与Ed Elson探讨AI需求泡沫

太疯狂了。

Insane.

确实疯狂。

It is insane.

所以这个我是几天前发现的。

So I found this couple days ago.

我一直在研究这个，我以为其他地方不会那么糟，结果我看了瑞银的斯蒂芬·朱伊斯。

I've been working on it, and I thought, well, it can't be that bad elsewhere, except I looked at UBS's Stephen Juice.

他是在三月份说的，巴克莱也是三月份。

He said this back in, so Barclays is back in in March.

斯蒂芬是六月份说的，他说40，抱歉。

Stevens was in June, and he said that 40 sorry.

谷歌云所有收入的27%，不是AI收入，是2026年谷歌云所有收入的27%，预计来自OpenAI和Anthropic，2027年则是44%。

It's 27% of all Google Cloud revenues, so not AI revenues, all Google Cloud revenues for 2026 are estimated to be OpenAI and Anthropic, and then 44% in 2027.

这简直太他妈的疯狂了。

That is completely fucking insane.

这对我来说简直是个巨大的丑闻。

Like, that is just this is this is to me such a huge scandal.

而且，我还可以提一下，富国银行的一份报告也说了微软的情况类似。

And, I mean, I can also mention that, well, the there's also a Wells Fargo report that says that Microsoft's the same deal.

他们AI收入的73%来自Anthropic和OpenAI。

73% is Anthropic and OpenAI of their AI revenues.

那边的Michael Turrin。

Michael Turrin over there.

我感觉自己快疯了，因为当你看到这些数字时，你会想，这些可都是主要投资分析师说的。

And it's like, I feel like I'm going a little insane because when you read these numbers, you're like, oh, these are from major investment analysts.

他们怎么不喊出来，嘿，Fargo先生。

How are they not calling like, hey, mister Fargo.

Fargo先生，我看了这些数字，你真的应该好好看看，因为我们可能有大麻烦了。

Mister Fargo, I read these numbers, and you should really check these out because, we might be in trouble.

而且我觉得，节目的朋友Kagawa真的帮我把这个弄清楚了。

And I think and friend of the showcase, Kagawa, really helped me nail this down.

我觉得这些人只能看到未来两个季度。

I think it's these people are only capable of thinking two quarters in the future.

他们能想到2028年。

Like, they can they can think of 2028.

他们能想到2027年，但仅限于最近两个季度的情况。

They can think of 2027, but they can only do so in terms of the last two quarters.

数字一直在涨。

Number has gone up.

OpenAI和Anthropic还没破产，所以数字还会继续涨。

OpenAI and Anthropic are still solvent, so Number will keep going up.

收入集中不是问题，因为数字一直在涨。

It isn't a problem that the revenue is centralized because Number keep going up.

但在我看来，这可能是科技史上最大的企业丑闻之一。

But to me, this is potentially one of the largest corporate scandals in tech history.

也许我有点夸张，但我们现在说的是超过9900亿美元的资本支出，其中大部分都投给了这两家公司，而它们根本付不起这笔钱，除非那三家公司中的一家给它们钱。

And maybe I'm being dramatic, but this is we're talking over $990,000,000,000 of CapEx now, and most of it is for these two companies that can quite literally not afford to pay unless one of those three companies gives them the money.

对。

Yes.

这是个非常非常严重的问题。

This is a huge, huge problem.

你之前说分析师和媒体没有尽到职责，没有提出一些根本性的问题，比如他们打算如何产生回报。

And you were saying that the analysts weren't and the media weren't doing their job in terms of asking meta questions about how they're going to generate a return.

我当时有点反驳你。

I sort of pushed back on you.

但这次，分析师和媒体都沉默得可怕。

On this one, the silence is deafening from the analysts and from the media.

对于世界上最大、最具系统重要性的公司，那些美国每个人的投资组合都完全依赖、整个股市都依赖的公司，有一个非常非常明显的担忧，那就是你的业务有多少依赖于两家公司：OpenAI 和 Anthropic。

There is a very, very obvious concern when it comes to the businesses of the largest and most systemic companies in the world, the companies on which the portfolios of every American are entirely predicated on, on which the entire stock market relies on, and that is how much of your business depends on two companies, OpenAI and Anthropic.

而你是少数几个在深入挖掘这个答案的人之一，你研究了富国银行、巴克莱、瑞银的一些研究报告，我们应该感谢那些分析师，他们辛苦计算了这些数字，努力弄清楚了这个问题。

And you are one of the few people who has been doing significant work into uncovering that answer, and you have looked into some of the research that has come out of Wells Fargo, Barclays, UBS, and we should give those analysts credit for crunching those numbers and doing the hard work of figuring it out.

但这些数字简直离谱。

But the numbers are insane.

OpenAI和Anthropic占了微软AI收入的74%，这非常非常不可持续。

OpenAI and Anthropic, 74 of Microsoft's AI revenue, that is very, very unsustainable.

我们谈过市场的多元化，以及拥有多元化投资组合的重要性。

We talked about diversification in markets, the importance of having a diversified portfolio.

商业也是一样的道理。

The same thing goes for business.

你需要多元化的收入来源。

You wanna have a diversified revenue streams.

那才表明强劲、健康、可持续的增长。

That indicates strong, healthy, sustainable growth.

这些大型科技公司在AI方面的收入来源，没有一家能接近多元化。

None of these big tech companies have anything close to diversified revenue streams when it comes to AI.

所以你必须问一个很明显的问题：如果OpenAI和Anthropic这两家公司中有一家倒闭了，会发生什么？

And so you have to ask the very obvious question, what would happen if one of those two companies, OpenAI and Anthropic, went under?

你还得问：OpenAI和Anthropic这两家公司倒闭的可能性有多大？

And you also have to ask the question, what is the likelihood that one of those two companies, OpenAI and Anthropic, could go under?

他们无法兑现那些支撑这些大型科技公司增长、特别是其收入增长的付款的可能性有多大？

What is the likelihood that they would not be able to make good on those payments that are propping up the growth of these big tech companies specifically when it comes to their revenue?

答案是，实际上这种可能性相当高。

And the answer is it's actually pretty fucking high.

对。

Yeah.

你看看他们的财务状况，就像你之前详细谈过的那样，确实很糟糕，我们可以深入讨论，但我想把话题交回给你。

You look at the financials, and they are ugly as you have talked about at length, and we can get into, but I wanna pass it back to you.

不。

No.

其实有个有趣的故事。

The thing is fun story.

几天前，亚马逊原本打算在OpenAI上市或达到AGI时再给他们350亿美元。

A couple days ago, Amazon, they were meant to give OpenAI another $35,000,000,000 if they went public or reached AGI.

当然，这两件事都没发生，但几天前他们还是给了那350亿美元。

Well, none of those none of those things neither of those things happened, of course, but, they still gave them the $35,000,000,000 a couple days ago.

这太疯狂了。

And that and that's crazy.

这真是个疯狂的数字，尤其是因为——我们来看看巴克莱的罗斯·桑德勒给出的这些数字。

That's a really crazy number, especially because that is well, let's just go and have a look at these numbers from Ross Sandler at Barclays.

350亿美元大约是——让我想想。

That is so $35,000,000,000 is roughly oh, let's see.

对。

Yeah.

这差不多是亚马逊今年从AI业务中赚到的钱。

It's a a little bit it's around how much AI revenue Amazon has made this year.

所以是14.1，每9，每8.5。

So 14.1, per nine, per 8.5.

实际上，是的。

Actually, yeah.

对。

Yeah.

这比亚马逊今年在AI上的预计总收入还多340亿美元，抱歉，是多340亿美元，这还包括了OpenAI和Anthropic的支出。

It's a $3,400,000,000 less sorry, more than Amazon is estimated to make on AI this year in totality, including OpenAI and Anthropic spend.

所以这挺好的。

So that's good.

他们就是在把钱送出去，直接送给OpenAI和Anthropic，然后他们再把钱送回来。

Just handing themselves money, just handing it to OpenAI and Anthropic so they could hand it back.

没错。

Exactly.

他们已经给了Anthropic50亿美元。

They already gave $5,000,000,000 to Anthropic.

他们今年本来还打算再给200亿。

They already were on course to give them another 20,000,000,000 sometimes this year.

谷歌今年已经给了Anthropic100亿，还欠着300亿。

Google already give 10 gave $10,000,000,000 to Anthropic this year, has another 30,000,000,000 they owe.

这根本不是真实的经济。

This is not a real this is not a real economy.

这不是增长。

This is not growth.

这里没有人真正在做生意。

No one's actually doing business here.

只是人们在互相给钱，然后说，对。

It's just people handing money to each other and being like, yep.

看我们多赚钱。

Look how profitable we are.

尽管这似乎也没有让利润率暴涨。

Even though that does not even doesn't seem like it's exploding margins either.

似乎根本没什么效果。

It doesn't seem like it's having much of an effect at all.

但真正令人担忧的是，大家对这些公司感到兴奋，是因为它们的云业务增长。

But what's really concerning is everyone's excited about these companies because of their cloud growth.

嗯。

Mhmm.

而且现在越来越像是，大部分的收入增长都来自这笔支出。

And it's really beginning to look like most of that revenue growth is coming from this spend.

是的。

Yes.

这太可怕了。

Which is terrifying.

几乎所有的增长都来自那两家公司。

It's almost all of it is coming from those two companies.

之前你说过，他们没有披露收入来自OpenAI和Anthropic，这算是一种丑闻。

Earlier, you said that that them not reporting that their revenue is coming from OpenAI and Anthropic alone, that that is something of a scandal.

我觉得那简直是不负责任。

I think that is just irresponsible.

我倒不觉得那算丑闻。

I think of it less of a scandal.

我觉得真正算丑闻的是那种循环融资的部分。

The part that I do think is a scandal is the circular financing aspect of it.

关键在于，这两家公司也正是大型科技公司投入数十亿美元以维持其运营的公司，而这也是它们能够存活下来的唯一原因。

It's the fact that those two companies are also the two companies in which big tech invested billions and billions of dollars in order to keep them afloat, And that is the only reason why those companies are surviving.

是的。

Yes.

这涉及硅谷和风险投资者，但其中很大一部分投资是来自大型科技公司的企业投资。

It's Silicon Valley and venture investors, but a lot of that investment is corporate investment coming from big tech.

这些投资来自微软。

It's coming from Microsoft.

来自谷歌。

It's coming from Google.

来自亚马逊。

It's coming from Amazon.

所以，最大的客户，也就是推动这些公司收入增长的主要客户，同时也是这些公司投资过的公司，这显然存在非常明显的利益冲突，到了这个地步，就开始真正进入丑闻的范畴了。

And so the idea that the largest customers, the customers that are driving the majority of these companies' revenue growth are also the companies in which those companies had invested in, that is a very, very obvious conflict of interest at which point it starts to get into actual scandal territory.

我认为把这种循环融资称为流行病，或者称之为丑闻，都有些过火了。

I don't think that calling the circular financing epidemic, I think calling that a scandal is overkill.

我觉得这也不算太夸张。

I don't think that's a stretch too far.

我认为这确实是个丑闻。

I think it really is a scandal.

而且我觉得黄仁勋在这件事上的态度真的很丢人，当被问到这种循环交易的问题时，他说：我觉得这根本不是循环。

And I think it's actually shameful the way Jensen Huang has talked about it where he was asked this question about the circular deal making, and he said, I don't think it's circular at all.

他就是直接否认了。

He simply denied it.

他没有解释为什么不是循环，也没有给出任何真正的理由，只是说：不，那是错的。

He didn't come up with an explanation as to why it wasn't circular, he didn't come up with any real excuses, he just said, No, that's wrong.

胡说八道。

BS.

我不喜欢你这么说。

I don't like what you're saying.

然后他就直接转移话题了，这真的是个问题。

And then he just moved on, and that is a real problem.

这时候你就会发现，这些高管要么是盲目无视正在发生的事情，要么就是在撒谎。

That's when you start to see that actually these executives either they're just blindly ignoring what is happening or they are actually lying.

我觉得，当黄仁勋说他不认为他们做的事情是循环交易时，他很可能是在撒谎。

I I think that it's reasonable to assume that Jensen Huang told a lie when he said that he didn't think that what they were doing was circular.

回到之前的问题，OpenAI 会不会倒闭？

Just to, like, go back to the question of would OpenAI go under?

因为那，是的。

Because that's Yeah.

这才是真正的，事情就是这样发展的。

That's the real that's how this goes down.

对。

Yes.

而且这也是现在人们热议的一个点。

And that's that is also a point that people are now readily discussing.

这不再是那个叫 Ed 的疯子提出的边缘经济理论了。

It's no longer fringe economic theory from a crazy person called Ed.

现在这已经成了大家想讨论并且担心的问题了。

It's now something that people want to discuss and are worried about.

是的。

Yes.

我觉得

I think

但就像，嗯，对。

But it's like yeah.

但他们讨论的方式就像，怎么说呢，就像一家社区餐馆倒闭了一样。

But they they discuss it in this way that is like, I don't know, like, if a neighborhood restaurant shut down.

就像，哦，你知道，这种事时有发生。

Like, it's like, oh, you know, do you have you know, these things happen.

生意难做嘛。

Business is tough.

我真希望我是在开玩笑，但我试过跟记者聊这个，试着说，嘿。

I'm I I wish I was kidding, but I've tried to talk to reporters about this and tried to be like, hey.

你看。

Look.

这家公司真的是在支撑着整个行业。

This company is really holding up the earth.

对。

Yeah.

就像，它们对多家公司的未来收入影响很大。

Like, they they are quite consequential to the future revenues of multiple companies.

然后他们就说，对。

And they're like, yeah.

但是，你知道，还有其他需求。

But, you know, there's other demand.

是的。

Yeah.

这就是问题所在。

And that's the thing.

等我这篇正在写的文章发布的时候，那个大事情就会出来了，就像，是的。

When the big thing that this that the piece I'm working on will be out by the time this goes out says is, like, yeah.

谷歌、亚马逊和微软的AI收入中，大约73%来自OpenAI和Anthropic，这带来的另一个问题是，而且需要说明的是，这个比例是随着年份估算而变化的。

The other problem with 73% or so of AI revenues for Google, Amazon, and Microsoft being OpenAI and Anthropic is and just be clear, this progresses over the years of estimates.

他们不是说，好吧。

It's not like they say, okay.

需求会趋于平稳。

Demand will even out.

那暗示着AI没有需求。

That suggests that there isn't demand for AI.

总的来说，就是没有需求，至少没有承诺的那么大，也绝对不足以填满所有这些数据中心。

Just writ large that there is not demand, at least not at the scale that's been promised and definitely not at the scale that will satiate all of these data centers.

你看这些数字，他们就知道这是不可能的。

Like, they get it's not possible when you look at these numbers.

这也不是我随便说说，哦，我会写些文章来解决问题。

And this is not me just being like, oh, I'll work some stuff out with some articles.

这些都是真正的金融分析师。

These are actual financial analysts.

对。

Yeah.

对。

Yeah.

这些确实是。

These are yeah.

这真的很吓人，因为不仅仅是有多糟糕。

And it's it's genuinely terrifying because it's it's not just how bad it will be.

而且，是的，我们一会儿会回到OpenAI倒闭的话题。

And, yes, we'll get back to OpenAI dying in a second.

问题是，每个人似乎都只是看着它，然后说，哎呀，没事的。

It's the fact that everyone's kind of just looking at it and be like, ah, it's alright.

所以我觉得，是的。

So I think Yeah.

我觉得你真正的想法其实是在接受彭博社采访时透露出来的。

I think it was really the actual philosophy was was revealed in your Bloomberg interview.

我不是对那位采访者不敬，我觉得那其实是个很棒的采访，他们做得非常好。

I don't mean this disrespectfully to the interviewer, I thought it was actually a great interview and I thought they did a really good job.

但我忘了她叫什么名字，那位采访者问了你一个很重要的问题，她说，我只是想确保你明白，我只是想确保我们达成共识，经营这些公司的人都是非常聪明的人。

But I forget her name, the interviewer said something quite important to you where she said, I just want to make sure you understand, I just want to make sure we're all on the same page here, the people who run these companies are very smart people.

而那是那个问题的开头，接下来她大概会说，你真的确定这一切吗？

And that was the beginning of the question which would go on to say something like, Are you really sure about all of this?

我觉得这就是你在和媒体打交道、去《纽约时报》时看到的那种犹豫。

And I think that is the hesitancy that you're seeing when you talk to media outlets, when you go to the New York Times.

我觉得这也是我们所有人的犹豫，那就是说这家公司可能不会成功，这绝不是疯狂的说法，只要看看他们的财务状况和数字就知道了。

I think that's the hesitancy among all of us, which is that to say that this company might not actually work out, which is not an insane thing to say by any stretch of the imagination when you look at their financials and when you look at their numbers.

但如果说它可能不会成功，就等于说经营这家公司的人和投资这家公司的人不是世界上最聪明的人。

But to say that it might not work out is to say that the people who are running that company and the people who are investing in that company aren't the smartest people in the world.

而在今天，说这样的话是需要勇气的。

And today that is a bold thing to say.

说萨蒂亚·纳德拉可能不知道自己在做什么，这话很大胆。

It is a bold thing to say that Satya Nadella might not know what he's doing.

说山姆·奥特曼可能不知道自己在做什么，这话很大胆。

It's a bold thing to say that Sam Altman might not know what he's doing.

说马克·扎克伯格可能不知道自己在做什么，这话很大胆。但我觉得挺疯狂的，因为就在几年前，他花了超过800亿美元去搞元宇宙，结果彻底失败，一出来就死了。

It's a bold thing to say that Mark Zuckerberg might not know what he's doing, which is quite crazy to me considering that just a few years ago, he spent more than $80,000,000,000 trying to build the metaverse, which was a total failure, total dead on arrival.

他把公司改名为Meta，因为他觉得这就是未来。

He renamed his company to Meta because he thought that this was the future.

他在这方面错得太离谱了。

He was so, so wrong about that.

所以，认为这些人不会犯错，或者他们不会乱花钱，或者他们有时可能并不是优秀的资本配置者——

So the idea that these people can't be wrong or that they can't spend money incorrectly or that sometimes they might actually not be great capital allocators.

这完全是胡说八道。

That is total lunacy.

这完全是异想天开。

It is a total fantasy.

但不管什么原因，人们就是不愿意相信这一点。

But for whatever reason, people don't want to believe that.

当你开始提出这种说法时，人们会感到不舒服。

People get uncomfortable when you start to suggest that.

而我觉得，这正是为什么人们听你说那些话时会不舒服，因为这暗示着也许这些人并不清楚自己在做什么。

And that, I think, is why people get uncomfortable when you say the things that you say because it suggests maybe these guys don't know what they're doing.

嗯，确实是这样。

Well, it's that.

而且，我们现在就直说吧。

And also, let's just let's just say this right now.

我眼前就摆着巴克莱的财报数据。

I'm looking at the Barclays numbers in front of me.

OpenAI明天就倒闭。

OpenAI dies tomorrow.

也就是说，我们看看，这相当于亚马逊明年AI收入的28%。

That is, let's look at it, 28% of Amazon's AI revenue next year.

没了。

Gone.

再见。

Goodbye.

就这么蒸发了。

Just just evaporated.

而且确实，未来两年也是27%。

And indeed, 27% for the next two years as well.

如果Anthropic和OpenAI中有一个倒闭了，那大概就是20%，让我看看。

If Anthropic and OpenAI if one of them dies, what that's like 20% let's see.

OpenAI，嗯，看起来大约是谷歌2027年AI收入的125亿美元。

OpenAI is a well, it looks like about $12,500,000,000 of Google's 2027 AI revenue.

就这么没了，蒸发了。

Just like gone, evaporated.

否则，他们其他的AI收入就相当平稳。

And otherwise, their other AI revenues are pretty flat.

看起来Vertex Vertex有所增长，但同样，那是通过销售Anthropic和OpenAI的模型实现的。

It looks like Vertex Vertex kind of increases, but, again, that's through selling Anthropic and OpenAI's models.

我觉得人们没有意识到这篇文章里还有一些很棒的图表。

It's I don't think people realize that there's there's a great charts coming in this article as well.

Azure的收入中有多少是OpenAI的支出？

How much of Azure's revenue as well is just OpenAI spend?

让我看看。

And it's let's see.

我们来看看这个Azure收入的图表。

Let's take let's take a look at this chart of Azure revenue as well.

62%。

62%.

让我看看。

Let's see.

对。

Yeah.

在2026财年大约是72%，68%到72%，到2027年可能高达74%。

It's about 72% 68 to 72% in fiscal year twenty twenty six and as much as 74% in 2027.

如果这些公司倒闭，他们大部分的AI收入就没了，直接蒸发。

If these companies die, they lose most of their AI revenue, Evaporated.

这还不包括通过Azure AI出租模型带来的收入。

And that's not including the amount of money that comes through Azure AI renting out their models.

对。

Yeah.

就像人们只是说，是的。

Like and people are just like, yeah.

你知道，如果发生这种情况，会有更多的需求。

You know, if that happens, there'll be more demand.

我认识一个人，他非常善意。

I have this guy who's a very well meaning guy.

他在推特上。

He's in on Twitter.

他是一位半导体分析师。

He's he's a semiconductor ball analyst.

他总是回应说，你看。

He always responds and say, look.

所有有用的算力都会被买走。

All useful compute will be bought.

所有有用的算力都会被买走。

All useful compute will be bought.

它总是会被买走的。

It it will always be bought.

我就说，对。

And I'm like, yeah.

但如果OpenAI倒闭了，它是AI算力最大的消费者，那会怎么样？

But if OpenAI dies, it's the largest consumer of AI compute, what happens to that?

嗯，它会被买走的。

Well, it'll get bought.

对。

Yeah.

这对我来说简直难以置信。

And it's just insane to me.

这真的让我觉得匪夷所思，没有人……我想，退一步看，就像你说的，他们不愿意把逻辑推到终点，因为那意味着所有这些聪明人都很愚蠢。

It is genuinely bonkers to me that no one like, I guess I guess what it is taking a step back is looking at what you said as well, is it's they don't want to follow the argument to the logical end point because it's all of these smart people were stupid.

没错。

That's right.

而且，所有人都错了。

And, also, everyone was wrong.

错得很难回头。

Wrong in a way that is going to be very difficult to back out of.

没错。

That's right.

没错。

That's right.

对。

Yeah.

他们倒是想。

They will like to.

不过先说清楚。

Just just to be clear.

对。

Yeah.

这就是问题所在。

That that's that's the problem.

所以在这点上我完全同意你。

Now so I'm in total agreement with you on all this.

我想把问题推到逻辑的终点，我要说一些话，但不知道你会不会感兴趣。

I wanna take things to their logical conclusion further, and I'm gonna say some but I don't know if you're going to be that excited about what I'm going to say.

那是

That's

好的。

fine.

因为我在想，如果AI真的就这么死了，会发生什么？

Because when I think about this, what would actually happen if AI were to basically just die?

如果OpenAI真的倒闭了，Anthropic也真的倒闭了，我觉得我们有理由相信这会发生，我们可以深入聊聊他们的财务状况，重新梳理一下，因为这些情况非常重要，而且对整个生态系统有系统性影响。

If OpenAI actually died and if Anthropic actually died, and I think there's good reason to believe that, which we can get into, we can get into their financials and just rehash them because they are very important and they're systemic to the ecosystem.

但我有时候会想，如果AI这事儿没成，世界会是什么样子？

But I sometimes like to think, okay, what would a world look like where the AI thing just didn't work out?

我得说，现在很多这种焦虑其实已经反映在大型科技公司的股价里了。

And I have to say, a lot of that anxiety actually is priced in right now when it comes to big tech.

我想让你看看价格，对。

And I just want to point you to the price Yes.

市盈率，2023年初大型科技公司的市盈率。

To earnings The price to earnings multiples of Big Tech at the beginning of twenty twenty three.

我回去查过这个。

I went back and looked at this.

那是在AI热潮开始之前。

This was right before the AI trade took off.

如果你看市盈率，微软当时是26倍。

If you look at the price to earnings multiples, you look at Microsoft, it was trading at 26 times earnings.

亚马逊是80倍。

Amazon was trading at 80 times earnings.

Meta是31倍。

Meta was trading at 31 times earnings.

现在再看市盈率，顺便说一下，我是上周看的，就在他们公布财报之后，所以现在可能已经变了。

If you look at the PE multiples today and by the way, I looked at this as of last week, at the end of last week, right after they reported their earnings, so they could have changed by now.

肯定变了。

I'm sure they have.

微软的市盈率比之前低了。

Microsoft's multiple is lower than it was.

现在是25倍。

It's at 25.

虽然低得不多，但确实是低了。

Not that much lower, but it is lower.

亚马逊的市盈率就低多了，而且这是真实的市盈率，我调整过的，没有算上投资Anthropic和OpenAI的收益，现在是30倍，比之前低了。Meta的市盈率也低了很多，现在是20倍，而AI热潮前是31倍。

Amazon is Amazon's is significantly lower, and this is the real PE, the one that I adjusted, not factoring in the the investment gains on Anthropic and OpenAI stakes, is trading at 30 times earnings, that's lower, and Meta's is significantly lower, it's trading at 20 times earnings compared to 31 before the AI boom.

所以我觉得我们应该认识到，在市场的一些地方，有些领域，我认为华尔街确实有些人很焦虑。

So I think one thing that we should recognize is that in some places, in some areas of the market, I think there are people on Wall Street who are actually quite anxious.

我觉得华尔街有很多人，Ed，都在读你的通讯。

I think there are a lot of people on Wall Street, Ed, who are reading your newsletter.

我真的这么认为。

I I truly believe that.

而且我觉得那种看跌的观点，以及承认这一切可能崩溃的看法，我认为它并没有完全主导市场。

And I think that that perspectivethe bearish perspective and the acknowledgement of how this could all implodeI don't think that it is completely driving the market.

我觉得乐观情绪多于悲观情绪。

I think there's more optimism than there is pessimism.

但我确实认为它在一定程度上有所反映，也就是说，如果这一切真的崩溃了，我不认为这些股票不会崩盘。

But I do think that it is somewhat being reflected, and that is to say if this all were to blow up, I don't think these stocks wouldn't crash.

它们肯定会崩盘。

They would certainly crash.

但我确实认为，不会像2008年那样可怕，当时房地产泡沫几乎没人察觉。

But I do think that it wouldn't be as horrible as it might have been, say, in 2008 when the housing bubble was truly elusive to almost everyone.

这是我的猜测，我肯定你会不同意，但我确实认为我们意识到了某种焦虑的程度，这种焦虑开始反映在一些资产价格上。

Is my conjecture, and I'm sure you might disagree with me on it, but I do think we recognize the extent to which there is some anxiety that is beginning to be reflected in some asset prices.

经营企业，却不考虑播客？

Run a business and not thinking about podcasting?

再想想吧。

Think again.

现在听播客的美国人比听Spotify和Pandora广告支持的流媒体音乐的人还多。

More Americans listen to podcasts than ad supported streaming music from Spotify and Pandora.

作为排名第一的播客平台，iHeart的规模是紧随其后的两家总和的两倍。

And as the number one podcaster, iHeart's twice as large as the next two combined.

所以无论你的客户听什么，他们都会听到你的信息。

So whatever your customers listen to, they'll hear your message.

而且，只有iHeart能把你的信息传递给广播电台的听众。

Plus, only iHeart can extend your message to audiences across broadcast radio.

觉得播客能帮到你的生意吗？

Think podcasting can help your business?

想想iHeart吧。

Think iHeart.

流媒体、广播和播客。

Streaming, radio, and podcasting.

拨打844844开始吧。

Call 844844 to get started.

就是844844。

That's 844844

我们都经历过。

We've all been there.

我们都有过动摇一切的季节。

We've had seasons that have shaken everything.

问题是，当震动停止时，还剩下什么？

The question is, what remains when the shaking stops?

嗨。

Hi.

我是加里·瓦伦西亚诺。

I'm Gary Valenciano.

在我的播客《与加里·维一起坚定不移》中，我会和许多了不起的人物坐下来聊聊那些几乎击垮他们的时期，以及那些低谷如何将他们推向高峰。

And on my podcast, Unshaken with Gary Vee, I sit down with amazing personalities and we talk about the seasons that nearly broke them, the lows that carried them to their highs.

在iHeartRadio应用、Apple Podcasts或任何你收听播客的地方，收听《与加里·维一起坚定不移》。

Listen to Unshaken with Gary Vee on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.

嗨，亲爱的。

Hi, my love.

开始新的一天可能很难。

Starting your day can be tough.

你醒来，可能按掉闹钟，然后刷一会儿手机，查看邮件，这一切都发生在你真正睁开眼睛、深呼吸之前。

You wake up, maybe hit snooze, and then scroll on your phone for a while, checking your emails, all before even giving yourself the chance to really open your eyes and breathe.

我想帮你打破这个循环。

And I wanna help you break that cycle.

我叫瑞安·韦斯，过去十五年里，我一直是一名情商教练和精神导师。

My name is Ryan Weiss, and for the past fifteen years, I've been an emotional intelligence coach and spiritual guide.

我来这里是想和你分享我的新播客《和瑞安一起醒来》。

And I'm here to share with you my new podcast, Waking Up With Ryan.

《和瑞安一起醒来》是一个每日音频视频播客，旨在打破你的恐惧、自我和评判的惯性，帮助你在一天的喧嚣开始之前与自己建立连接。

Waking Up With Ryan is a daily audio video podcast to interrupt the inertia of your fear, your ego, and judgment, here to help you connect with yourself before the noise of the day takes over.

所以，让我们一起以平静的时刻开始新的一天，一个反思和冥想的时刻，一个只属于你的时刻。

So let's start our days together with a moment of calm, a moment of reflection and meditation, a moment that's just for you.

在iHeartRadio应用、Apple播客或任何你收听播客的地方收听《和瑞安一起醒来》。

Listen to Waking Up With Ryan on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.

在音乐会上，我们随着同一首曲子起舞。

At concerts, we dance to the same tune.

在比赛中，我们拥抱陌生人。

At games, we hug total strangers.

然而在这些时刻，我们也有不做的事情。

And yet in these moments, there's also something we don't do.

我们不会根据他们的外表、他们爱谁、他们来自哪里或他们信仰什么来行动。

We don't act based on how they look, who they love, where they're from, or what they believe.

我们行动只是因为那些让我们不同的东西永远不会改变我们作为人的本质。

We just act because the things that make us different will never change what makes us human.

请访问lovehasnolabels.com了解更多信息。

Go to lovehasnolabels.com to learn more.

由Love Has No Labels和eye council为您呈现。

Brought to you by Love Has No Labels and the eye council.

有一个反对意见。对。

So one pushback Yep.

我认为这是一个关键点，那就是尽管市盈率等指标反映出一些下降和焦虑，但价格仍然被计入指数，而指数则考虑了这些公司将继续增长的因素。

And I think it's a critical one, is that even though the price through earnings and all that reflects a little bit lower and showing some anxiety, it still is a price factored into the I which factors in the idea that these companies will keep growing.

是的。

Yes.

而且我认为这是一个关键问题，因为我不确定，至少就他们的云业务而言，如果去掉OpenAI和Anthropic，这些公司是否还在增长。

And I think that that is a critical problem because I'm not sure, at least in the case of their cloud businesses, whether these companies are growing anymore when you remove OpenAI and Anthropic.

对。

Right.

我不是说微软、谷歌、亚马逊和Meta会倒闭，虽然我觉得Meta可能在十年内会完蛋。

I don't think it like, I'm not just talking like, I don't think Microsoft, Google, Amazon, and Meta, though I think Meta could die in ten years.

我这么想已经有一阵子了。

I've thought that for a while.

我觉得他们只是在分崩离析，但我不认为他们会倒闭。

I think that they're just it's falling apart, but I don't think they're dying.

我觉得他们的股票会被打得很难看。

I think their stocks are gonna get their asses beat.

但问题不仅仅是AI没起作用。

But the problem is going to be not just the AI didn't work.

那将是第一个问题。

That's what's gonna be the first problem.

第二个问题是，当他们说，对。

The second one is when they go, right.

好吧，OpenAI不再付钱了，突然你的收入就下降了。

Well, OpenAI isn't paying you anymore, and suddenly your your revenues are down.

是的。

Yes.

而且你的增长也放缓了。

And your growth has slowed.

嘿。

Hey.

谷歌云已经停止增长了。

Google Cloud has stopped growing.

它几乎完全停止增长了。

It stopped growing pretty much at all.

实际上看起来有点持平。

It's in fact looks kinda flat.

这是怎么回事？

What's the deal with that?

这是因为Anthropic和OpenAI的算力支出撑起了增长。

And it's because Anthropic and OpenAI's computes compute spend was propping up the growth.

所以，虽然华尔街可能会说，哦，如果AI不行了，我觉得他们的意思是他们认为AI不会成为下一个大热门，但从中仍然会有大的营收业务，而事实并非如此。

So while Wall Street might be saying, oh, well, if AI doesn't work out, I think what they mean is that they think AI won't work out and be the next big thing, but there will still be big revenue generating businesses from this, which is not the case.

我去看了Turin的富国银行预测。

I went and looked at the Wells Fargo estimates from Turin.

对微软来说，有趣的是，在2027财年，也就是明年，他们估计Microsoft 365 Copilot的年收入将达到100亿美元，这太可怜了。

And what's interesting for Microsoft is the in fiscal year twenty twenty seven, so next year, they estimate the Microsoft three sixty five Copilot will make $10,000,000,000 a year in in annual revenue, which is pathetic.

是的。

Yeah.

那就像，什么？

That's like a a what?

大概只有智能云的三分之一？

Like a third of intelligent cloud?

就是说，这比他们预估的Azure AI收入要少一点，而Azure AI的收入主要是靠卖OpenAI的模型撑起来的。

It's let I mean, it's a little bit less than they estimate the Azure AI will make and which is predominantly made up of selling OpenAI's models.

我是说，他们在财报电话会议上吹嘘的一个点就是他们有3000万付费的Copilot用户。

I mean, they they said one of their bragging points on that earnings call was that they have 30,000,000 Copilot paying Copilot users.

但他们没怎么提的是，这只是更大的企业套餐的一部分，而那个套餐只占他们Microsoft 365总订阅用户的8%左右。

But what they don't really talk about is the fact that that's part of a larger enterprise bundle, which only makes up about 8% of their total Microsoft three sixty five subscribers.

所以基本上，他们告诉我们的，他们想炫耀，但最终我读到这个，我就想，哦，所以你只转化了你用户群的8%。

So basically, what they're telling us, they're they're trying to brag, but ultimately, I read that and I go, oh, so you only converted 8% of your user base.

那是你用户中愿意尝试这个你号称要改变世界的AI产品的比例。

That was the share of your users who were down to try this AI product that you say is going to change the world.

所以说到这一点，我觉得他们在消费者AI业务上展现出来的实力，远没有他们想让我们相信的那么强。

So to that point, I don't think that the strength that they are projecting projecting in in their their consumer AI business is anywhere near, what they'd like us to believe.

而且，富国银行的预估，预估，预估，肯定有人在评论区挑我这个毛病。

Well, on top of that, the Wells Fargo estimates estimates estimates, someone will get me on that one in the comments.

他们预估整个2026财年，Microsoft 365 Copilot只赚了38.5亿美元。

They only estimate that they made 3,850,000,000 for the entire fiscal year 2026 from $3.65 Copilot.

你知道，三千万付费席位这事儿。

You know, the thing with 30,000,000 paid seats.

他们把这折扣打到天边去了。

They're discounting this to Timbuktu.

就是说，他们把这折扣打到地底去了。

Like, this is this is being they're discounting it to hell.

而且，基本上微软卖的任何东西都有折扣，尤其是Microsoft 365。

And I mean, pretty much anything like like, pretty much anything that Microsoft sells has discounts, especially when it comes to Microsoft three sixty five.

但话说回来，那些分析师还是照单全收。

But again, the hogs lapped it up.

他们就喜欢这种糊弄人的东西。

They love their slob.

他们喜欢有个新的大数字可以解读。

They love being given a new big number so that they can interpret it.

不过，我真不知道这些分析师在干嘛，因为我不清楚。

When they again, not really sure what these analysts were doing because I don't know.

再说一次，我觉得我之前已经说过了。

Again, I think I've said this already.

比如，如果我在富国银行工作，手里有这些数据，我会直接去找我老板的老板。

Like, if I was working at Wells Fargo and I had these numbers, I would be calling my boss's boss.

我会说，嘿，各位。

I'd be like, hey, guys.

嗨。

Hi.

我们遇到问题了。

We have a problem.

因为就像我之前说的，不仅仅是微软、谷歌和亚马逊在OpenAI和Anthropic之外的AI收入惨淡。

Because like I said earlier, it's not just the Microsoft, Google, and Amazon have pathetic AI revenues outside of OpenAI and Anthropic.

问题是，如果他们收入惨淡，如果他们作为AI算力的最大持有者，拥有最强的定价权，表面上最好的团队，以及最丰富的建设AI数据中心和提供AI算力的经验，他们却几乎凑不出什么？

It's that if they have pathetic revenues, if they, as the largest holders of AI compute, as with the ones with the best pricing power, with the best teams ostensibly, with the most experience building AI data centers and providing AI compute, if they can barely scrape together what?

比如什么？

Like what?

比如说，我们来看看。

Like, let's see.

所以这占了他们收入的70%，大概73%、74%。

So it's 70 like 73, 74% of their revenues.

所以，他们那点可怜的AI收入，大部分都来自多元化收入。

So, like, they have this meager little little shrivel of their AI revenues is from diverse revenue.

如果是这样的话，那就说明对AI算力没有需求。

If that's the case, then there is not demand for AI compute.

不是说只有一点点。

It's not that there's only a little.

而是说根本没有。

It means there's none.

这意味着实际上只有几十亿，因为，再看看其他供应商，CoreWeave，他们最大的客户是谁？

It means there's actual like, there is a couple billion because, again, when you look at the other providers, CoreWeave, who are their biggest customers?

Anthropic、Meta和OpenAI。

Anthropic, Meta, and OpenAI.

Zyrin最大的客户是谁？

Who's Zyrin's biggest customer?

微软。

Microsoft.

Nebius。

Nebius.

微软。

Microsoft.

微软在用AI做什么？

What's Microsoft doing with AI?

没做太多。

Not very much.

实际上真没做多少。

Not actually very much at all.

他们只是给OpenAI提供算力，然后忽悠人们买Copilot。

They are providing compute to OpenAI and conning people into buying Copilot.

对。

Yeah.

就是说，这真的很惊人，因为这比他们还要大。

Like, it's just it's astonishing because this is bigger than them.

比他们还要大。

It's bigger than them.

这是计划中的190吉瓦的数据中心容量。

It's the 190 gigawatts of planned data center capacity.

所有这些数据中心都在各地建设。

It's all of these data centers being built everywhere.

我，我，我说话都结巴了，因为这带来的后果正是我一直担心的：AI需求并没有大规模存在，而且大多数这些数据中心不会得到支付。

I like, I I'm stuttering because it's like the ramifications of this are everything I've kind of feared, which is that AI demand does not exist at scale, and most of these data centers don't go paid for.

即使Anthropic和OpenAI能成功撑过2027年，情况也是如此。

And that's even if Anthropic and OpenAI, like, succeed in surviving past 2027.

嗯，我觉得你之前提到的增长很重要，那就是如果AI这玩意儿不行了，微软、亚马逊、谷歌和Meta并不是说就没有业务了。

Well, I think you said something important earlier about the growth, and that is if if this AI thing doesn't work out, it's not as if Microsoft and Amazon and Google and Meta won't have a business anymore.

他们在AI之前就有业务，而且那个业务规模很大，做得也很好。

They had a business before AI, and the business was pretty big and doing pretty well.

但这意味着这些公司实际上不会再增长了。

What it does mean is that these companies won't actually be growing anymore.

不再是增长型公司了。

Won't be growth companies.

我觉得这触及了AI的核心问题，那就是在2022年底ChatGPT发布时，大型科技公司突然意识到，天哪，创新正在我们之外发生。

And I think this gets to the heart of what AI has been all about, and that is that the big tech companies in late twenty twenty two when ChatGPT was released to the world, they suddenly realized, oh my gosh, innovation is happening without us.

有其他公司出现了，有其他技术，而我们没有参与其中。

There are other companies on the scene, there are other technologies and we don't have a part in it.

所以他们决定，好吧，我们该怎么办？

And so they decided, okay, what are we going to do?

我们要尽一切可能，把那个公司里任何一点技术苗头都拿过来，然后嫁接到我们自己的公司上。

We're going to do whatever we can to take whatever semblance of technology exists in that company and then bolt it onto our own companies.

我们基本上要把那种乐观情绪像打肉毒杆菌一样注入到我们自己的脸上，让市场相信我们确实又年轻了，我们在增长，我们很热门，我们很有创新力，你们应该来投资我们。

We're basically going to inject that optimism into our own faces as if it were Botox to convince the market that we are in fact young again, that we are growing, that we're hot, that we're innovative, and that you should come and put your money in us.

因为我们不仅仅是一家继续产生现金的公司。

Because we're not just a company that's going to continue to generate cash.

我们是一家划时代的公司。

We're a generational company.

我们是一家在未来十年、二十年、三十年里利润会以指数级速度大幅增长的公司。

We're a company that's going to significantly increase profits at an exponential rate over ten, twenty, thirty years.

这就是他们想向市场讲述的故事。

That is the story that they wanted to tell the markets.

他们非常迫切地想讲这个故事，以至于愿意真的去投资像OpenAI、Anthropic，还有SpaceX这样的公司——我们之前没怎么提到SpaceX，但它其实在其中扮演了角色。

And they wanted to tell it so desperately that they were willing to actually go and invest in those companies like OpenAI and Anthropic and also SpaceX, which we haven't mentioned that much, but it actually plays a role in all this.

对，就是这样。

That's what yeah.

然后让那些公司回过头来买我们的芯片，把钱还给我们。

Then have those companies come back and then and then buy our chips, give us back the money.

然后我们再用这个作为故事，告诉全世界我们其实又年轻了，我们在增长。

And then we're gonna use that and use that as a story to tell the world that actually we're young again, that we are growing.

并且通过这种方式说服所有人，为了跟我们的投资公司竞争，把钱花在AI计算上。

And convince everyone by proxy to spend money on AI compute to compete with our.

没错。

Exactly.

跟我们以及我们投资的公司竞争。

With us in with the companies that we invest in.

在这个过程中，我们的资产负债表会被摧毁，那时情况会变得非常糟糕。

We will obliterate our balance sheets in the process, and that's that's where things are gonna get really ugly.

那些真正开始亏损现金的公司，我们已经开始看到了。

The companies that really start to to lose cash, and we're starting to see it.

亚马逊，自由现金流为负。

Amazon, negative free cash flow.

谷歌，本季度自由现金流也为负。

Google, negative free cash flow as of this quarter.

但这正是他们在这里所做的。

But that's really what they're doing here.

他们想重回年轻。

They wanna be young again.

但如果你看看这些公司，它们已经很老了。

But if you look at these companies, these companies are old.

它们是成熟的公司。

These are mature companies.

它们的鼎盛时期已经过去，但它们不愿相信这一点。

Their heyday is behind them, but they don't want to believe that.

它们想永续存在。

They want to live forever.

它们就是科技界的布莱恩·约翰逊们。

They are the Brian Johnsons of the world of technology.

所以我认为，如果这事不成，而且我认为不成的可能性很大，因为OpenAI和Anthropic的财务状况，以及整个世界对它们的依赖程度，我认为会发生的是，这些大科技公司不会倒闭。

So I think what's gonna happen if this thing doesn't work out, and I think there's a high likelihood that it won't because of the financials of OpenAI and Anthropic and how dependent the whole world has become on them, I think what will happen is the companies won't go under, the big tech companies.

我认为会发生的是，华尔街会看着它们说，你们不是成长型公司。

I think what will happen is Wall Street will look at them and they'll say, you're not a growth company.

你是一家成熟的公司。

You're a mature company.

这意味着我们将对你的收益给予一个成熟的市盈率。

And that means that we're going to assign a mature multiple to your earnings.

我们不会再看到那种日子了，我觉得那会是在15到20倍市盈率的范围内。

We're not gonna have the days of I think that'll look like something in the range of 15 to 20 times earnings.

所以，顺便说一下，我对Meta并不太担心，因为Meta实际上就在这个范围内。

So that's why, by the way, Meta I'm not so worried about because Meta is actually in that range.

我觉得如果你看看这家公司，华尔街正在看着它说，这家公司的增长前景充其量是不太可能的。

I think if you look at the company, I think Wall Street is looking at it and saying, this is a company whose growth prospects are unlikely at best.

所以我对Meta并不太担心。

So I'm not so worried about Meta.

我更担心的是像谷歌这样的公司，它的市盈率在30倍以上。

The companies I'm more worried about is a company like Google, trading above 30 times earnings.

顺便说一下，从市盈率的角度来看，苹果已经远远超出了合理范围，这真的让我很惊讶。

By the way, a company that's gotten way out over its skis from a multiple perspective, which is really surprising to me, is Apple.

我觉得这是因为人们对AI交易感到紧张了，所以他们说，好吧。

And I think that's because people have actually gotten nervous about the AI trade, and so they said, okay.

亚马逊的资产负债表上还有现金。

Amazon still has cash on the balance sheet.

苹果的市盈率接近40倍。

Apple is trading at nearly 40 times earnings.

他们一直能骗市场相信他们是成长型公司，但十年没出过新产品，尝试推出的那些也完全失败了。

The way that they have been able to dupe the market into thinking that they are a growth company when they haven't come out with a new product in ten years and the ones that they've tried to come out with and tried to sell have completely flopped.

他们取消了苹果汽车，Vision Pro也是一场彻底的灾难。

They canceled the Apple car, the Vision Pro was a total disaster.

我的意思是，这显然不是一家成长型公司。

I mean, that's that's very clearly not a growth company.

但我觉得问题就在这里。

But I think that's the question.

苹果有趣的地方在于，他们靠对AI交易不眨眼就撑下来了，但AI交易因为内存成本上升反而坑了他们。

The funny thing is with Apple as well is that they've managed to keep going by just not blinking on the AI trade, except the AI trade fucked them over because of the increasing cost of RAM.

对。

Right.

因为大家都在买，我是说，这是另一回事。

Because of everyone buying into I mean, that's the other thing.

就像这个增长，这是另一件事，所以实际上，这正好接上你刚才说的。

It's like the increasing this is the other so actually, this is a good bridge from what you were just saying.

嗯。

Mhmm.

这个问题还有另一个方面，就是，你说这些公司会被赋予像那些老掉牙的、无聊的公司那样的增长倍数。

The other problem with this as well is, yeah, you're saying these companies will be assigned growth like the multiples associated with old old boring companies.

对吧？

Right?

无聊、老套、成熟。

Boring, old, mature.

但再说一次，我认为这没有考虑到，第一，它们没有增长，第二，它们现在拥有数千亿的资产。

But again, I don't think that factors in, a, the fact they're not growing, but b, the fact that they have hundreds of billions of assets now.

他们只是背负着债务，无尽的债务。

They're just burdened with debt, endless debt.

微软在债务管理上还算聪明。

Microsoft has been kind of smart in the amount of debt it's taken on.

它远没有那么沉重，但谁知道呢？

It's nowhere near as encumbered, but who knows at this point?

但谷歌进行了大规模的股权出售。

But Google's done a massive equity sale.

他们计划进行大规模的股权出售。

They plan to do a massive equity sale.

他们越是急于追求这种重返青春的美容式未来，就越危险。

The more on tilt they get in order to pursue this Botox future of being young again This is really dangerous

你得把这句话记下来。

You gotta write that out.

他们越是危险，对他们整个存在的威胁就越大。

The more dangerous the the the larger the threat to their entire existence becomes.

所以我同意你的看法。

So I will agree with you.

如果他们继续沿着这条路走下去，不只是用银行账户里现有的现金来支付肉毒杆菌，而是背负大量债务、大量杠杆和保证金去追求那个梦想，因为我太执着于追求它，我必须再次年轻。

If they continue down this road of, I'm not just gonna pay for the botox with the cash that I already have in my bank account, I'm going to take on significant amounts of debt, significant amounts of leverage and margin in order to pursue that dream because I'm so obsessed with pursuing it and I need to do it because I have to be young again.

如果他们像你所说的那样，把这条路走到逻辑的尽头，那么，是的，我们可以开始讨论这些公司是否会崩溃。

If they take that to its end, to its logical conclusion as you say, then, yeah, we can start talking about will these companies implode.

我认为甲骨文就是一个例子，他们已经这么做了。

And I think Oracle is an example of a company where they have done that.

他们确实这么做了，而且它的前景确实令人恐惧。

They have genuinely done that, and its prospects are genuinely terrifying.

我们开始在其他一些公司身上看到这种情况。

We're starting to see it with some of these other companies.

我们开始……我很震惊谷歌竟然愿意公布一个负自由现金流的季度。

We're starting to I was shocked that Google was willing to come out with a quarter of negative free cash flows.

我很震惊亚马逊竟然愿意这么做。

I was shocked that Amazon was willing to do that.

所以看起来他们在这方面开始失去理智了。

So it does appear that they're starting to lose their marbles on this front.

他们太执着于这个使命了。

They're so obsessed with the mission.

他们太执着于AI的未来，以至于实际上在采取不负责任的行动，但我认为我们还没到那一步。

They're so obsessed with the AI future that they are actually acting irresponsibly, but I don't think that we're quite there yet.

我认为迹象表明他们会继续这样做。

I think the signs show that they're gonna continue to do this.

我的意思是，如果他们继续用债务来充实资产负债表，谁知道未来会是什么样子呢？

Mean, if they continue to load up their balance sheets with debt, who knows what the future will actually look like?

但是

But

目前来看，他们还没有到完全崩溃和瓦解的地步，这是我对当前情况的看法。

this is So far, they're they're not quite in a in a place where they're they're fully going to collapse and implode would be my read of the situation at present.

顺便说一句，我完全同意。

I fully agree, by the way.

我其实不觉得三大巨头会倒闭。

I don't actually think the big three collapse.

我觉得Meta已经在走下坡路了。

I think Meta is already on the down spiral.

对。

Mhmm.

我觉得他们的产品正在崩溃，而且我从那家公司不断听到的消息表明，他们的文化已经烂透了。

I think they real I think their product is collapsing, and the things that I continually hear from that company suggest that, like, the culture is screwed.

但尽管如此，至少短期内，未来五年，我觉得他们还是会没事的。

But nevertheless, for the short term at least, next five years, I think that they will be fine.

当这一切结束时，他们不会好过的。

They're not gonna have a fun time of it when this ends.

再说，谈到逻辑终点和人们不愿讨论的那些事，甲骨文现在把全部收入都花在——是的。

Again, when it comes to, like, logical endpoints and the way that people don't wanna discuss them, Oracle is now spending all of its revenue Yeah.

甚至超过收入，都花在AI资本支出上，就为了OpenAI。

Just more and more than its revenue on on AI CapEx just for OpenAI.

而且，顺便说一句，纽约时报简直就是在抄袭我。

And, I mean, the New York Times just just fucking ripped me off, by the way.

比如，他们那篇关于堕落天使的文章里有一句话，跟我写的几乎一模一样。

Like, there is a sentence in there about fallen angels that is, like, bordering on identical to my piece.

真是服了纽约时报了。

Fucking New York Times, man.

如果你们有纽约时报的人在听，去看看那篇关于甲骨文的文章吧。

If you're listening from the Times, go read the Oracle piece.

去看看我在OpenAI泡沫那篇里写的关于堕落天使的部分。

Go read my fallen angel bit from the OpenAI bubble.

你们说说，是不是我疯了。

You tell me if I'm crazy.

不过，现在大家都在讨论，哇，甲骨文居然不赚钱。

But, nevertheless, people are having this conversation like, wow, Oracle, they're kind of unprofitable.

这可不太好。

That's not good.

嗯。

Mhmm.

哇。

Wow.

那可不太好。

That's not good.

但他们和其他人一样，没有说出他们不想说的话。

But then they stop short of saying the thing that they don't want to say, much like everyone else.

现在我想谈谈这一点，是的，如果甲骨文不行，AI交易就完了。

Now I wanna talk about this bit, which is, yeah, if Oracle doesn't work out the AI trade, it dies.

对。

Yes.

经通胀调整后，其收入十五年来一直持平。

Its revenue has been flat for fifteen years adjusted for inflation.

疯狂的是，我几周前就写过这个。

It has the crazy thing is I wrote about this a couple weeks ago.

甲骨文最疯狂的地方在于，他们其实已经在走下坡路了。

The crazy thing about Oracle is that they actually would be on the decline already.

但每当他们业绩放缓，他们就会收购更多公司。

But whenever they slow down, they buy more companies.

他们就是不停地收购公司。

They just acquire companies.

他们真的很奇怪。

They're like, they're really strange.

对。

Yeah.

但尽管如此，没人愿意谈论甲骨文会倒闭的事实。

But nevertheless, like, no one wants to talk about the fact that Oracle will die.

说实话，看到这些数字，是的，甲骨文会倒闭。

And honestly, seeing these numbers, yeah, Oracle will die.

这确实是一个真实的可能性。

It's it is a it is a genuinely it it is a genuine possibility.

这不是过度炒作。

It's it's not overhype.

这不是悲观主义。

It's not doomerism.

这真的有可能发生。

It's it's it could genuinely happen.

我觉得我们应该感谢一群人，他们似乎确实在定价中考虑到了这种可能性，那就是信用违约互换交易员。

I think we should give credit to one group of people who actually do seem to be pricing in that possibility, and that is the credit default swap traders.

对。

Yeah.

CDS把那些CDS推高到了现在的水平。

CDSs driven those CDSs up to where they are.

我不知道它们现在是什么水平。

I don't know where they are.

这只是给听众们说的。

And that's just for the listeners.

那只是在赌甲骨文会不会倒闭。

That's just betting on whether Oracle will die

而且有点在赌它会不会慢慢死掉。

and have sort dying.

具体来说，是在赌它会不会债务违约。

Specifically betting on whether it will default on its debts.

对。

Yes.

就是说，你为它可能违约的可能性买保险。

Whether it was it's the insurance that you pay on the on the possibility that it will default.

而这份保险的价格正在上涨。

And the price of that insurance is going up.

在那个市场里，人们开始意识到这种可能性是真实存在的，它真的可能发生。

People are beginning to realise in that market specifically that the possibility is real, that it could actually happen.

而且我觉得，还是那个问题，他们是不是反应太慢了。

And I I think there's a question as to, again, whether they're too late on that.

之前难道没有迹象吗？

Weren't the signs there before?

我们难道不能更早看到这一点吗？不是今天，而是六个月、十二个月或十八个月前？

Couldn't we have started to see this coming not today, but six to twelve to eighteen months ago?

我想你的读者和听众会这么认为，他们从你那里听到这个消息已经很久了，然后突然大家都觉得，哦，也许这确实是个问题。

I think that would be the contention of your readers and your listeners who have been hearing about this from you for a very long time, and then suddenly everyone decides, oh, maybe it's a problem.

也许我们会开始写一些相关的文章。

Maybe we'll start to write some articles on this.

也许我们会开始看到CDS掉期的利差扩大。

Maybe we'll start to see those spreads widening on the CDS swaps.

但我们应该给他们应有的肯定，你知道的。

But we should give it give them their props, you know.

一些投资者似乎已经意识到了这个情况。

Some investors seem to be acknowledging the situation.

对。

Yeah.