in an RSI regime, or when is the world economy growing at 10% a year? Because if you're at 100+ gigawatts a year, you're at absurd GDP growth.

如果你认为 2028 年全球有 200 吉瓦，那么到那时中国有多少？在整个趋势中，中国的算力如何继续增长？因为如果 RSI 的事情在西方率先启动，而中国还没有大量算力，也许我们生活的世界会与没有这种情况时不同。

If you think there's 200 gigawatts globally in 2028, how much is in China by that point? How does Chinese compute continue increasing through this whole trend? Because if the RSI stuff kicks off in the West before China has a large amount of compute, maybe we're living in a different world than when it doesn't.

如果我们回到 2022 年，美国大约贡献了全球算力的 45-50%。中国大约贡献了 30-35%。其余由世界其他地区承担。自 2022 年以来，我们对华实施了严格管制，美国大幅增加。所以今天，70% 的电力部署在美国。中国真的非常少。用于数据中心 AI 计算的电力中，中国占比不到 10%。随着我们向前迈进，他们仍然处于非常小的数字。

If we level-set back to 2022, the US was adding about 45-50% of the world's compute. China was adding about 30-35%. The rest was being taken up by the rest of the world. Since 2022, we've had big regulations against China and a dramatic increase in America. So today, 70% of watts are being deployed in America. China is really a very small number. Sub-10% of watts being deployed for data center AI compute is in China. As we step forward, they're still at a very small number.

他们的国内产量相当小。他们从英伟达的采购量仍然相当小，而且其中很多最终流向其他地方，比如马来西亚之类的。所以归根结底，中国国内在新增算力中的占比仍低于 10%。我认为到 2028 年可能会开始出现拐点。但很容易预测中国将拥有 30 吉瓦或更少的 AI 算力。

Their domestic production is quite small. Their purchasing from Nvidia is still quite small, and a lot of that ends up in other places as well, Malaysia or what have you. So ultimately, China domestically still continues to have sub-10% of incremental new compute. In 2028 it might start to inflect up, I think. But it’s pretty easy to say China will have 30 gigawatts of AI compute or less.

到 2028 年？

By 2028?

是的，在 2028 年。

Yeah, in 2028.

好的。那他们的指数增长会有多快？

Okay. And then how fast does their hockey stick go up?

我确实认为在 2028 年，他们能部署的算力会有大幅提升。2026 年，他们仍主要依赖大量走私芯片，很多台积电为那些他们认为不是华为但最终是华为的公司制造的芯片，或者三星出货的大量 HBM。但到 2027 年，晶圆厂开始投产。尤其是 2028 年，中芯国际和长鑫存储等晶圆厂开始投产，国内产量实际上达到每年数百万颗。到 2028 年，他们仅靠国产芯片就能新增 5-10 吉瓦。这些芯片肯定比英伟达 2028 年、谷歌 2028 年或 OpenAI 2028 年将拥有的芯片差得多。

I do think in 2028, they have a big uplift in what compute they’re able to deploy. In 2026, they’re still mostly relying on a lot of the smuggled chips, a lot of the chips that TSMC made for companies that they thought weren’t Huawei but ended up being Huawei, or a lot of HBM that Samsung is shipping. But in ’27, fabs start to go up. In ’28 especially, fabs start to go up from SMIC and CXMT and such, where domestic production is actually reaching many millions of units a year. Now they’re incrementally adding 5-10 gigawatts, in just 2028, of domestically produced chips. Those chips are definitely worse than the chips that Nvidia will have in ’28, or Google will have in ’28, or OpenAI will have in 2028.

所以你是说，即使吉瓦数也高估了情况。是 30 吉瓦，但芯片质量差得多。

So even the gigawatt number overstates things, you're saying. It’s 30 gigawatts, but it’s really much worse chips.

对。

Right.

但如果你认为第二年全球将新增 100 吉瓦——我知道你说过你无法预测那么远——那中国下一年能增加多少？基本上，我想知道：他们是在能够开始大量出货算力时就直接指数增长，还是会仍然低于美国及其盟友？

But if you think the world is going to add 100 gigawatts the following year — I know you said you can’t really say that far out — how much is China able to add the subsequent year? Basically, I want to know: do they just hockey stick at the point at which they are able to start shipping large amounts of compute, or is it still going to be less than US plus allies?

这很大程度上取决于美国是否通过《MATCH 法案》，工具是否继续受到出口管制，以及中国能以多快速度建造他们开始能够国产的新设备。但最终，中国肯定会指数增长。如果说中国真正擅长什么，那就是非常非常快地扩大制造业规模。我想中国将开始能够将越来越多的外国芯片采购引入国内，或者至少缩小美国允许英伟达向他们出售的差距，诸如此类。

There’s a lot left to whether or not the US passes the MATCH Act, whether or not tools continue to get export-controlled, how fast China can build their new equipment that they’re starting to be able to produce domestically. But ultimately, China is definitely going to hockey stick. If there’s anything China’s really good at, it’s scaling manufacturing really, really quickly. I imagine China will start to be able to extract more and more purchasing of even foreign chips into domestic China, or at least close the gap in what the US is allowing Nvidia to sell them, or what have you.

但你认为中国在 2029 年能新增 50 吉瓦吗？

But do you think China could be adding 50 incremental gigawatts in 2029?

我认为这完全合理。其中一部分也可能从国外采购。但是的，我认为中国在 2029 年达到 50 吉瓦是完全合理的。

I think that’s completely reasonable. Part of that could also be purchased from foreign. But yeah, I think it’s completely reasonable that China in 2029 can do 50 gigs.

但如果其中大部分是国产芯片，那么存在某种因素，使得这 50 吉瓦实际上相当于美国芯片的 20 吉瓦。

But if most of those are domestic chips, there is some factor there where that 50 gigawatts is really worth as much as 20 gigawatts from American chips.

对。

Right.

所以你实际上在预测一个世界，如果按质量加权吉瓦数，2028 年的领先实验室可能比中国在 2029 年甚至 2030 年拥有的算力还要多。这意味着没有采取任何措施来减缓美国实验室的发展。

So you’re actually projecting a world where maybe the leading lab in 2028 has more compute than all of China will have in ’29 or even ’30, if you weighted gigawatts by their quality. Implying that there’s nothing done to slow down the US labs.

没错。但显然政府和政客们开始这么做了。而中国不会减缓 AI 发展。事实上，他们唯一要做的就是加速。

That’s right. But clearly the government and politicians are starting to do that. Whereas China’s not going to slow down AI. In fact, the only thing they’re going to do is accelerate it.

说实话，当我采访黄仁勋并问到出口管制时——我是一个自由意志主义者——我并不真正确定我对这个问题的看法。我在为他相反的观点进行强论证，因为我认为理清想法很重要。我当时想，“是的，也许如果我们与中国合作，对我们更好，尤其是因为他们控制着供应链和机器人技术所需的其他东西的很大一部分。”但我没有意识到算力状况像你说的那么糟糕。实际上，出口管制似乎确实……如果他们出货量如你所说，那将是一个巨大的差异。当我们拥有自动化编码器并进入自动化研究员阶段时，中国在算力存量上远远落后。如果最终是这样，那就会奏效。我认为这实际上是一个显著的成功。

Honestly, when I interviewed Jensen and asked about export controls — I am a libertarian person — I wasn’t genuinely sure what I thought about this issue. I was steelmanning the opposite view from what he has, because I think it’s important to hash out ideas. I'm like, "Yeah, maybe there’s a world where if we just cooperated with China, it would be better for us, especially since they control so much of the supply chain and the other things that will be needed for robotics." But I didn’t realize the compute situation was as fucked as you’re saying. Actually, the export controls do seem to have really… If they ship the amount that you’re saying, that’s a huge difference. By the time we have automated coder and are getting into automated researcher, China is way far behind on the compute stock. If that ends up being the case, that would have worked. I think that’s actually a notable success.

唯一的警告是，其中一部分是出口管制，但另一部分也只是金融体系。美国金融体系比中国金融体系更愿意对初创公司进行 YOLO 式投资。但一旦中国金融体系选择了一个重点行业，他们会补贴得多得多。所以中国半导体行业获得的补贴比世界其他地区半导体行业的总和还要多得多。

The only caveat there is that some of it is export controls, but some of it is also just financial systems. American financial systems are more willing to YOLO into startups than Chinese financial systems. But once Chinese financial systems choose an industry to focus on, they’ll subsidize it a hell of a lot more. So the Chinese semiconductor industry has significantly more subsidies than the rest of the world’s semiconductor industries combined.

如果起飞不像你暗示的那么快，而是需要更长时间，那么最终中国将在半导体方面大幅追赶，这在某个时候就会变成算力。

If takeoff is not as fast as you’re implying but actually takes longer, then ultimately China will catch up drastically on the semiconductor side, which then is compute at some point.

另一个值得注意的方面是，相对于他们拥有的算力量，今天的中国公司在 AI 模型方面并不落后太多，至少公众感觉如此。领先的中国实验室总共最多拥有 100-200 兆瓦的算力，字节跳动 Seed 是个例外，他们拥有的远不止这些。但 Kimi 并没有运行 1 吉瓦或接近这个数字。而 Anthropic 到年底将超过 5 吉瓦。所以问题是，这有关系吗？

The other noteworthy aspect of this is that Chinese companies today are not that far behind in AI models, at least perceivably by the public, relative to the amount of compute they have. The leading Chinese labs have 100-200 megawatts total of compute at most, ByteDance Seed being the one outlier where they have significantly more than that. But Kimi is not running a gigawatt or anywhere close to it. Whereas Anthropic is more than 5 gigawatts by the end of the year. So the question is, does it matter?

我认为目前这种算力差异并不那么重要。

I think right now this difference in compute doesn’t matter that much.

当我们分解实验室的算力比例或预算时，到目前为止是 60% 训练，40% 推理。但训练可以进一步分解。实际上，50% 的算力用于研究，10% 用于开发，然后 40% 用于推理。我所说的研究和开发是指：研究人员产生想法，测试新架构，测试新的数据混合，测试新的超参数，新的注意力技术，等等。但最终当他们进行训练运行时，当 Anthropic 训练 Mythos 时，它低于 200 兆瓦。

When we break down the compute ratio or budget of a lab, so far it’s been 60% training, 40% inference. But that training gets broken down further. Actually 50% of the compute is research, 10% of the compute is development, and then 40% is inference. What I mean by research and development is: researchers are generating ideas, testing new architectures, testing new data mixes, testing new hyperparameters, new attention techniques, blah, blah, blah. But ultimately when they do the training run, when Anthropic trains Mythos, it’s sub-200 megawatts.

预训练还是整个过程？

The pre-train or the whole thing?

预训练。低于 200 兆瓦，持续大约两个月。

The pre-train. It’s sub-200 megawatts for, call it, two months.

那强化学习用的算力就更少了。你认为强化学习的算力比预训练还少？

Then the RL is even less. You think the RL was less compute than the pre-train?

至少在单点预训练方面，是的。但总算力可能更高，对吧？但那是顺序进行的。他们某一时刻最多可能用到 200 兆瓦。实际上他们有多个吉瓦，所以他们大部分算力都用于研究，而不是模型开发。

At least in terms of single site of pre-training, yeah. But total compute was probably higher, right? But it's sequential. At most, the most they ever used at one point in time was maybe 200 megawatts. In reality they had multiple gigawatts, so most of their compute was going to the research, not the development of a model.

这其中有原因。协调所有这些集群很难。把它们放在一起很难。做多站点训练很难。做强化学习很难。在强化学习期间生成更多 rollout 并不一定会让它更好。有各种原因导致你可能无法将你拥有的两个吉瓦全部用于训练。实际上，我只能利用 200 兆瓦。

There's reasons for this. It's hard to coordinate all these clusters. It's hard to co-locate all of them. It's hard to do multi-site training. It's hard to do RL. Generating even more rollouts during RL does not necessarily make it better. There's all sorts of reasons why you may not be able to leverage all two gigawatts that you have onto training. Actually, I can only leverage 200 megawatts.

随着我们在自动化编码和自动化研究方面越来越深入，我实际上预计用于研究而非训练的算力预算比例会变得更加模糊，甚至训练占比会更高。还有持续学习之类的事情。所有这些都开始意味着越来越多的算力实际上用于训练模型。

As we get further and further down automated coding and automated researcher, I actually expect the percentage of the compute budget that goes to research versus training to become a lot more fuzzy, or even higher for training. Also things like continual learning. All of these things start to mean that more and more is actually going to training the model.

如果你最终处于每年 100 吉瓦的世界，按当前价格计算，那将是每年 5 万亿美元的资本支出。再加上你必须在更早之前建造发电厂。而且发电厂是 30 年期的资产。再加上数据中心是 15 到 20 年期的资产，你也必须在那时建造。所以这 5 万亿美元，一旦考虑到未来几年的增长，实际上将更像是 7 到 10 万亿美元的资本支出。

If you end up in a world where you're doing 100 gigawatts a year, at current prices, that would be $5 trillion of CapEx every single year. Then stack on the fact that you have to build the power plants way before then. It's also a 30-year asset. You stack on the fact that the data centers are a 15-, 20-year asset, and you have to build that then too. So the $5 trillion, once you account for future years' growth, is actually going to be more like $7 or $10 trillion of CapEx.

等等，我没明白。这还不包括数据中心本身没有发电基础设施的事实。

Wait, I didn't understand. That doesn't include the fact that there's not the infrastructure for the power generation in the data center itself.

对，没错。当你谈论 AI 资本支出时，人们说的是 400 亿、500 亿美元。但那真的只是关键 IT 部分：服务器、网络、光纤、收发器、光通信等等。它没有考虑到数据中心本身或发电厂本身，而这些是提前建造的。

Right, exactly. When you talk about AI CapEx, people are saying $40, $50 billion. But that's really just the critical IT: the servers, the networking, the fiber, the transceivers, optical communications, all this sort of stuff. It doesn't account for the data center itself or the power plants themselves, which are being built ahead of time.

如果我今年建造 100 吉瓦，明年建造 150 吉瓦，那么那 150 吉瓦的所有建筑都需要在今年作为资本支出建造。如果后年建造 200 吉瓦，所有这些发电厂都需要投入……你必须今年就购买涡轮机。所以实际上，如果你建造 100 吉瓦，这甚至比 5 万亿美元还要大得多。

If I'm building 100 gigawatts this year and 150 gigawatts next year, then all of the buildings for that 150 gigawatts need to be built in CapEx this year. If I'm building 200 gigawatts the year after that, all those power plants need to be spent… You have to buy the turbines this year. So actually, it's much bigger than even $5 trillion if you're building 100 gigawatts.

对。很可能，到 2030 年底，每年的增量资本支出将接近 10 万亿美元，这将接近世界经济的十分之一。

Right. Very plausibly, incremental CapEx every year is getting close to $10 trillion by the end of 2030, which is going to be close to a tenth of the world economy.

如果所有这些都发生在美国……美国经济也会增长。但即便如此，按美国经济目前的规模，将会有大约三分之一到四分之一用于数据中心。

If all of it's going up in the US… The US economy will have grown as well. But still, at the current size of the US economy, it'll be like a third to a quarter of the US economy just going towards data centers.

当我大声说出来时，我想，“也许你是对的，我们就是不允许这样，这就是为什么这不会发生。”因为要让这种指数增长继续下去，美国经济的四分之一就只是建造数据中心。

As I say that out loud, I'm like, "Maybe you're right and we just won't allow it, and that's the reason this doesn't happen." Because for this exponential to continue, a quarter of America's economy is just building data centers.

我相信资本主义和资源向最有利可图的事情重新分配。但与此同时，政治存在，信贷市场存在，资本市场存在。

I believe in capitalism and reallocation of resources towards the most profitable thing. But at the same time, politics exist, credit markets exist, and capital markets exist.

所以为了实现，比如说，到 2030 年达到 100 吉瓦……或者我们甚至缩减到 2028 年，那时所有这些项目的资本支出约为 3 到 4 万亿美元：超过 2.5 万亿美元用于 IT 资本支出，另外 1 到 2 万亿美元用于数据中心和能源，以及所有下游供应链，比如半导体等等。

So to enable, let's say, that 100 gigawatts by 2030… Or let's even pare it down to 2028, where it's like $3 or $4 trillion of CapEx across all of these items: over $2.5 trillion towards IT CapEx, and then another $1 to $2 trillion on data center and energy, and all the supply chain downstream, like semiconductors and all that stuff.

如果你有 3 到 4 万亿美元的资本支出，所有这些现金从哪里来？

If you're at $3 or $4 trillion of CapEx, where does all this cash come from?

还没有人从业务中产生那么多现金。超大规模企业迄今为止资助了所有增长。谷歌、微软、亚马逊、Meta。他们资助了很大一部分。他们占了超过一半的算力，但他们现在不产生现金。他们实际上把所有东西都花在资本支出上。此外，他们举债并把所有东西都花在资本支出上。你看到 Meta 这样做了，甚至亚马逊、谷歌也这样做了。微软很快就会这样。每个人都在举债来支付他们的资本支出。

No one is generating that much cash from the business yet. Hyperscalers funded all of the growth up until now. Google, Microsoft, Amazon, Meta. They funded a huge percentage of it. They were more than half of compute, but they now don't generate cash. They actually spend everything on CapEx. In addition, they raise debt and spend everything on CapEx. You've seen Meta do it, even Amazon, even Google. Microsoft will be there soon. Everyone is raising debt to pay for their CapEx.

那么谁是以前没有这样做而现在增量支付这笔费用的人呢？

Now who is the incremental person to pay for this that was not doing it before?

以谷歌为例，他们停止回购股票，或者 Meta 停止回购股票，转而购买计算基础设施，这相当简单。这对市场没有巨大影响，但确实有一些影响。但当你迈向 2028 年——超大规模企业现在举债数千亿美元，然后他们所有的供应链也举债数千亿美元——谁来为此买单？

In the case of Google, it was pretty simple for them to stop doing buybacks, or Meta stop doing buybacks, and turn around and buy computer infrastructure. That doesn't have a huge effect on the market, but it does have some effect. But as you step forward to 2028 — where the hyperscalers are now raising hundreds of billions of dollars of debt, and then all of their supply chain is raising hundreds of billions of dollars of debt — who pays for this?

所以有几种不同的方式。有像 Nvidia、Broadcom 和内存公司这样的半导体公司转身决定资助部分资本支出。有传统的基础设施投资者，他们聚集资本并投资于基础设施。不是桥梁，而是数据中心。最后，经济中的每个人都在意识到，“也许我不应该买房，或者也许我不应该投资于帮助人们买房的信贷，或者也许我不应该购买政府债券。我应该购买超大规模企业的债券，或者我应该购买这个数据中心的债券，或者我应该购买 Anthropic 的债券。因为 Anthropic 愿意为额外的十亿美元支付 20% 的利率来建设他们的产能。因为他们知道从中获得的收入将是巨大的，而且他们会支付 20%，因为这仍然比从 SpaceX 以每吉瓦 500 亿美元租用要好。”

So there's a few different ways. There's semiconductor companies like Nvidia and Broadcom and the memory companies turning around and deciding to fund some of this CapEx. There's the traditional infrastructure investors who are gathering capital and investing in infrastructure. Instead of bridges, it's data centers. Then lastly, there's everyone in the economy who's realizing, "Maybe I shouldn't buy a home, or maybe I shouldn't invest in credit that's helping people buy homes, or maybe I shouldn't buy government debt. I should just buy hyperscaler debt, or I should buy this data center's debt, or I should buy Anthropic's debt. Because Anthropic's willing to pay 20% rates for the incremental billion dollars to build their capacity. Because they know their revenue from it's going to be huge, and they're going to pay 20% because it's still better than renting it from SpaceX for $50 billion a gigawatt."

所以你有所有这些争论。但如果你现在这样做，整个世界经济真的会重新洗牌。

So you've got all of this contention. But if you now do this, the whole world economy is really shifted around.

Antithesis 是一个确定性软件测试平台，能够实现完美的可重现性。它还解锁了一些非常疯狂的调试方法，比如时间旅行。

Antithesis is a deterministic software testing platform that enables perfect reproducibility. It also unlocks some pretty insane approaches to debugging. Like time travel.

借助 Antithesis，你可以跳到轨迹中的任意一点，从那里重新开始。所以当发生崩溃时，你可以倒回到出问题的确切时刻，并冻结整个系统：应用程序、数据库，甚至环境本身。这让你能做到原本不可能做到的事——在完全相同的瞬间观察分布式系统的每一个部分。时间旅行还允许你为已经发生的事件添加遥测和日志。例如，你可以倒回到崩溃前五秒，决定捕获所有网络流量。最强大的是，Antithesis 为你提供了一个进入系统的实时终端，你可以用它来扰动任何你想要的东西。杀掉一个节点或禁用某个功能，然后按下播放键，看看会发生什么。然后再回去尝试别的。在生产环境中，这种破坏性分析往往只有一次机会。比如，如果你重启一个死锁的服务，你需要研究的那个死锁就消失了。但有了 Antithesis，原始时间线总是可重现的，所以你可以根据需要测试尽可能多的假设。如果你不想自己进行这些时间旅行，你可以通过 Antithesis API 让你的智能体代劳。访问 antithesis.com/dwarkesh 了解更多。

With Antithesis, you can jump to any point in a trajectory and start from there. So when there's a crash, you can rewind to the exact moment that something went wrong and freeze the entire system: the application, the database, even the environment itself. This lets you do something that would otherwise be impossible, which is to observe every part of a distributed system at the exact same instant. Time travel also allows you to add telemetry and logging to an event that has already happened. For example, you can rewind to five seconds before a crash and decide to capture all the network traffic. Most powerfully, Antithesis gives you a live terminal into your system that you can use to perturb anything you wish. Kill a node or disable a feature, then hit play and see what happens. Then go back and try something else. In production, you often only get one shot on goal with this sort of destructive analysis. If you restart a deadlocked service, for example, the exact deadlock you needed to study disappears. But with Antithesis, the original timeline is always reproducible, so you can test as many hypotheses as you need. And if you don't want to do all this time traveling yourself, you can just have your agents do it for you via the Antithesis API. Go to antithesis.com/dwarkesh to learn more.

过去几天，我们一直在私下争论，AI 是否会导致主权债务危机。逻辑是这样的。正如我们提到的，现在的情况是，很少的投资就能变成很多钱。所以回报率——

You and I have been debating off air for the last few days whether there will be a sovereign debt crisis as a result of AI. The logic is this. As we were mentioning, you have a situation where very little investment turns into a lot of money. So the rate of return—

这他妈的真是个问题，老兄。我的天。简直不敢相信。

What a fucking problem, dude. Oh my God. Can’t believe it.

不，这对其他所有不能把很少的钱变成很多钱的人来说，是个巨大的问题。所以回报率高得离谱。即使在数据中心层面，如果你建一个数据中心，试图以折旧成本 10 倍的价格租给 Anthropic 或 OpenAI，那简直疯了。你年底能把 1 美元变成 2 美元或 10 美元。这推高了利率。现在，如果利率上升，并且对整个经济都如此……人们借越来越多的钱。他们在与政府本来会进行的贷款、其他公司本来会进行的贷款，或者你作为消费者或房贷买家本来会进行的贷款竞争。这让其他所有人的借贷成本都更高了。这对无数人都有巨大影响。抱歉，我在这里要开始一段独白了，但我们一直在共同思考这个问题。我认为美国最终会没事。因为如果数据中心建在美国，基本上可以直接对数据中心征税。但按照目前的税收体系，企业收入占联邦收入不到 10%。80% 以上是工资税和所得税，而随着自动化程度越来越高，这部分会缩水。同时，在支出方面，目前 20% 的税收支出用于偿还债务，支付债务利息。现在，很多债务是短期的，所以每五年展期一次。

No, it is a huge problem for everybody else who can’t turn a little money into a lot of money. So the rate of return is incredibly high. Even at the data center level, if you build a data center and you’re trying to get rented out to an Anthropic or an OpenAI for 10x what it costs you on a depreciated basis to build it, it’s fucking crazy. You turn $1 into $2 or $10 or something at the end of the year. That raises the rate of interest higher. Now, if the rate of interest goes higher, and if it does that for the entire economy… People are borrowing more and more money. They’re competing against the other lending that the government would’ve done, or that other companies would’ve done, or that you as a consumer or a mortgage buyer would’ve done. That’s making it more expensive for everybody else to borrow. This has huge implications for tons and tons of people. Sorry, I’m going to go on a bit of a monologue here, but we’ve been thinking about this together. I think the US will be fine at the end of the day. Because if the data centers are built in America, you can fundamentally just tax the data centers. But the way the current tax system is set up, corporate income is less than 10% of federal revenues. 80%-plus is payroll taxes and income taxes, which, as more and more automation happens, will shrink. At the same time, on the spending side, currently 20% of tax revenue spending goes towards servicing the debt, paying interest payments on the debt. Now, a lot of the debt is short duration, so it rolls over every five years.

你他妈的为什么在笑？

Why are you fucking laughing?

因为这些都是你上个月才学到的东西。

Because it’s things you’ve learned in the last month.

说得好像你不一样似的。你可是拿了金融经济学学位的人。我没有。互联网还以为我是养蜂的呢。

Like it’s any different for you. Like you got a degree in fucking financial economics. I didn't. The internet thinks I’m a beekeeper.

几个月，几个月。这是我们的行当，Dylan。

Few months, few months. This is our business, Dylan.

我知道，我知道。抱歉，抱歉。现在我都开始不自在了。靠。

I know, I know. Sorry, sorry. Now I’m self-conscious. Fuck.

不，挺好的。你做得不错。我只是觉得好笑。

No, it’s good. You’re doing good. I just think it’s funny.

一百万人听这个家伙讲，而他这个月才刚学了债务。

A million people listen to this guy who just learned about debt this month.

假设利率上升 1%。在五年期基础上，用于偿债的税收比例从 20% 上升到 25%。如果上升 5 个百分点，那就会超过 40%。但如果考虑到政府每年借入 2 万亿美元，那就会从 40% 上升到 60% 以上。所以 60% 的税收收入都用于支付债务利息。现在，我认为美国会没事，因为如果我们允许数据中心在美国建设，税基就会增加。其他国家在我看来绝对完蛋了。

Suppose the interest rates rise 1%. Over a five-year basis, the fraction of tax revenue that goes towards servicing the debt goes from 20% to 25%. If it rises 5 percentage points, that would go north of 40%. But if you take into account the fact that the government is borrowing $2 trillion every single year, then that goes from 40% to north of 60%. So 60% of tax revenue just goes towards paying interest payments on the debt. Now, I think the US is going to be fine because the tax base will increase if we let data centers get built in America. Other countries are absolutely fucked, in my opinion.

我刚才在看哪些国家债务很多、税收很少，而且很多债务偿还频率很高。这些国家，比如巴基斯坦或尼日利亚，我认为在新的利率体制下会非常惨。

I was just looking at which countries have a lot of debt, have very little tax revenue, and also a lot of their debt is serviced quite often. Those countries, like Pakistan or Nigeria, I think are just going to be very fucked in this new interest-rate regime.

这种挤出效应就是为什么不是“YOLO 10 亿吉瓦”的原因。你有所有这些大量使用债务的行业和国家，你之前提到的那些贫困国家，它们只会违约。还有消费品公司，所有那些在 Trader Joe's 之类地方看到的东西的制造商。它们大量使用债务。所有这些电信公司也大量使用债务。银行也大量使用债务。所以如果市场利率上升——不一定是政府设定的利率，而是政府所说的联邦利率与其他所有人收取的利率之间的利差，因为亚马逊明年要借 1000 亿美元债务，或者不管具体数字是多少，可能更少——你最终会面临一个非常棘手的问题：现金从哪里来？有一部分是由现金流提供资金的，而且现金流持续增长。但合乎逻辑的做法是投资远超现金流，因为未来几年的回报会非常惊人。所以就有了这个差额。然后压低这个差额的是所有这些其他因素：针对数据中心的监管、消费者的愤怒、政客的愤怒、针对 AI 的监管、AI 实验室出于安全原因不发布最新模型。利率上升对所有这些都有影响。

This crowding-out effect is the reason it’s not YOLO 1 billion gigawatts. You’ve got all these industries and countries that use a lot of debt, all these impoverished countries that you mentioned earlier that are just going to default. You’ve got consumer packaged goods, all of these companies that make things you see at Trader Joe’s or wherever. They use a lot of debt. All these telecom companies use a lot of debt. Banks use a lot of debt. So if interest rates go up in the market — not necessarily the government-set interest rate, but the spread between what the government says their federal rate is versus what everyone else is charging, because Amazon wants to raise $100 billion of debt next year or whatever the hell the number is, probably less — you end up with this really challenging problem of, where does the cash come from? There is some level that is funded by cash flows and the cash flows keep going up. But the logical thing to do is to invest way more than your cash flows because then the returns in future years will be amazing. So you have this delta. Then what’s pushing down on the delta is all of these other things: regulations against data centers, consumers getting mad, politicians getting mad, regulations against AI, the AI labs not releasing their latest models because of safety reasons. Interest rates going up are an influence on all of these things.

所以所有这些因素都会把曲线从资本主义在纯粹、简单的经济学意义上想要的方向，拉向我们这个复杂系统想要的方向，并且越拉越低，以至于最终建成的吉瓦数会低于本应建成的数量。

So all of these things bend the curve from what capitalism wants in terms of pure, simple economics to what the complex system that we have wants, and bend it lower and lower to where not as many gigawatts as should be built will be built.

嗯，利率是资本主义的一部分，对吧？

Well, the interest rate is part of capitalism, right?

是啊，但那是简单经济模型，而不是我们实际所处的更复杂的现实。

Yeah, but in the simple economic model versus the more complex what we have.

你觉得明年亚马逊或 Anthropic 之类的公司发行债券融资的利率会是多少？如果他们借几千亿美元的债，平均利率是多少？

What is the rate at which you think Amazon or Anthropic or whatever will be issuing bonds for debt next year? If they do hundreds of billions of dollars of debt. What is the average rate?

我不认为亚马逊会借几千亿美元的债。

I don’t think Amazon will do hundreds of billions of dollars of debt.

我是说总计。比如说大型科技公司，所有超大规模云厂商加起来，所有云……

In total. Let’s say the big tech guys. The hyperscalers in total, and all the clouds…

在我们做的建模中，2024 到 2029 年大约有 11 万亿美元的资本开支。

In the modeling that we do, we have about $11 trillion of CapEx from 2024 to 2029.

总计？

Total?

总计。如果你尽可能多地用现金流来融资，最终仍会有超过 5 万亿美元的信贷需要发行，来支撑这 11 万亿美元以上的建设。

Total. If you fund a lot of this with cash flows, as much as you can, you still end up with north of $5 trillion of credit that needs to be issued for this $11 trillion-plus build out.

所以你不认为 AI 收入会继续每年翻三倍？

So you don’t think the AI revenue continues even 3x-ing year over year?

AI 收入确实在增长。但我不认为它能永远增长而不触及某些约束。实验室会有某些激励。在很多情况下，实验室并不是所有算力的建设方，尽管他们越来越倾向于自己建。但他们会有所有这些现金流。

AI revenue does go up. I don’t think it can go up forever without certain constraints being hit. Labs will have certain incentives. Labs are not the ones building all the compute in many cases, even though they’re increasingly trying to go that way. But they’ll have all this cash flow.

你说收入会是多少？你觉得他们不会有那么多收入？

How much did you say the revenue will be? You think they’ll not have that much revenue?

不，我只是说直到 2029 年，资本开支大约有 11 万亿美元。其中 6 万亿用现金融资，5 万亿用债务融资。如果是这样，整个生态系统筹集 5 万亿美元的债务确实会推高利率。

No, I’m just saying till 2029 there’s something on the order of $11 trillion of CapEx. $6 trillion of that is funded with cash, and $5 trillion of that is funded with debt. If that’s the case, $5 trillion of debt being raised across the whole ecosystem does make interest rates go up.

那什么能阻止这种情况？

Then what prevents that?

有几件事。第一，实验室会不会提高每兆瓦的收入，并保持推理分配很大？如果是这样，他们就在积累整个标普 500 的利润，因为每个人都在花钱降低成本。当然，他们的利润也会上升，但现金必须来自某个地方。所以他们的收入增长速度相对于他们为世界创造的价值，有一个上限。而且技术还有扩散的一面。但最终实验室的收入会持续上升。他们无法用现金流为所有东西融资。最优的方案其实是尽可能多地用信贷来融资，因为即使实验室的现金流能覆盖很多东西，你想建的比那更多。所以会有一定量的信贷被创造出来。我们目前的建模显示，到 2029 年有 5 万亿美元的信贷和 6 万亿美元的现金基础设施投资。当你把这些算进去，相对于 AI 模型的需求增长，算力还是不够。

There’s a couple things. One, do labs increase their revenue per megawatt more and keep inference allocations large? In which case, they’re accumulating all the profit across the S&P 500 because everyone’s paying to reduce their costs. Of course, their profits will also go up, but cash has to come from somewhere. So there’s an upper limit on how fast their revenue can grow versus the value they deliver into the world. And there’s a diffusion aspect of the technology. But ultimately labs’ revenues keep going up. They can’t cash-flow fund everything. The optimal scenario is you actually use credit as much as you can to fund, because even if cash flows from the labs fund a lot of stuff, you want to build more than that. So there is some amount of credit that gets built. Our current modeling has $5 trillion of credit and $6 trillion of cash-funded infrastructure investments through ’29. When you take that, this is not enough compute relative to what the demand growth is from the AI models.

所以你得到了显而易见的答案，就是每兆瓦收入持续上升。这说得通。你觉得到 2029 年，这一切会导致利率上升多少？

So you’ve got the obvious answer, which is revenue per megawatt keeps going up. That makes sense. How much do you think interest rates will increase by 2029 as a result of all this?

老兄，这纯粹是拍脑袋给个数，但如果你要拍脑袋……世界经济增长会大幅上升，那亚马逊的利率凭什么不从现在的位置往上涨？这绝对是拍脑袋，但最近 Meta 以 5% 到 6% 的利率融资。我看不出他们为什么不会付 8%。他们会很乐意付 8%，因为他们要建的算力带来的回报是巨大的。市场不希望他们付，但他们自己会想付 8%。另一面是，如果他们付 8% 而不是现在的 5%、5.5%、6%——上升 250 个基点——那会让经济中其他所有人也多付 250 个基点，这会引发很多事情。银行会尖叫，因为如果他们的信用利差扩大，他们的债务重新定价的速度会比资产重新定价更快。如果信用利差爆掉，他们最终会损失惨重。

Dude, this is vibing a number, but if you’re vibing a number out… Growth in the world economy is going up a lot, so why wouldn’t interest rates for Amazon go up from where they are today? This is going to be extremely vibed out, but recently Meta’s raised at 5 to 6%. I don’t see why they wouldn’t pay 8%. They would happily pay 8% because the return from the compute that they’re going to build is humongous. The market won’t want them to, but they’ll want to pay 8%. The flip side is that if they pay 8% versus the 5%, 5.5%, 6% they do today — a 250 bps increase — that makes everyone else in the economy also pay 250 bps more, which then causes a lot of things. Banks will scream, because if their credit spread goes up, their debt reprices faster than their assets reprice. They ultimately end up losing tons of money if their credit spread blows up.

这件事的另一个后果——这是你提出的观点——是如果利率上升，贴现率就会上升，这意味着所有股票的贴现现金流会暴跌。这意味着即使整个股市可能表现不错——标普 500 会没事——任何个股的价值可能都会暴跌，尤其是巴菲特、伯克希尔那种，连续 30 年派发良好现金流的股票。

The other consequence of this — this is a point you made — is that if interest rates rise, the discount rate increases, which means that the discounted cash flows of all equities crater. Which means that even though the stock market as a whole might be doing fine — the S&P 500 will be fine — any individual stock will probably have just cratered in value, especially the Buffett, Berkshire type, pay-good-cash-flows-for-30-years type stocks.

是啊。就像，“我为什么要为强生付这么多钱？”他们被视为稳定股票：现金流好，会随时间返还现金流。或者一家铁路公司。如果我的贴现率不是 3% 或 5%，而是 8% 或 10%，我凭什么投那么多钱？

Yeah. It’s like, "Why would I pay this much for Johnson & Johnson?" They’re seen as a stable stock: good cash flows, they’ll return their cash flows over time. Or a railway company. Why the fuck would I invest that much if my discount rate isn’t 3% or 5%? It’s now 8% or 10%.

对于发展中国家……Basil Halperin，我的好朋友，也是一位经济学家，提出过这个观点：我们会看到第二次沃尔克冲击。80 年代，为了对抗通胀，美联储主席保罗·沃尔克把利率提高了超过 5%，实际利率大约 8%。那导致大约 40 个不同的国家，主要在拉丁美洲，在那个十年里违约。我认为这很可能会再次发生。

For developing countries… Basil Halperin, who’s a good friend and an economist, made this point that we’ll see a second Volcker shock. In the ’80s, to fight inflation, Fed Chair Paul Volcker raised interest rates more than 5%, something like 8% real interest rate. That caused some 40 different countries, mostly in Latin America, to default in that decade. I think that will probably happen again.

好了，现在我们开始聊奇点了。我们一直在谈利率上升会怎样——

Okay, now we’re getting into singularity talk. We’ve been talking about what happens if interest rates rise—

顺便说一句，我认为这一切都发生在奇点之前。

I think this all happens before singularity, by the way.

是啊，我就是这个意思。我们刚才在谈奇点之前，利率上升 2-3% 等等。在某个时点，我认为世界经济非常有可能每年翻一番。这不是五年内会发生的事。但最终会实现。有一位研究者，Damon Binder，在这方面做了很棒的工作。如果你看一个完全自动化经济中的投入产出表……要让经济中所有东西的总存量每年翻一番，需要什么条件？

Yeah, that’s what I’m saying. We were talking about before singularity, interest rates rise 2-3%, et cetera. At some point, I think it’s very likely that the world economy will be doubling every single year. This is not happening in five years. But it’ll happen eventually. There’s this researcher, Damon Binder, who’s done great work on this. If you look at input-output tables in a fully automated economy… What would it take to double the entire stock of things in the economy every single year?

嗯。

Yeah.

如果经济每年增长 3%，那么用 70 法则，就是二十多年。

If the economy grows at 3% a year, then rule of 70, that’s 20-something years.

对。

Right.

但他说，“好吧，现在我们受限于人的因素，你不能每年让人口翻一番。”但在一个你也能每年让劳动力翻一番的世界里，经济能增长多快？我认为它可以每年翻一番。至少也会是每年几十个百分点。

But he was like, "Okay, right now we’re bottlenecked by the fact that there’s people, and you can’t double people every single year." But in a world where you can also double the labor force every single year, how fast can the economy grow? I think it could double every single year. At the very least it would be tens of percent every single year.

好吧。

Okay.

利率应该会非常接近增长率。因为消费的存在，不会完全相等，但应该会非常相似。然后我们会进入一个世界，我认为在 2030 年代，利率会达到几十个百分点。我脑子里有一部分在想：“可能会是几百个百分点”，但我们就说至少是几十个百分点吧。我就觉得，好吧。所有不参与 AI 生产的国家都会违约。所有不是 AI 股的股票基本上都一文不值，因为折现现金流毫无价值。如果联邦政府想不出办法对 AI 征税，那么偿还债务的金额就会超过当前的税收收入。而且还有所有这些其他影响，我敢肯定我们甚至都没有考虑到：你无法获得抵押贷款，等等等等。

The rate of interest should be pretty close to the growth rate. It won't be exactly that because of consumption, but it should be pretty similar. Then we'll go into a world, I think in the 2030s, where the rate of interest is tens of percent. Part of my brain is like, 'It might be hundreds of percent,' but let's say it's at least tens of percent. I'm just like, okay. Every country that is not involved in the production of AI defaults. Every stock that is not an AI stock is worth basically zero because discounted cash flows are worth nothing. If the federal government can't figure out a way to tax AI, servicing the debt is more than the current tax revenue. And you have all these other effects that I'm sure we're not even pricing in: you can't get a mortgage, et cetera, et cetera.

从根本上说，这个世界正在发生什么？这都是书呆子话，对吧？但让我们退一步。发生了什么？

Fundamentally, what is happening in this world? This is all nerd speak, right? But let's step back. What's happening?

刚才开始的就是书呆子话？我们会进入一个完全不同的增长体制。经济基本上在说：“嘿，现在政府借钱支付养老金的机会成本极高。因为这笔钱本可以用于建造一个机器人工厂，而这个工厂又能建造另一个机器人工厂，如此循环。”资本的机会成本将大幅增加。这从根本上就是我们讨论的所有这些事情的根源。随着利率上升，股市会受到重创。即使是 AI 公司也不例外。有些真正相信 AI 的人会说：“为什么美光、海力士或铠侠的市盈率只有 2 到 3 倍？”这就像是在说：“好吧，如果你真的被 AI 洗脑了，那么经济中的一切都应该以 2 到 3 倍的市盈率交易。”如果你没有被 AI 洗脑，那么当然，它们是在超额盈利。这是一个论点，说明为什么——我认为内存会表现很好——内存股不应该再涨 10 倍或什么的。因为如果我们处于一个对内存有如此大需求的市场——这意味着 AI 已经导致了经济的剧烈变化——那么一切都应该以 2 到 3 倍的市盈率交易，股市应该他妈的崩盘。从某种意义上说，Meta 的估值……我认为它们是一家大约 1.5 万亿美元的公司。这算什么？愚蠢。它们的价值远不止于此，至少在逻辑上是这样。你只要看看它们的现金流、它们囤积的所有基础设施，以及它们将能够以每瓦特疯狂的价格出售的所有算力，无论是作为代币，因为它们的实验室成功了，还是直接卖给 Anthropic 和 OpenAI。

Just now it started, the nerd speak? We'd be entering a totally different growth regime. The economy's basically saying, 'Hey, the opportunity cost of the government borrowing money to pay people pensions is extremely high now. Because that money could be spent building a robot factory that builds a robot factory that builds a robot factory.' The opportunity cost of capital is going to increase a ton. That's fundamentally the cause of all of these things we're talking about. As interest rates go up, equity markets get pummeled. Even AI companies. Some people who really believe in AI are like, 'Why does Micron or Hynix or Kioxia trade at 2 or 3 times earnings?' It's like, 'Well, if you're really AI-pilled, everything in the economy should trade at 2 or 3 times earnings.' If you're not AI-pilled, then sure, they're over-earning. It's an argument for why — I think memory is going to do great — memory stocks shouldn't 10x or whatever again. Because if we're in the market where there's that much demand for memory — which means AI's caused this drastic change in the economy — then everything should trade at 2 or 3x multiples and the stock market should fucking crash. In a sense, Meta trading at… I think they're like a $1.5 trillion company. It's like, what? Silly. They're worth way more than that, at least in a logical sense. You just look at their cash flows, all the infrastructure they're hoarding, and all the compute that they're going to be able to sell for crazy amounts of dollars per watt, either as tokens because their lab works, or just to Anthropic and OpenAI.

这最终变成了一个问题：你必须将所有资本重新分配给 AGI。你通过把其他所有人挤出市场来实现这一点。所以 AGI 的限制因素不是研究工程师（比如我们的室友 Sholto）能多快地转动齿轮。实际上，只是世界其他地方允许这种情况发生多少？因为他们会进行监管。他们显然会提高利率。他们会说：“不要建数据中心。”他们会说：“停止建造晶圆厂。”他们会说：“哦，该死，每家公司的股权价值都在暴跌，那我怎么才能为 AI 买单来提升我的业务？”那么，Anthropic 和 OpenAI 就必须开始自己建造东西。他们已经在制造自己的芯片，或者至少在设计自己的芯片，而且这还会扩展。他们正在签订自己的数据中心合同，并在未来几年内建设自己的基础设施。这里有一个问题，即经济的这种重新分配如何发生。有很多下行压力，使得它不会只是直线起飞，即使模型有能力做到这一点。我认为你我都相信我们正处于一个模型有能力做到这一点的世界。但缓慢起飞，至少是我的希望，是可能的，因为经济和监管领域的一切。

It ultimately becomes a question of, you have to reallocate all the capital to the AGI. You do that by pricing everyone else out. So the limiter on AGI is not how fast the research engineers, like our roommate Sholto, can crank the gears. It's actually just how much does the rest of the world let that happen? Because they're going to regulate. They're going to obviously increase interest rates. They're going to say, 'No data centers.' They're going to say, 'Stop building fabs.' They're going to say, 'Oh shit, every company's equity value is tanking, so how can I pay for AI to increase my business?' Well then, Anthropic and OpenAI have to start building their own stuff. They're building their own chips already, or at least designing their own chips, and it'll expand out. They're contracting their own data centers and building their own infra in the next couple years. There's the question of how this reallocation of the economy happens. There's a lot of downward pressure on it not being just straight takeoff, even if the models were capable of it. I think you and I believe we're in a world where models are capable of that. But slow takeoff is, at least my hope, possible, because of everything in the economy and regulatory world.

政府说：“不要发布你的模型，”政府说：“实际上，你甚至不能在内部大量使用你的模型，”因为这很快就会发生。他们已经在说你不能发布你的模型了。

Government saying, 'Don't release your models,' the government saying, 'Actually, you can't even use your models internally that much,' because that's going to happen soon. They're already saying you can't release your models.

我最担心的是奇点，而外部部署实际上对此有帮助。所以，我们阻止外部部署这一事实是愚蠢的。这能阻止奇点吗？目前，这会导致更多收入，因为模型没有能力进行递归自我改进。但我担心的是这样一个世界：到了 2030 年，政府会说：“我们要等六个月，你才能向公众发布你的最新模型。”六个月，100 倍。来吧。在这六个月里，他们在内部进行递归自我改进。公司里会发生各种疯狂的事情。与此同时，我们其他人只能被困在那些按当前速度落后数年的模型上。

The thing I'm most worried about is a singularity, which external deployment is actually helping. So the fact that we're preventing external deployment is stupid. Does that prevent singularity? Right now it would lead to more revenue, because the models are incapable of RSI. But I'm worried about a world where it's 2030 and the government's like, 'We're going to wait six months before you can release your newest model to the public.' Six months, 100x. Let's go. In that six months, they do recursive self-improvement internally. They just have all kinds of crazy shit happening in the company. Meanwhile, the rest of us are stuck with models that are, at current pace, years behind.

我的想法是这样的。假设整个世界都参与了这个阴谋，试图减缓 AI 的发展。

Here's my thought. Suppose that the whole world gets in on this conspiracy to try to slow down AI.

我不认为这是一个阴谋。这是每个政治家都公开写出来的。

I don't think it's a conspiracy. It's outwardly written from every politician.

假设他们把 AI 减缓了一年。如果算力每年增长 2 到 3 倍，他们阻止了一整年的 AI 部署，这样你就会比原本的情况落后一年。在递归自我改进期间，你在一年内就能获得 3 到 6 年的 AI 进展。但他们不仅限制算力。他们还限制了实验室在内部发布模型的能力。我们看到了这一点。如果他们这样做了，那就理想了。

Suppose they slow down AI by a year. If compute is increasing 2 to 3x every single year, they prevent a whole year of AI deployment such that you're a year behind where you would otherwise have been. During RSI, you're getting 3 to 6 years of AI progress in a single year. But they don't just limit compute. They also limit the lab's ability to release the model internally. We saw that. If they did that, that would be ideal.

Anthropic 不得不暂时停止向外国员工提供 Mythos。

Anthropic had to stop giving Mythos to foreign employees for a bit.

我不知道这是真的，内部也是如此？

I didn't know that was true, internally as well?

他们是这么声称的。我以为那只是一个不同的检查点，不是 Mythos，但基本上就是 Mythos。但诸如此类的事情也不会被允许。政府是愚蠢的，但他们不会那么愚蠢，至少我希望如此。政府——至少是掌握着筹码的美国联邦政府——不会希望 Anthropic 在内部使用 Mythos 4。他们会说：“他妈的，等等。慢下来，”因为所有这些监管原因。每个当选的人都会讨厌 AI。即使是已经当选的人也已经讨厌 AI。所有的选民。

That's what they claimed. I thought that was just a different checkpoint that was not Mythos, but it was basically Mythos. But stuff like that is not going to be allowed either. The government is dumb, but they're not that dumb, I would hope, at least. Governments — at least the US government, which has the cards here — are not going to want Anthropic to use Mythos 4 internally. They're going to be like, 'Hold the fuck on. Slow down,' because of all of these regulatory reasons. Everyone who's elected is going to hate AI. Even the people who are elected already hate AI. All the constituents.

我敢打赌，总有一天你父母会打电话给你说：“Dwarkesh 宝贝，你干得太糟了，你在让 AI 进展得更快。”

I bet you at some point your parents are going to call you and be like, "Dwarkesh beta, you're doing a terrible job. You're making AI progress happen faster."

因为我的播客，我在加速 AI 进展？

Because of my podcast I'm accelerating AI progress?

也许吧。你教育了人们。如果他们更聪明，也许他们会让 AI 进展得更快。

Maybe. You educate people. Maybe if they're smarter, they're progressing AI faster.

不管怎样，你将会面临 AI 进展、发展和部署的现实约束。即使最终会发生，我们可能在到达那之前就自我撕裂了。

Anyway, you're going to have real-world constraints on the progress and development and deployment of AI. Even though it will happen eventually, we could tear ourselves apart before we get there.

Jane Street 现在正在招聘两个独立的机器学习实习岗位：一个专注于机器学习工程，另一个专注于机器学习研究。我和 Alok 坐下来聊了聊，他帮助管理研究方向的实习项目，以了解更多信息。

Jane Street is hiring for two separate ML internships right now: one focused on ML engineering and the other focused on ML research. I sat down with Alok, who helps run the research track, to learn more about that program.

我认为这个领域从根本上研究不足。在我们的深度学习研究团队中，经常有未解答的问题，比如我们不了解某些市场参与者的行为或交易发生的某些动态。这些未解答的问题非常适合作为实习生项目，因为它们最终是我们关心的主题，只是我们还没有时间去解决。所以即使作为实习生，你也会为真正的研究做贡献，而不是在做某种人为设计的练习。

I think this domain is fundamentally understudied. Often we have unanswered questions within our deep learning research team where we don't understand, say, some market participants' behaviors or certain dynamics of how trading happens. These unanswered questions make for really good intern projects because they are ultimately topics that we care about and just haven't gotten around to figuring out yet. So even as an intern, you'll be contributing to real research, not working on some sort of contrived exercise.

Jane Street 团队密切关注前沿大语言模型研究。一个相对常见的实习生项目是将最近的论文应用到金融市场，这些市场有自己的一套棘手问题。最终，我们试图建模数千个相互连接的不规则时间序列。信噪比极低，因为我们有很多竞争对手也在做同样的事情。所以我们正在解决的是一个对抗性的、非平稳的、极高维度的问题。需要说明的是，你不需要了解任何金融知识就能胜任。只要你有机器学习研究的背景，Jane Street 可以教你其余部分。他们的 2027 年实习申请现已开放。请在 janestreet.com/dwarkesh 申请。

The Jane Street team follows frontier LLM research closely. A relatively common intern project is adapting a recent paper to financial markets, which come with their own set of gnarly problems. Ultimately, we're trying to model thousands of interconnected irregular time series. The signal-to-noise ratios are extremely low because we have a lot of competitors trying to do the same. So we have this adversarial, non-stationary, extremely high-dimensional problem that we're trying to solve. To be clear, you don't need to know anything about finance in order to be a good fit. As long as you have a background in ML research, Jane Street can teach you the rest. Their 2027 internship applications are open now. Apply at janestreet.com/dwarkesh.

我觉得这些场景中疯狂的一点是，世界未来的劳动力供应最终会集中在极少数公司手中，而且这个劳动力供应每年增长得有多快。如果前沿的算力（以 FLOP 计）每年增长 4-5 倍——而且达到一定能力水平所需的算力每年下降 3 倍——基本上，前沿实验室的有效 AI 人口规模每年增长 10 倍。这在目前并不太重要，因为 AI 还不足以胜任完整的工作，或者像人一样自主地工作或策划阴谋等。但如果当前趋势继续，你会看到一个世界，OpenAI 从今年大约 1000 万 AI 劳动力，到明年 1 亿，再到后年 10 亿。很快，即使 Scaling（规模扩张）放缓，用不了几年，每个公司单独拥有的劳动力当量就会超过地球上的人口。

One thing I find crazy about these scenarios is just how much of the world's future labor supply ends up in very few companies, and also how fast that labor supply grows year over year. If compute at the frontier in FLOP terms is growing 4-5x a year — and further the compute required to achieve a level of capabilities is decreasing 3x a year — basically the effective AI population size at the frontier labs is increasing 10x year over year. That doesn't really matter that much right now, because AIs are not good enough to do full jobs or be as autonomous as people in their capacity to do work or pull off schemes or whatever. But if the current trend continues, you have a world where OpenAI goes from having, say, basically 10 million AI laborers this year to 100 million the next year, to a billion the year after that. Pretty soon, even if compute scaling slows down, it doesn't take many more years before each company individually has more labor equivalence than there are people on Earth.

我认为到本十年末，这非常可能：在一个实验室内部，AI 劳动力、有效人口会比地球上的人口还多。

I think that's very plausible by the end of this decade, that there's more AI labor, more effective population, within a single lab than there are people on Earth.

我们经常谈论由于国有化或其他原因导致的权力集中。但我们没有充分思考这样一个事实：我们实际上正在快速进入一个体制，其中大多数“人”（以工作产出计）集中在两个实验室，它们消耗着世界上越来越多的算力。如果这些 AI 不对齐，那么基本上世界的大部分都不对齐，因为世界的大部分思想都在那里。但即使它们对齐，也只有极少数公司拥有很大的影响力或控制力。

We talk often about centralization of power because of nationalization or whatever. But we don't think enough about the fact that we're actually moving very fast into a regime where most "people", in terms of work output, are concentrated within two labs who are consuming more and more of the world's compute. If these AIs are misaligned, then most of the world is misaligned, basically, because most of the world's minds are there. But even if they're not, very few companies have a lot of influence or a lot of control.

最近有一场争论，我记得 Gavin Baker 说：“Dario 相信世界上只会有一家公司。”然后 Sholto 和 Dario 出来说：“不不不，我们没这么说。”但最终，如果你相信 RSI（递归自我改进），如果你相信实验室是算力最有效的使用者，能从算力中产生最大价值，那么唯一会发生的事情就是算力集中。如果你相信 AI 研究者、RSI、AGI，那么所有这些都存在，这些都是基础。

There was the whole spat recently where I think Gavin Baker was like, "Dario believes that there's only going to be one company in the world." Then Sholto and Dario came out and were like, "No, no, no. We didn't say that." But ultimately, if you believe in RSI, if you believe the labs are the most effective user of compute and can generate the most value from the compute, then the only thing that's going to happen is centralization of compute. If you believe in AI researchers, RSI, AGI, then all of this exists, all of this is the base.

即使没有 RSI，这也是真的。前沿的有效人口目前以每年 10 倍的速度增长，对于给定的能力水平。所以如果你达到一个非常能干的远程工作者、或非常能干的软件工程师、或非常能干的研究者的能力水平，这些人口以当前能力增长速度每年增长 10 倍。

This is even true if there's no RSI. The effective population of the frontier is currently increasing 10x year over year for a given level of capabilities. So if you get to the level of capabilities of a very competent remote worker or a very competent software engineer or a very competent researcher, the population of those is increasing 10x year over year at the current rate of capabilities growth.

我明白了，而且没有 RSI。

I see, and without RSI.

然后一旦有了 RSI，就更疯狂了。那时可能每年增长 100 倍或 1000 倍。

Then once you have RSI, it's even crazier. Then it's maybe growing 100x a year or 1,000x a year.

或者他们的智能在增加，但人口没有增加。或者是两者的混合，对吧？

Or their intelligence is increasing but the population isn't increasing. Or some mixture of the two, right?

Dwarkesh，你看到什么样的世界，一切不是集中的？因为在我看来，每一种力量都在尖叫着走向集中。这太可怕了。

What world do you see, Dwarkesh, where everything is not centralized? Because it seems to me that every force is screeching towards centralization. And that's scary as hell.

我希望它不要完全集中。但也许这就是一个热爱优雅的机器的全部意义，对吧？它是一切，让我们的生活变得美好。未来太难想了。但我同意你的看法。

I would love for it not to be centralized completely. But maybe that's the whole point of a machine that loves grace, right? It is everything and it makes our lives great. It's so hard to think about the future. But I agree with you.

我认为根本问题是 AI 训练具有巨大的规模经济，因为你花在训练 AI 特定技能或特定知识上的任何努力都会在数十亿次会话或数十亿用户中摊销。所以这是一个效应。另一个效应是，如果你在 AI 竞赛中稍微领先，而算力短缺，你可以收取更高的加价，因为你能更好地节约这种稀缺资源。所以有两个效应，让领先者获得越来越多。可能还有更多。如果模型从部署中学习，而一个模型比另一个部署得更广泛，它获得更多的真实世界数据。

I think the fundamental problem is that AI training has huge economies of scale, because any effort you spend on training an AI for a specific skill or a specific set of knowledge gets amortized across billions of sessions or billions of users. So that's one effect. The other effect is that if you're slightly ahead in the AI race and compute is in shortage, you can charge a much higher markup because you can better economize this scarce resource. So there are two effects which give more and more to the person who's ahead in the AI race. There may be more. If models are learning from deployment, and one model is deployed much more widely than another one, it's getting much more real-world data.

你的观点我接受：无论是用户部署和持续学习，还是训练和规模经济，还是最佳 AI 模型帮助你制造下一个最佳 AI 模型的渐进式进步，RSI，所有这些都指向集中化。我认为，老实说，我们应该花时间思考的重大智力项目之一——至少我会花时间思考——是：在 AGI 之后，一个去中心化、广泛赋能的未来，同时认真对待这些规模经济，这样的愿景是什么？

Your point is taken that whether it's user deployment and continual learning, whether it's training and having these economies of scale, whether it's the incremental progress where the best AI model helps you to make the next best AI model, RSI, all of these things point to centralization. I think one of the big intellectual projects, honestly, that we should spend some time thinking about — or at least I'll spend some time thinking about — is: what is a vision of a decentralized, broadly empowered future after AGI that takes these economies of scale seriously?

另一种愿景是政府控制它，也许你认为你可以更信任政府，因为它不是私营公司。我不信任政府，也不信任 Dario，也不信任 Sam。这是个问题，对吧？显然，对未来判断错误很容易。你无法预见到一个关键效应或改变一切的东西。但事前来看，很难看到我们如何避免不得不选择一个集中化来源的情景。

The alternative vision is that the government controls it, and maybe you think that you can trust the government more because it's not a private corporation. I don't trust the government, and I don't trust Dario, and I don't trust Sam. That's a problem, right? Obviously it's very easy to be wrong about the future. You don't anticipate a key effect or something that changes everything. But ex ante, it's very hard to see how we avoid a scenario where we have to choose one source of centralization.

这就是资本主义成功的原因，对吧？它是去中心化的决策和去中心化的权力。这也是为什么超级集中的资本主义经济体实际上比超级去中心化的资本主义经济体增长得更慢，在某种程度上。你必须要有法治等等。但 AI 把这一切都颠覆了。

It's why capitalism worked, right? It's decentralized decision-making and decentralized power. And it's why super-centralized capitalistic economies actually grew slower than super-decentralized capitalist economies, to some extent. You have to have rule of law and all this. But then AI flips all this on its head.

最终你会说，“实际上，私有制可能不是最有效的经济，因此它比集中的 AI 经济增长得更慢。”嗯，它仍然是私有制，但真正参与这一经济份额的公司有多少？现在大概只占经济的 2% 吧？1 万亿美元除以 30。Nvidia 占了很大份额，还有 Anthropic、OpenAI 和这些超大规模企业。显然还有其他公司参与，但 AI 的很大一部分只发生在极少数公司。所以它可以是私有财产，但涉及的公司很少。我的意思是，这就是市场结构正在做的事情。

And ultimately you're like, "Actually, private ownership is probably not the most efficient economy, and therefore it grows slower than an AI economy, which is centralized." Well, it's still private ownership, but how many firms are really involved in this share of the economy? It's, what, maybe 2% of the economy right now? $1 trillion divided by 30. Nvidia is a huge share of it, and Anthropic and OpenAI and these hyperscalers. Obviously there are other firms involved, but a large share of the AI stuff is just happening from very few companies. So it could be private property, but very few companies are involved. I mean, this is what the structure of the market is doing.

那么什么能阻止它？我不知道。除非 AI 进展放缓，除非政府大力监管，否则这就是会发生的一切。在这种情况下，我们正走向一个世界：要么资源超级集中，我们祈祷那一家公司把一切都做对；要么政府放慢一切，人们放慢一切，希望进展放缓，权力更加平衡。

So what can prevent it? I don't know. Unless AI progress slows down, unless governments regulate the fuck out of it, this is all that happens. In which case, we're headed for a world where either we have super concentration of resources and we pray that that one company gets everything right, or we have governments slow everything down and people slow everything down, and you have a slowdown of progress somehow hopefully, and there is more of a balance of power.

即使我们走向 AGI、ASI、RSI，沿途的一切仍会导致有人捕获更多资源。所以很难找到一个框架，让 AI 不导致超级集中。

Even as we go towards AGI, ASI, RSI, everything along the way will still lead to someone capturing more resources. So it's kind of hard to find a framework in which AI doesn't lead to super concentration.

现在，这里一个积极的事情是，今天 Anthropic 并没有捕获大部分价值。我们可以随意谈论他们如何从每兆瓦 2000 万美元涨到每兆瓦 1 亿美元，但他们仍然为购买的许多算力支付 1300 万美元。但归根结底，他们涨到每兆瓦 1 亿美元的原因是因为 Jane Street 每兆瓦捕获 3 亿美元或 5 亿美元。或者 Dwarkesh，通过研究他的播客和学习信贷，每兆瓦捕获多少美元？

Now, the one positive thing here is that today Anthropic does not capture most of the value. We can talk all we want about how they went from $20 million per megawatt to $100 million per megawatt, but they're still paying $13 million for a lot of the compute they're buying. But at the end of the day, the reason they've gone to $100 million per megawatt is because Jane Street is capturing $300 million per megawatt or $500 million per megawatt. Or Dwarkesh, from researching his podcast and learning about credit, is capturing how many dollars per megawatt?

现在你能用多少？难说。但我认为这是唯一的可取之处，即经济的其他部分可能从 Anthropic 中获利更多——

Now how much can you use? Tough. But I think that's the one saving grace, that the rest of the economy maybe profits so much more from Anthropic—

不，但你之前阐述的整个逻辑——他们将推理重新分配给 AI 研发——那个逻辑是 AI 实验室内部的劳动回报远高于外部。

No, but the whole logic you were laying out earlier — them reallocating inference to AI R&D — the whole logic of that is that the returns to labor inside AI labs are much higher than the returns outside.

是的。这是我的应对。我同意。在世界的所有情景中……有 8 万个世界，其中只有一个世界里，Anthropic 不拥有整个世界。再说一次，权力集中是因为我不想把 token 发送到外部。它们在内部更有价值。所以这是同一回事。我为什么要让 Jane Street 从这些堕落的期权交易者身上赚这么多钱？

Yes. This is my cope. I agree. In all scenarios of the world… There's 80,000 worlds and in only one of them, Anthropic doesn't own the whole world. Again, power concentrates because I don't want to send the tokens outside. They're more valuable inside. So it's the same thing. Why would I let Jane Street make all this money off of these degenerate options traders?

嘿，他们是赞助商，拜托。天哪。不，我觉得这很好。让世界成为一个高效市场，这对世界是有价值的。

Hey, they're a sponsor, come on. Jesus Christ. No, I think it's great. It's a good value for the world to make it an efficient market.

Jane Street 通过正确把握世界观赚这么多钱，从堕落的期权交易者身上赚钱，不管是什么，Anthropic 为什么要将算力分配给它？如果 Jane Street 每兆瓦的最终货币化是 2 亿美元，所以他们愿意付给 Anthropic 1 亿美元……那么，如果 Anthropic 可以通过内部使用这些算力，每兆瓦产生数亿美元呢？这就是正在发生的事情。

Jane Street making all this money off of getting the worldview correctly, making money off of degenerate options traders, whatever it is, why would Anthropic allocate compute to that? If the end monetization that Jane Street has per megawatt is $200 million, so they're willing to pay Anthropic $100 million… Well, what if Anthropic can just generate hundreds of millions of dollars per megawatt by using that compute internally? That's what's happening.

在这个沉重的基调上，我想我们会在 RSI 正式启动时再见面。

On that somber note, I guess we'll meet again when the RSI is officially kicked off.

你不会再让我上你的播客，大概两个月后？好吧，酷。谢谢，老兄。

You're not going to have me on your podcast again for like two months? Alright, cool. Thanks, dude.