加文·贝克——暗流涌动的风暴——[投资致胜，第260期]

它们的相对基本面应该是最好的。

Their relative fundamentals should be the best.

在短期内，由于经济放缓以及我们讨论的所有因素，而在十二到十八个月的短期时间内，相对基本面通常驱动相对表现。

In short timeframes, because of the economy slowing and everything we talked about, and in short, twelve to eighteen months timeframes, relative fundamentals generally drive relative performance.

我很想深入探讨一下这个大趋势背后的细节。

I'd love to dive in a little bit just underneath that big trend.

你提到了半导体、软件和互联网这三个子行业。

You mentioned the three subsectors of semis, software, Internet.

也许我们还可以讨论第四个类别，即目前已成为综合性企业的五大或十大科技公司。

Maybe we could also talk about a fourth category, which is the big five or the big 10 technology companies that are sort of conglomerates at this point.

也许我们可以先从半导体开始。

Maybe we'll start with semis.

我过去已经跟你聊过这个话题。

You know, I've talked to you about this in the past.

我对半导体行业完全着迷。

I'm just totally fascinated by the semiconductor industry.

我知道这是你起步的地方，从你职业生涯开始以来你就一直关注这个行业。

I know this is where you cut your teeth a lot, followed it since its beginning and since the start of your career.

两年前，很多人还没听说过台积电，但现在由于各种原因，更多人知道了——不仅仅是短缺和供应链的重要性，还有地缘政治因素。

A lot of people had never heard of Taiwan Semiconductor two years ago, and I think a lot more people have now for a variety of reasons, not just the shortages and the importance of supply chain, but also the geopolitical stuff.

请谈谈你对当前半导体行业的看法，这一子领域在科技行业中有哪些变化和关键点，毕竟它在第一季度如此重要。

Walk us through your take on semis today and what's evolved and what matters in that subsector in tech since it's such a Q1.

让我们退一步，回顾一下过去十五年半导体行业的发展。

Let's step back and look at the last fifteen years of STEMIs.

这个行业已经完全整合，如今几乎每个领域都形成了垄断或准垄断、双头垄断的局面。

The industry has completely consolidated to where you almost have these monopolies and almost monopolies are duopolies.

在半导体的每个细分领域，你要么面对垄断，要么是双头垄断，最差的情况也是三家寡头垄断。

In every sector of semis, you either have a monopoly, duopoly, or in the worst case, an oligopoly of three.

就连半导体公司的设备供应商，也都是垄断或双头垄断的局面。

And then even their suppliers to capital equipment companies, they're all either monopolies to duopolies.

因此，在过去十五到二十年里，这个行业经历了前所未有的高度整合，这种整合本不该被允许发生。

So the industry is massively consolidated over the last fifteen to twenty years in a way that maybe should have never been allowed to happen.

事实上，这些市场之所以如此，有很多原因。

Although the fact is that a lot of these markets, they do for a lot of reasons.

它们主要得益于软件代码的网络效应和规模经济。

They do mostly because of network effects around software code and in economies of scale.

它们往往趋向于形成垄断。

They do tend towards being a monopoly.

垄断或双头垄断。

A monopoly or duopoly.

在基带领域，存在双头垄断。

So in basebands, there's a duopoly.

在CPU领域，存在双头垄断。

CPUs, there's a duopoly.

在GPU领域，存在双头垄断。

GPUs, there's a duopoly.

现在由于英特尔进入该行业，变成了寡头垄断。

Now it's an oligopoly because Intel is entering that industry.

内存方面，NAND和DRAM都属于寡头垄断。

Memory, it's an oligopoly for both NAND and DRAM.

模拟芯片领域，几乎逐个部件来看，通常都是双头垄断。

Analog, almost part by part, it's generally a duopoly.

FPGA也是如此。

Same thing for FPGA.

因此，这是一个高度集中和整合的行业。

So it's a very consolidated concentrated industry.

我认为，需求已经结构性上升。

And you have had demand, I think, structurally shift up.

这一直是一个长期增长的行业。

This has always been a secular growth industry.

它一直以全球GDP的1.5到略高于两倍的速度增长。

It's always grown, I call it, between 1.5 to low twos multiple of GDP, global GDP.

所以，它一直是一个长期增长的行业，但这个乘数已经上升了。

So, it's always been a secular growth industry, but that multiplier shifted up.

它之所以上升， broadly speaking，是因为人工智能。

The reason it shifted up is, broadly speaking, because of artificial intelligence.

人类在编写软件代码时，会付出很大努力来最小化资源使用，至少优秀的程序员会这样做，比如计算和内存资源。

Human beings, when they write software code, they make a big effort to minimize their use of, at least good programmers do, use of resources like compute and memory.

在云计算出现之前，你必须有一个预算限制，只能使用有限的内存和存储空间。

You used to have to have a budget you had to work with before the dates of cloud computing, Only so much memory and only so much storage.

云计算彻底打破了这种限制，而人工智能则恰恰相反。

The way cloud computing can throw that all out the window, and AI is just the inverse.

提升人工智能性能的方法就是用更多的数据来训练它。

The way you make AI better is you train it on more data.

就是这样。

That's it.

其实就这么简单。

It's really just that simple.

这里有一个非常好的经验法则，微软十年前——或者可能是十二年前——在一篇研究论文中就提到过。

And there's just a really good rule of thumb, and Microsoft wrote about this in a research paper ten years ago, or maybe not twelve years ago.

给定AI算法的质量会随着训练数据量增加十倍而翻倍。

The quality of a given AI algorithm doubles with every 10x increase in the amount of data you use to train that algorithm.

马克·安德森十年前写过一篇评论文章，称软件正在吞噬世界。

Marc Andreessen wrote this op ed, whatever it was, ten years ago about how software is eating the world.

现在，人工智能正在吞噬软件。

Now AI is eating software.

这意味着世界正变得越来越依赖计算能力和半导体。

That just means that the world is getting much, much more compute and semiconductor intensive.

此外，归根结底，汽车是一个巨大的消费市场。

Then on top of that, you have all these at the end of the day, cars are a massive, massive consumer market.

随着汽车转变为电动车以及下一代电动车，汽车中的半导体含量正在急剧增长。

And as those become EVs and next EVs, the semiconductor content for car is really exploding.

将这两点结合起来，世界正变得越来越依赖半导体。

You put those two things together, the world is just becoming a lot more semiconductor intensive.

令人担忧的是，我现在对半导体行业的看法可能是多年来最谨慎的，因为它仍然是一个周期性行业。

The bummer, and I would say I'm probably as cautious as I've been on semiconductors in a long time right now, it is still a cyclical industry.

如果你看看这个行业在80年代和90年代的历史，就会发现存在产能周期，这些周期是由一个我记不起他姓什么的人推动的，但他特别搞笑。

If you look at the history of the industry in the '80s and '90s, you have these capacity cycles, and they're driven by the fact that God, I can't remember his last name, but he was hilarious.

TJ，他曾经掌管Cypress和McNulty公司。

TJ, he ran Cypress and McNulty.

他曾经有名地说过：真正的男人拥有晶圆厂。

He famously said, Real men own fabs.

因为当时有一种趋势是转向代工模式，但在80年代和90年代，如果你产能不足，唯一能做的就是建一个新的晶圆厂。

Because there was this trend of going fabulous, but it used to be in the '80s and '90s, if you ran out of capacity, well, the only thing you could do was build a new fab.

每个人都不得不在同一时间用尽产能。

Everybody had to run out of capacity at the same time.

所以所有这些晶圆厂都会在同一时间投产。

So all these fabs would come on at the same time.

因此，你可以把需求看作是平稳的、真实的底层需求——一条相对平滑的线，而产能则以阶梯式的方式增加。

And so you could think of demand as being the smooth underlying true demand, the relatively smooth line, and then capacity comes on in the stair step pattern.

你会经历这些恶性循环，公司总是不断倒闭。

You would have these vicious cycles, and companies were always going out of business.

但后来，世界转向了无晶圆厂模式，只有少数例外。

But then the world moved to fabless with few exceptions.

如今，英特尔、三星、台积电是全球仅有的几家能够制造先进逻辑芯片的公司。

Today, Intel, Samsung, Taiwan, Semi, they're the only companies in the world that can make leading edge logic.

全球只有三家公司的先进DRAM制造能力，NAND可能有四家。

There's only three companies that can make leading edge DRAM, maybe four for NAND.

因此，他们在平滑整合产能方面变得更好了。

And so they got much better about aggregating capacity smoothly.

因此，过去二十年中看到的周期只是库存周期。

As a result of that, the cycles you've seen in the last twenty years are just inventory cycles.

其原因是支配SIMIs的基本方程：客户库存必须等于交货周期。

And the reason for that is the fundamental equation that governs SIMIs is customer inventories must equal lead times.

因为如果两者不相等，你作为采购经理就会被解雇。

Because if they don't, then you're a purchasing manager, you get fired.

因此，无论交货周期是多少，客户库存就会是多少。

And so, whatever lead times are, that is what customer inventories are.

这就导致了一个疯狂的正反馈循环：当交货周期延长时，库存就会增加，而这又进一步推高了交货周期，使库存继续累积。

And that leads to this crazy positive feedback loop where if lead times are going up, inventories are building, which causes lead times to go up, which causes inventories to build further.

而一旦有任何变化，所有这些都会迅速逆转。

And then as soon as something changes, all of that unwinds.

于是交货周期开始缩短，库存被消耗。

And then lead times are going down, you're burning inventory.

因此，如果你是一家半导体公司，你永远看不到真实的终端需求。

So if you're a semiconductor company, you're never seeing true end demand.

你看到的要么是高于市场真实需求的水平——因为交货周期拉长、库存累积；要么是低于真实终端需求的水平。

You're either seeing above market demand because lead times are going out and inventories are building or below market demand are below true end demand.

因此，在过去二十年里，你一直经历着这种持续不断的库存周期。

And so you've had these inventory cycles really consistently for the last twenty years.

你现在所经历的，我认为是一次大规模的库存周期。

What you have right now is, I think, a massive inventory cycle.

我们之前讨论的一切——经济放缓、甚至在美联储加息之前，个人消费支出就已从商品转向体验——我认为对半导体的需求几乎不可避免地会略有放缓，而这将导致这一库存周期的消解。

And everything we talked about, the economy slowing, even before the Fed hikes hit, PCE shifting away from goods towards experiences, I think demand for simmies is almost inevitably going to decelerate a little, and that's going to lead to an unwind of this inventory cycle.

然后所有新增的产能可能让情况变得更糟。

Then all the capacity that's being brought on probably makes it worse.

但当你度过这一阶段后，你会发现这个行业从过去以两倍于名义GDP的速度增长，变成了现在可能以三倍于名义GDP的速度增长。

But then I think you get to the other side of that, and you're left with an industry that used to grow at 2x nominal GDP to one that probably now is a 3x nominal GDP grower.

你仍然拥有这种高度集中的供应链结构。

You still have that super consolidated supply structure.

现在有些人说半导体公司应该像评估软件公司那样来评估自己。

You're now having people saying semiconductor companies should be evaluating software companies.

不，他们不应该。

And no, they shouldn't.

你有一大堆人。

You have a bunch of people.

我认识的每一个管理资产低于500亿美元的基金，都在疯狂寻找半导体分析师，或者真正懂芯片的人。

Every fund I know that's under 50,000,000,000 is frantically looking for a semiconductor analyst or somebody who's really good at simmies.

这个领域里有很多投机者。

Have a lot of tourists in the sector.

当这些公司失误时，失误往往非常严重。

And it's just when these companies miss, they miss big.

回看一下2018年。

Just go back to the 2018.

你可以看到一些非常、非常、非常大的失误。

You could see some really, really, really big misses.

因此，我会说补贴应相对谨慎。

So I would say relatively cautious subsidies.

上次我们交谈时，大家对软件公司在新冠疫情期间如此高的客户留存率和净收入扩张率是否可持续，提出了很大疑问。

Last time we talked, I think there was a big question about how secure these incredible retention numbers and net dollar expansion numbers were for software companies in the depths of COVID.

我想知道你对此的看法是否有所改变。

Curious how your thinking there has changed.

以Zoom为例，它曾经历疯狂上涨，如今又经历疯狂下跌。

I mean, use Zoom as like the obvious example of just insane run up and now insane run down.

正如我们所有人已经采用它一样，这家公司的基本面非常出色。

The fundamentals of the business are pretty amazing as we've all adopted it.

我们现在正在讨论这个问题。

We're talking on it right now.

跟我讲讲你对软件行业的看法，以及它经历过的过山车般的起伏。

Talk me through your views on software and kind of the roller coaster that it's been on.

是的。

Yeah.

首先，我得说，我错了。

Well, first, I have to say, mea culpa, I was wrong.

预测未来真的非常非常困难。

Predicting the future being very, very hard.

我想上一次我上你的播客时，曾说过类似软件不会比第一顺位债务更好的话。

I think the last time I was on your podcast, I said something to the effect of software is not going to be better than first lien debt.

好吧，我错了。

Well, I was wrong.

软件的表现优于第一顺位债务。

Software was better than first lien debt.

它只是全面地表现出了这一点。

It just was really across the board.

所以我真的、真的错了。

So I was really, really wrong.

而今天，我认为你现在看到的是，我确实对那些产生现金流的公司有偏爱。

And today, I think now you're looking at, I do have this bias for companies that are generating cash flow.

这对我来说是一条明确的界限。

That's a big line in the sand for me.

当我看到一家公司估值达到销售额的10倍、15倍时，如果它能实现2%、2.5%、3%的自由现金流收益率，与那些仍在烧钱的公司相比，两者之间有着天壤之别。

There's a huge difference to me when I look at a company over 10 times sales, over 15 times sales, between having a two, two and a half, 3% free cash flow yield looking out to 23 and a company that's still burning cash.

我对软件行业非常非常看好。

I'm really, really bullish on software.

尤其是在增长放缓的世界里——我相信我们确实正处于这样一个阶段——软件就像是科技领域的必需消费品。

It is particularly in a world of slowing growth, which I do think for sure we are in for, it's the consumer staple of tech.

它应该具备最好的相对基本面。

It should have the best relative fundamentals.

在很多方面，它们的基本面甚至比必需消费品还要稳定。

In a lot of ways, their fundamentals should be more stable than even staples.

我对这一点非常非常看好。

I'm really, really positive on it.

我确实有一个偏向，相对于应用软件公司，我更倾向于基础设施公司。

I do have a bias, I would say, towards infrastructure companies relative to application software companies.

有很多应用软件公司并没有太多真正的技术。

There are a lot of application software companies that don't have a lot of true technology.

我认为，随着在云上构建自己的应用、使用自己的数据且不共享数据变得越来越容易，它们面临一定的风险。

I do think they're a little bit at risk from as it gets easier and easier and easier in the cloud to build your own applications, use your own data, not share it.

过去，首席信息官们害怕上云，因为他们担心会失去所有预算。

It used to be like the CIOs were scared of going to the cloud because they thought they'd lose all their budgets.

他们会失去乘坐公司专机飞行、去看超级碗比赛等特权。

They'd lose being able to fly around on the corporate jet and go to the Super Bowl or whatever it was.

这些只是名义上的借口，安全问题；而几次高调的黑客事件之后，大家意识到，从安全角度来看，不上云的风险反而比上云更大。

They're nominal excuses, security, and then couple of high profile hacks that everybody realized it was more of a risk to not be in the cloud than to be in the cloud from a security perspective.

但现在我认为，在这些大型云服务商那里，你真的可以做到。同样，人们曾以为IT服务会被摧毁。

But now I think you really can, at these big cloud vendors, Similarly, thought IT services were going be destroyed.

这些IT服务公司现在报告了创纪录的季度业绩和创纪录的待办订单。

These IT services companies here are reporting record quarters, record backlogs.

这对它们来说非常好。

It's great for them.

原因在于，如果你是财富500强公司的CIO，你现在可以自己构建应用、掌控自己的数据。

And the reason is is now if you're that CIO at a Fortune 50 company, you can build your own apps, control your own data.

这比以往任何时候都容易多了。

It's way easier than it ever was.

归根结底，这正是许多CIO真正想做的事情。

At the end of the day, that is what a lot of these CIOs want to do.

因此，我确实非常偏向于那些建立在这些云计算巨头之上的基础设施软件。

So I do have a big bias towards infrastructure software that rides on top of these cloud computing giants.

在终值方面也是如此吗？

Is that true in terminal value too?

所以，如果你只是看一下前10到15家软件公司，它们大多数仍然以应用程序为主或占据主导地位，比如Salesforce、Adobe、Intuit，还有许多类似这样的公司，还有刚刚被微软收购的动视暴雪，我们不妨聊聊这个，挺有意思的。

So if you just look at, I don't know, the top 10 or 15 software companies, most of them are still application heavy or dominant, you know, Salesforce, Adobe, Intuit, lots of companies like this, Activision Blizzard, which got bought today by Microsoft, which we should riff on and have fun with.

但目前在市值最高的那部分市场，似乎仍然以应用为主，而其中一些公司仍然是私营的。

But it does seem very app heavy still at the top end of the market cap spectrum, whereas some of these companies are still private.

但像Stripe、Twilio这样的公司，还有云服务提供商、Datadog等这类更侧重基础设施的公司，仍在崛起中。

But the Stripes of the world, the Twilios of the world, the cloud providers, the Datadogs, all these kinds of companies that are more infrastructure focused are still up and coming.

你如何看待这一切的最终形态？

How do you think about, like, the end state of all this?

你的这种观点是否不仅限于当前的机会，还延伸到了未来基础设施占据主导地位的状态？

Does that opinion of yours extend beyond just the current opportunity set to some future state where infrastructure is dominant?

当然，这种观点延伸到了未来状态。

For sure, it extends to a future state.

我确实相信，在某种程度上，谷歌是第一家明确这样说的云服务商：好吧。

And it is definitely a belief that at some level I do think Google was the first cloud to say, Okay.

马克·比尼亚夫会讲一个关于AWS的故事，说他们在尽职调查软件公司时，根本无法克制自己。

Mark Biniaff would tell this parable about AWS them diligencing software companies about how they just couldn't help themselves.

这是一个蝎子和大象的寓言。

It was a parable of the scorpion and the elephant.

大象要渡过一条汹涌泛滥的河流，而蝎子也需要过河。

The elephant get across, like, a raging river that's flooded, and the scorpion has to get across it.

它身后有火灾或洪水，蝎子说：嘿。

There's a fire or flood behind it, and the scorpion says, hey.

你能带我过河吗，大象？

Will you take me across, elephant?

不然我会死的。

Because I'll die otherwise.

大象说：当然。

And the elephant is like, yeah.

好的。

Sure.

我想做个好人。

I wanna be nice.

我怎么知道你不会蜇死我？

How do I trust you that you're not gonna sting me to death?

然后蝎子说：如果我蜇死你，我自己也会死。

And then the scorpion's like, well, if I sting you to death, I'm gonna die.

那样的话，我们都会死。

Then we'll both die.

所以，我当然不会那么做。

So, of course, I'm not gonna do that.

然后大象说：好吧。

And then the elephant's like, okay.

爬到我背上来。

Get on my back.

蝎子爬上了它的背。

Scorpion gets on its back.

它们正过河时，蝎子却蜇了大象。

They're going across the river, and then the scorpion stings the elephant.

大象说：‘你为什么要这么做？’

The elephant's like, Why'd you do that?

当他们双双濒死时，蝎子说：‘这是我的天性。’

As they're both dying and then a scorpion says, It's in my nature.

我忍不住。

I couldn't help it.

是啊。

Yeah.

我喜欢这个。

I love that.

所以比迪经常讲这个故事，说别让AWS的审慎拖垮你的公司，因为他们本性难移。

So Biddy often tells that story about how don't let AWS diligence your company because they cannot help themselves.

他们不会收购你。

They're not going to buy you.

他们会试图自己打造，你已经看到他们试图自己构建了

They're going to try and build you've seen them try to build their own

这和弹性突发有关。

That with elastic surge.

真的发生了。

It happened.

你可以说它在

You could say it in a

一大群人中，没错。

bunch Exactly.

是啊。

Of Yeah.

很多很多次。

Many, many times.

他们试图分叉开源基础设施，自己构建。

They try and fork the open source infrastructure build themselves.

谷歌是第一个意识到：嘿，这有点傻。

Google was the first one to realize, Hey, this is kind of silly.

我们之外正在发生大量的创新。

There's a lot of innovation happening outside of us.

其次，仅仅提供这些基础组件就能赚到很多钱。

B, there's a lot of money to be made just in providing these primitives.

让我们停止直接在云之上与这些独立的基础设施软件提供商竞争。

Let's stop trying to compete with these independent infrastructure software providers right on top of the cloud.

他们可以成为我们的渠道。

They could be a channel for us.

他们可以为我们推动需求。

They can drive demand for us.

顺便说一下，现在很多这些公司也非常重要且具有战略意义，因为其中一些公司也有本地部署业务，而本地部署现在就意味着私有云。

And by the way, a lot of these companies now are also very important and strategic because some of them have on prem businesses too, and on prem just means private cloud now.

如果你采用这些公司在其数据中心所使用的相同基础设施软件，就能让他们更容易迁移到公有云。

You make it easier for these companies to migrate to the public cloud if you embrace the same infrastructure software that they're using in their own data centers.

但我确实认为，这是一种观点，即软件世界不仅仅会由AWS、GCP和Azure主导。

But I do think it is a little bit of a view that, Hey, the world of software is not just going to be AWS, GCP, and Azure.

他们不会向上游发展并消费所有的基础设施功能。

They're not going to go up the stack and consume every infrastructure function.

这是一种观点。

That is a view.

我觉得这很有趣。

I do think it's interesting.

你提到了终值，但 broadly speaking，这种观点认为，那些拥有传统企业市场策略的应用提供商，仅仅提供一个漂亮的用户界面，以及大量甲骨文和 SAP 的销售人员在推销一款取代电子表格或其他类似产品的软件，如今这些公司的终值倍数要低得多。

You touched on terminal value, but broadly speaking, this view that these app providers who have a traditional enterprise go to market and they just have a pretty UI and a lot of Oracle and SAP salespeople selling a product that replaces a spreadsheet or whatever it is, those terminal multiples today are much lower.

我认为市场明白这是一种风险。

I do think the market understands that this is a risk.

但你确实必须相信，软件的世界不会成为一个寡头垄断，而我相信这一点。

But, yeah, you for sure have to believe that the world of software is not going to be an oligopoly, and I believe that.

我认为 AWS 是最后一个慢慢放弃这一观点的公司。

And I think AWS is the last one that's slowly giving up on this.

GCP 是第一个拥抱这一点的，然后是 Azure。

GCP was the first to embrace it, then then Azure.

如果Azure正在拥抱这些第三方基础设施软件提供商，那么所有人都在这么做，因为他们拥有比任何人都更强大的软件能力。

And it's like if Azure is embracing these third party infrastructure software providers, everybody is, because they have more software capability than anyone.

另一个有趣的事情是我有自己的偏见。

And then the other interesting thing is I do have a bias.

我并没有持有这些公司的任何股份，但如果你要投资一家应用软件公司，我更倾向于那些服务中小企业的公司。

I don't own any of these, but if you are going to own an application software company, I have a bias towards ones that serve SMBs.

因为这些中小企业无法在AWS上自行开发应用程序，因此他们免受这种影响。

Because those SMBs are not going to be able to roll their own apps in AWS, so they're shielded from that.

网络安全是另一个我原本对此持开放态度的领域，但我一直从结构上看空网络安全，因为网络安全是市场中。

Cybersecurity, that's another thing I was, I guess, was open minded to this, but I was always structurally bearish on cybersecurity because cybersecurity is the area of the market.

从零做到十亿美元总是最容易的，但之前几乎不可能突破一千五百亿到两千亿的规模。

It was always easiest to go from zero to a billion, but basically, it was previously impossible to go over 15 to 20,000,000,000.

原因在于，成功反而埋下了自我毁灭的种子，因为攻击者会开始专门针对你进行优化。

And the reason was success sowed the seeds of your own destruction because the attackers would start to optimize for you.

风险投资人都明白这一点。

The VCs knew this.

他们一直在资助竞争对手。

They were always funding competitors.

它的规模和体量，是科技领域中少数几个规模越大反而越有害的领域之一。

Its size and scale, it was one of the only areas of tech where it hurts you rather than it helped you.

嗯，人工智能改变了这一点。

Well, AI has changed that.

现在，如果您的模型在学习，如果您真正是一家以人工智能驱动的网络安全公司，规模对您是有利的。

Now that attack surface, if your models are learning, if you're truly an AI driven cybersecurity company, scale is good for you.

这就是为什么有史以来第一次，很长一段时间以来，别管确切数字了，你从未见过任何网络安全公司的估值超过200亿美元。

That's why for the first time ever, for a long time, forget the exact numbers, but you never had a cybersecurity outcome over 20,000,000,000.

现在，已经有公司突破了这一界限，从市值角度看仍在快速增长，类似的情况也发生在这一领域。

Now, you have companies that have blown through that, still going fast from a market value perspective and kind of similar thing for this.

这就是一家独立公司能在自身重压下崩溃前达到的最大规模。

This is how big an independent could get before they collapse under their own weight.

人工智能真正改变了网络安全。

AI has really changed cybersecurity.

最后一个要简要提及的类别是最近的新闻，这为我们提供了一个契机，来聊聊一些事情，然后在结束前再讨论几个有趣的话题。

The last category to touch on briefly, recent news is an excuse to talk about a few of the things and then maybe talk about a couple couple other fun topics before we close.

首先是互联网和软件之间的区别。

The first is the difference, I guess, between Internet and software.

如果它们是两个独立的类别，那么它们之间的差异似乎显而易见。

If those are two separate categories, it's kind of obvious like what defines the difference.

但也许可以更详细地定义一下这种区别：它们的交易方式、关键变量，以及你今天的看法。

But maybe define the difference in a little more detail, how they trade, the variables that matter, your view on them today.

再多谈一点关于互联网作为一个类别的情况。

Say a bit more about Internet specifically as a category.

我认为，广义上讲，互联网公司大致有三种，其中一种在很多方面其实就是软件公司。

I'd say there's broadly speaking three kinds of Internet companies, and one in a lot of ways is software companies.

这三种互联网公司分别是电子商务公司、广告公司，以及像Netflix和Spotify这样的订阅制公司。

So the three kinds of Internet companies are ecommerce companies, advertising companies, and then subscription companies like Netflix and Spotify.

我想说的是，从定义上讲，电子商务和广告业务对GDP的敏感度远高于软件。

And I would just say, definitionally, ecommerce and advertising are way more GDP sensitive than software.

因此，他们正感受到当前经济放缓的影响。

So they're feeling this slowdown that's happening right now in the economy.

此外，他们还承受着新冠疫情的后遗症，而软件公司则不会。

And on top of that, they are suffering a COVID hangover that I just software is not.

价格总是会催生出各种叙事。

Price always leads to invented narratives.

如今，由于软件表现疲软，人们开始说MuleSoft表现不佳，因为它是一个领先指标，涉及API集成。

So now that software has been weak, there's all this, oh, MuleSoft was weak, and they're a leading indicator because API integrations.

不对。

No.

这只是一个人为编造的叙事。

This is an invented narrative.

但几乎所有互联网公司都确实存在真实而根本性的疲软，可能只有搜索业务除外。

But there is real legitimate fundamental weakness in almost all of the internet names, maybe with the exception of search.

这仅仅是因为，搜索可能是互联网领域中唯一与服务支出复苏紧密关联的部分。

And that's just because probably search is the only part of the internet that is levered to this resurgence in services spending.

你确实能看到相当广泛的。

You do see pretty broad based.

如果我在播客里这么说，有人就会这么做。

If I say this on the podcast, somebody will do this.

在我眼里，我总是把软件公司、订阅制互联网公司、任何特许经营模式和任何数据商业模式都看作是一样的。

In my mind, I always just look at software companies, the subscription Internet companies, any franchise business model, and any data business model, they're all the same to me.

它们都有持续的收入。

They have recurring revenues.

它们拥有定价权，值得思考的是相对的基本面、相对的资本回报率和相对的估值。

They have pricing power, and it's interesting to think about the relative fundamentals, the relative ROICs, and the relative valuations.

它们之间的差异并不大。

There's not a lot of differences.

如果你是一家特许经营的餐厅或酒店公司，你只是收取特许权使用费。

If you're a franchised restaurant or hotel company, you just have a royalty.

虽然这是一种对经济敏感的特许权使用费，但它仍然是特许权使用费。

Now, it's an economically sensitive royalty, but it's a royalty.

你的资产非常轻。

You're very asset light.

人们实际上通过超额使用来设置订阅模式。

People effectively set up for subscriptions with overage.

在这方面，特许经营者看起来更像是新一代不销售合同、而是按使用量收费的软件公司，这在运作良好时非常有利。

In that respect, franchisors, they look more like a next gen software company that doesn't sell contracts, but they sell on consumption, which is good when it's working for you.

但当模式运作良好时，你又希望当初签了合同。

Then when setup's working for you, wish you had those contracts.

但我想说，互联网整体上正遭受着新冠疫情后的余波，而软件行业似乎并没有这种问题。

But I would just say internet, broadly speaking, is just suffering from a COVID hangover that it doesn't appear software is.

摩根大通刚刚发布报告称，我们正急于上云。

Mean, JP Morgan just came out and said, We're in a panic to go to the cloud.

我们将花费比任何人预期都多得多的钱。

We're going spend way more money than anyone thought.

我认为2022年将是有许多公司表现最好的一年。

And I think you're going to see 2022 be the best year from a lot of companies.

顺便说一下，有很多来自CIO的调查支持这一点。

By the way, there's all sorts of surveys that support this from CIOs.

也许我们可以聊聊最近的一些事件，谈谈一些感兴趣的话题。

Maybe a couple recent events to just talk about some favorite topics.

我先从Facebook更名为Meta以及整个元宇宙趋势说起。

I'll start with the change of Facebook's name to Meta and just the entire metaverse trend.

我想我们最初认识，是因为讨论科幻小说。

I think the way that you and I originally got to know one another was talking about sci fi.

我知道，科幻小说在你职业生涯早期坚持从事科技行业，以及你对科技的普遍兴趣中发挥了巨大作用。

And I know sci fi has played a huge role in your sticking with tech in the early part of your career and just your interest in technology, generally speaking.

重要的是要记住，我不记得是谁说过，很多关于元宇宙的科幻小说都是反乌托邦的，而不是乌托邦的。

Important to remember, I can't remember who said it, that a lot of the sci fi novels with the metaverse, was dystopian, not utopian.

但跟我谈谈元宇宙吧。

But talk to me about the metaverse.

我的意思是，这是一个非常庞大的话题，我之前做过几期相关内容。

I mean, it's such a huge topic that, you know, I've done some episodes on.

人们对这个想法着迷。

People are fascinated by this idea.

这是一个宏大的构想。

It's a big idea.

世界上最大的公司之一正在大力押注于此。

One of the biggest companies in the world is betting on it in a major way.

你对突然间人人都知道并关心元宇宙这件事有什么反应？

What's been your reaction to the all of a sudden, everyone knowing what the metaverse is and caring?

现在我用这个词时感到有点尴尬。

Well, I feel embarrassed to use the word now.

会是。

Would be.

是的。

Yeah.

是的。

Yeah.

这很烦人。

It is annoying.

我喜欢在像消费产品公司这样的领域里当个早鸟和极客。

I like to be early and nerdy when you have, I don't know, consumer products companies.

他们说，我们要打造自己的元宇宙，我们要做这个。

They say, we're gonna have our own metaverse, and we're gonna do this.

对我来说，这感觉有点过时了，但它仍然是真实的。

It feels a little overdone to me, but it's still real.

它依然在发展中。

It's still coming.

我对PlayStation VR头显感到兴奋。

I'm excited for the PlayStation VR headset.

我很期待苹果会做什么。

I'm excited to see what Apple does.

Meta会继续前进，不断迭代。

Meta is going to keep going, keep iterating.

这只是一种趋势，会一直持续下去，直到我们所有人都拥有脑机接口，可以闭上眼睛做任何事情。

It's just this trend that's going to continue until we all have brain computer interfaces and we can close our eyes and whatever.

当一个精灵战士，或者做我想做的任何事。

Be an elven warrior or do whatever I want to do.

所以我仍然非常、非常相信它。

And so I still really, really believe in it.

顺便说一句，我要说，这是我之前犯错的另一件事。

And by the way, I would say that's another thing I've been wrong about.

对我来说，唯一的真正元宇宙就是今天，它们是什么？

To me, the only real metaverse is today, what are they?

军事元宇宙存在，而且基本上由《使命召唤》主导。

Well, there's military metaverse, and it's pretty much dominated by Call of Duty.

还有两个卡通风格、适合朋友和孩子的元宇宙，那就是《Roblox》和《堡垒之夜》。

There's two cartoony, friends fiddly kids friendly metaverses, Roblox, and Fortnite.

然后是幻想元宇宙，那就是《魔兽世界》。

Then there's fantasy metaverse, and that's World of Warcraft.

这些确实是发生商业活动、人们生活其中的元宇宙。

And these really are metaverses where commerce happens, people live their lives.

科幻元宇宙。

Science fiction metaverse.

那就是《命运》。

That's destiny.

我花了很多时间在上面。

I spend a lot of time in that.

你还是得玩这个，帕特里克。

You still gotta play that, Patrick.

但这些股票，首先，我认为从广义上讲，还没有真正反映在它们的基本面或价格上。

But those stocks, a, it hasn't, I would say, broadly speaking, really filtered through to their fundamentals and or their prices.

但我仍然认为，就我们的元宇宙而言，它们就是电子游戏，而这些就是当今真正的元宇宙。

But I do still think that, like, in terms of where our metaverse is, well, they're video games, and these are real metaverses today.

不过，没错，元宇宙我觉得仍然在来的路上。

But, yeah, the metaverse, I think it's still coming.

你认为，用今天另一个头条来说，动视被微软收购，是否正是因为它们想掌控人们花费数字时间的场所？

And do you think that, to use the other headline of the day, you know, Activision joining Microsoft, is a lot of this just about wanting to own the place where people spend their digital time?

从根本上说，这股趋势的走向是朝着积极参与发展，这意味着视频游戏将胜过社交媒体或Netflix或流媒体娱乐。

And fundamentally, that's just the arc of that bends towards active participation, which means video games versus social networks or Netflix or streaming entertainment.

你怎么看待这一趋势的长期走向？为什么各大科技公司突然都需要制定元宇宙战略？

How do you think about the long arc of this trend and why it seems like the big technology companies all of a sudden need to have a metaverse strategy?

我不确定微软收购动视是否就是为了这个目的，但确实很有趣。

And I don't know if that's what Microsoft's doing with Activision, but it sure seems interesting.

你怎么看这个问题？

How do you think about that?

是的。

Yeah.

毫无疑问。

It is for sure.

我认为微软收购动视、贝塞斯达、《辐射》和《上古卷轴在线》的原因就在这里。

I think what is happening with Microsoft and Activision, it's why they bought Bethesda, Fallout, the Elder Scrolls Online.

这些也都是小型元宇宙。

Those are also little metaverses.

这就是他们收购Minecraft的原因。

That's why they bought Minecraft.

这是另一个适合孩子的元宇宙，我本该把它和《堡垒之夜》和《Roblox》放在一起说。

That's another kid friendly metaverse that I should have thrown in with Fortnite and Roblox.

微软的情况很有趣，因为它们有HoloLens合同。

Microsoft, it's just interesting because of their HoloLens contract.

它们几乎肯定会拥有最纯粹的AR/VR收入。

They're almost certainly going to have the most pure ARVR revenue of any company.

它们掌控着这个视频游戏平台。

They control this video gaming platform.

几乎所有的PC游戏都在Windows电脑上进行。

Almost all PC gaming happens on Windows PCs.

我认为它们正缓慢地试图将用户从Steam拉回来。

And I think they're slowly trying to bring that back away from Steam.

然后他们还有这么多优秀的第一方视频游戏。

And then they have all these great first party video games.

所以我认为他们拥有非常丰富的资产组合。

So I think they have a very interesting collection of assets.

所有这些公司，我只是觉得它们本质上是全球GDP的杠杆型特许权收入。

All those companies, I just think of them, they're just levered royalties on global GDP.

你几乎可以做一个群体分析，这对所有这些公司来说都极其看涨，尤其是取代奈飞。

You could almost do a cohort analysis, and it's so bullish for all of those companies taking out Netflix.

奈飞无疑在某种程度上是全球GDP和消费者闲暇时间的杠杆型特许权收入。

Netflix is for sure, at some level, a levered royalty on global GDP and consumer free time.

如果你想想互联网广告与云计算公司，那就像是按公司成立年份来分类：我打赌五十年前成立的公司，它们在广告或云计算上的支出大概只占总收入的2%，我就随便猜的。

If you think about the internet advertising with cloud computing companies, it's like if you just bracket companies by the year that they were created, It's like I bet companies that were created fifty years ago, I bet they're spending, I don't know, I'm going make this up, 2% of their total revenue on either advertising or cloud computing with those companies.

明白吗？

Okay?

如果你是一家过去三年内成立的公司，一个很好的例子是，你可能将50%的收入花在了这些公司上。

If you're a company that was created in the last three years, a good case is you're only spending 50% of your revenue with those companies.

这是一个很好的例子。

That's a great case.

每年，世界都在向它们靠拢。

Every year, the world moves towards them.

如果半导体以三倍于GDP的速度增长，它们的整体增长可能会更高，而且过去十年确实如此，但只要它们不被拆分，比如TikTok，除非出现真正的平台范式转变，否则这种基于全球GDP的杠杆化收益将继续存在，并由我刚才描述的同侪分析所支撑。

If semiconductors grow at 3x GDP, they may collectively grow at a higher multiple of that, and they certainly have the last ten years, but that may continue as long as they could be broken up, TikTok, unless you have a real paradigm shift to platform shift, that levered royalty on global GDP is going to continue supported by that cohort analysis that I just described.

这就是为什么它们也专注于元宇宙。

And that is why they're also focused on metaverse.

每个人都想拥有下一个平台。

Everybody wants to own the next platform.

每个人都认为元宇宙是下一个平台。

Everybody thinks the metaverse is the next platform.

大家都在为此努力。

Everybody's working on it.

我想起我们第一次播客时说的，我认为苹果和安卓将拥有显著优势，因为它们控制着每个人口袋里的平台，而最终，元宇宙很可能需要具备便携性，VR和AR将融合为XR扩展现实。

And I go back to our first podcast, I think it's going to be Apple and Android, some pretty profound advantages because they control the platform that's in everybody's pocket, and ultimately, the metaverse are probably going to want it to be portable, and VR and AR are going to move it, fuse into XR extended reality.

很难不看到主导的移动操作系统具有明显优势。

It's hard not to see how the dominant mobile operating systems aren't advantaged.

顺便说一句，这令人兴奋，因为我确实认为这意味着，嘿，你想参与元宇宙吗？

And by the way, that is something exciting because I do think it means, like, hey, you want to play in the metaverse?

你很可能需要一个移动操作系统。

You probably need a mobile operating system.

所以我认为你会看到这些公司开始产生交集。

So I think you're going to see some of these companies start to collide.

苹果将会做搜索。

Like Apple's going to do search.

他们最终一定会做。

They're eventually going to do it.

我认为你会看到更多的移动操作系统竞争。

I think you'll see more mobile OS competition.

这些生态系统将以一些有趣的方式发生碰撞。

Some of these ecosystems will collide in ways that are interesting.

我所不

One of the things I don't

认为你我讨论过的是，这一切对那些五十年前成立、如今仍只将2%或类似比例的预算投入这些现代平台的公司所产生的影响。

think you and I have talked about is the impact of all of this on those companies that were founded fifty years ago and are spending the 2% or whatever it is on these modern platforms.

你如何看待这一机遇？无论是作为投资者还是对企业而言，如何将这2%提升到20%、30%或更多？

How do you think about that opportunity, both as an investor and for businesses, to bring the 2% to 20% or 30% or whatever?

我认为这至关重要。

I think it's pivotal.

当我对2020年至2021年期间的长期增长持谨慎态度时，我非常关注那些廉价的消费周期性股票，我认为它们的商业模式已实现真正结构性且持久的改善。

When I was cautious on secular growth from, call it, 2020 to 2021, I was very focused on cheap consumer cyclicals that I thought had really structurally and durably improved their business models.

我认为有很多零售商和品牌确实做到了这一点。

I think there's a lot of retailers and brands that did that.

顺便说一句，你会经历很多阶段。

By the way, there's so many phases you go through.

在足够长的时间线上，每一家互联网公司最终都会卖广告。

On a long enough timeline, every internet company sells ads.

这些零售商都将开始在线销售广告。

Well, all these retailers, they're going start selling ads online.

它们要成为现代互联网企业，前方还有大量空间。

There's so much still in front of them from becoming a modern internet business.

我认为仍然存在很多机会。

Still think there's a lot of opportunity.

总的来说，对于那些站在变革正确一侧的零售商，我仍然对许多这类企业非常看好。

I would say for retailers that are on the right side of change, broadly speaking, I'm still very excited about a lot of those businesses.

现在我可能对软件的兴趣比对这些企业更大，但我认为它们不会回到过去那种长期被看空的状态。

Probably now much more excited about something like software than I am about those, but I don't think they're going back to being perma shorts that they were.

我认为有很多人本能地只想做空零售商，但这些零售商中的许多企业将会长期存在。

I think there's a lot of people who reflexively just want to short retailers, but a lot of these retailers, they are here to stay.

它们已经以非常巧妙的方式将线上业务与实体店网络整合在一起。

They do have big online businesses integrated the store network with their e commerce operations in really cool ways.

顺便说一句，这正是贝索斯过去二十年来一直担心的问题。

And by the way, this is what Bezos was worried about for twenty years.

在线购买，店内自提。

Buy online, pick up in store.

全球最大的电子商务公司都在开设实体店。

All the world's largest ecommerce companies, they're opening stores.

实体店是有价值的，但前提是它们与现代IT系统深度融合。

Like, stores have value, but if they're really well integrated with a modern IT system.

顺便说一下，在软件的背景下，必须意识到这些资产正面临一场疯狂的私募股权收购热潮。

By the way, in the context of the software, do think it's important to realize there is a crazy underlying private equity bid for these assets.

这将为价格提供一个底部支撑。

It just it will put a floor under.

Tomo Bravo 刚刚筹集了一笔200亿美元的基金。

Tomo Bravo, they just raised a $20,000,000,000 fund.

sidelines 上有大量的闲置资金。

There's a lot of dry powder on the sidelines.

顺便说一下，帕特里克，我还想强调一下关于价值与增长的另一个重要观点。

By the way, Patrick, another point I just wanna make about value versus growth that I think is so important.

每次增长表现开始落后时，都会有很多人惊呼：天哪。

Every time growth starts to underperform, there are a lot of people who are like, oh my god.

这太棒了。

This is amazing.

泡沫破裂了。

It's the bubble crashing.

我们将回到2001年到2008年价值股主导的时代。

We're gonna go back to the way it was for value from o one to o eight.

辉煌的日子又回来了。

The glory days are here again.

太惊人了。

It's amazing.

顺便说一下，正如我所说，我并不总是看多增长股。

And by the way, like I said, I'm not always bullish on growth.

我几乎整整一年都极度偏向价值股。

I was tilted very, very value heavy for nearly a year.

但我认为，回到股票收益的两个驱动因素——估值和盈利，这种分析遗漏了关键点。

But I think what's missing from that analysis, going back to the two components, what drives stock returns, multiples, and earnings.

价值股在科技泡沫破裂后表现优异，原因不仅仅是估值上升，更是因为盈利出现了爆炸式增长。

The reason value was so good after the tech bubble burst wasn't just that multiples expanded, but it was that earnings exploded.

为什么盈利会爆炸式增长？

Why did earnings explode?

因为当时有吉姆·奥尼尔，还有那些砖块。

Because you had Jim O'Neill, the briquettes.

中国加入世贸组织之后。

Had post China entering the WTO.

中国经历了大规模的工业化，这导致对石油、铁矿石以及你能想到的每一种原材料的需求出现了质的飞跃。

You had just this massive industrialization in China that was a true step function in demand for oil, for iron ore, for every material you could think of.

他们正在建设那么多机场。

They're building so many airports.

他们购买了大量飞机、汽车和各种东西。

They're buying so many airplanes, cars, everything.

而这正是真正推动它的原因。

And that was really what drove it.

我认为有时人们忽略了，在那段时期，需求出现了巨大而根本性的上升，推动了众多这些价值股的表现。

And I think sometimes people miss that is there was a massive, massive fundamental upshift in demand that drove so many of these value stocks in that period.

我今天看不到这种情况。

I don't see that today.

即使当我想到一个高通胀的世界，存在工资通胀时，我也看不出像当年中国入世时那样，对这些传统重工业产生类似的 Demand 冲击。

Even when I think of a higher inflation world where you have wage inflation, it's like I just don't see any comparable demand impulse for a lot of these traditional smokestack industries like you had with the bricks during the February.

你与这些澳大利亚、南美洲的大型公司会面，但你会发现，我们不确定世界上是否有足够的铁矿石。

You meet with these Australian, South American giant companies, but it's like, we're not sure there's enough iron ore in the world.

别提曾经流行的‘石油峰值’理论了。

Forget peak oil, which was a popular theory.

我们担心的是‘一切的峰值’。

We're worried about peak everything.

事实证明，地球很大，而那确实是一次性的建设浪潮。

Well, it turns out the earth is a big place, and it really was a one time build.

我认为，我们第一次上节目时，曾讨论过价值投资难以开展的一个原因，那就是算法确实面临职业风险。

I think maybe the first time we were on, we discussed one reason value investing has been hard is that algorithms, they do have career risk.

但它们只是不知道而已。

They just don't know it.

它们根本不在乎。

They don't care.

它们没有情感。

They don't feel emotion.

人们曾经谈论的价值投资所面临的职业风险，现在已经消失了。

That career risk that people would talk about being with value investing, that's gone.

但同样地，这种疯狂的需求冲击，我真的看不到任何类似的情况。

But also, for sure, this crazy demand impulse, I just don't see anything like that.

有两个有趣的观点让我想到，我知道我们想讨论这个，所以这真是个很好的契机。

There's two interesting points that makes me think of, and I know we want to talk about this, so it's a great excuse to do so.

一个是私募市场对这一切的影响。

One is the impact of private markets on all of this.

需求在公司这一侧是一回事。

Demand is one thing on the side of the companies.

而在想投资这些公司的投资者一侧，又是另一回事。

It's another thing on the side of the equity for people that want to invest in these companies.

你看到有高达200亿美元的基金正在募集。

And you've got just incredible $20,000,000,000 funds being raised.

你看到了这种疯狂的脱节现象，这正是我想听听你看法的地方——私人市场估值，即使是后期阶段的，与公开市场估值，即使是最近上市的公司，之间的巨大差异。

You've got this crazy disconnect, which is really what I'd like to hear your view on between private market multiples, even late stage ones and public market multiples, even recently IPO ed ones.

这似乎是一种奇怪的过渡：当你从私人市场转向公开市场时，估值大幅下跌。

So it seems like this weird handoff where you go from private to public and your valuation drops by a bunch.

是什么推动了这种现象？

What's driving that?

你认为这一切将如何影响私募股权？这里的私募股权既包括后期阶段的收购，也包括早期阶段的风险投资？

Like, how do you think all of this will affect the private equity, both private equity in the sense of later stage buyouts, but also early stage venture?

你看到这里正在发生什么？

What do you see unfolding here?

因为现在确实存在这种奇怪的不匹配。

Because it does seem to be this weird mismatch right now.

所以我认为，从宏观层面来看，正在发生的是：如果你是资金配置者，你希望获得最高夏普比率的资产。

So I think at a high level, what is happening so if you're an allocator, you want the highest Sharpe ratio asset you can have.

长期以来，什么是高夏普比率的资产？

What's been a high Sharpe ratio asset for a long time now?

任何私募资产。

Anything private.

原因是它们不按市价估值。

And the reason is is because they don't mark to market.

我向你保证，私募股权基金、风险投资基金以及其他所有机构，都会按市价估值。

Like, I promise you, private equity funds and venture capital funds and everybody else, mark to market.

你知道吗？

You know?

换句话说，如果每一家风险投资公司都坐下来，拿出一份报告，说：这是你的报表。

In other words, like, if every venture capital firm sit down a thing, be like, Here's your statement.

因为这些中型科技公司的估值倍数下降了60%，我们自身也下调了40%，我们已经改变了很多东西。

Because the multiples of these mid cap tech companies are down 60%, we're marking ourselves down 40%, we've changed a lot of things.

这种情况不会发生。

That's not going to happen.

所以过去曾存在流动性溢价。

And so what has happened is it used to be there was a liquidity premium.

现在则出现了非流动性溢价。

Now there is an illiquidity premium.

这甚至体现在单个基金层面。

This is even at individual fund level.

明白吗？

Know?

如果你是一个共同基金，身处一个大型复杂机构中，能接触到私募资产，并且你5%的资产是按月或按季度估值的私募，那么你的夏普比率就会高于那些只投资公开市场的竞争对手。

If, like, you're a mutual fund and you're in a big complex and you get access to privates and you have 5% of your assets and privates that are marked monthly or quarterly, you're going to have a higher Sharpe ratio than your competitors who only do publics.

确实存在非流动性溢价，我并不是说这很理性，但这是一个事实，而且我看不出它会改变。

There is for sure, and I'm not saying it's rational, there is an illiquidity premium, and that is just a fact, and I don't see it changing.

这是整个系统所希望的。

It's what the entire system wants.

整个系统希望减少估值频率。

The entire system wants less frequent marks.

整个系统的储蓄和养老需求需要越来越大的私募市场来配置资本。

The entire system savings and retirement needs bigger and bigger private markets in which to deploy capital.

你可以获得这些收益，它们很平稳，内部收益率达到30%。

You can get these they are smooth, 30% IRRs.

所以我认为这种流动性溢价将长期存在。

So I think this liquidity premium is here to stay.

我不认为世界会回到过去。

I don't think the world is going back.

有两点。

Two things.

第一，当你变得流动性时，流动性溢价就会消失，但你最终必须变得具有流动性。

One, illiquidity premium goes away when you become liquid, but you do have to eventually become liquid.

因此，这就是这一切的最终制约因素。

So that is the ultimate governor on this.

公开市场的倍数才是最终的现实。

Public multiples are the ultimate reality.

最终，如果公开市场的疲软或强劲持续六到九个月，正如我向你保证的，只要这种疲软持续，私人市场的估值就会降温。

And at the end of the day, if public market weakness or strength persists for six to nine months, like I promise you, this weakness persists, private valuations will cool off.

它们就是会降温。

They just will.

我从事风险投资已经二十年了。

I've been doing venture for twenty years.

这种关系非常稳定。

This is a very consistent relationship.

顺便说一句，当公开市场开始复苏时，风险投资的估值也需要一段时间才能恢复。

And by the way, and then when public start to recover, it'll take the venture valuations a while to recover.

事实上，即使在2020年，也存在许多绝佳的投资机会，你可以在2020年3月的低点进行风险投资，因为风险投资市场在上下调整时都反应迟缓，但最终，公开市场确实需要并影响着风险投资的估值。

It's like there are all these great opportunities even into the 2020 where you could make venture investments at the March 2020 low just because the venture market had slower to adjust both up and down, but ultimately, the public market does really need and inform venture valuations.

因为怎么可能不是呢？

Because how could it not?

一个令人震惊的数据是，这些风险投资支持的公司退出方式中，战略收购与IPO的比例。

One of the stats that is staggering is the percentage of exits for these venture funded firms that are strategic acquisitions versus IPOs.

以前全是IPO。

Like, it used to be all IPOs.

现在则主要由战略收购主导，这使得情况更加不透明。

Now it's dominated by strategic acquisitions, which puts yet more like opacity around it somehow.

以Twilio的业务板块为例，你根本无法了解太多关于某个板块的信息，或许只有当它成为Twilio的较大板块时，你才会稍后了解到。

Twilio by segment, you never really get to learn a lot about segment or maybe much later, you know, once it's a bigger segment of Twilio.

但正是这种从上到下全面缺乏明确估值的现象，才是系统所期望的。

But it's this interesting lack of marks all the way up and down is what the system wants.

我喜欢你这样表达。

Like, I love how you put that.

这让我好奇，你认为风险投资的竞争格局将如何演变？因为我和你都大量参与私人科技投资。

It makes me wonder how you think then the competitive landscape of venture will evolve since you and I both do a lot of private technology investing.

显然，我们在这里是极其相关的利益方。

Obviously, we're super interested parties here.

鉴于这样的竞争环境，你认为这类投资者的成功或失败会由什么驱动？

What do you think will drive success or failure amongst those style of investors given this playing field?

风险投资领域正在以惊人的速度被拆解和重新架构。

The world of venture is being unbundled and re architected at an incredible rate.

这种变化正在以多种方式发生。

It's happening in many different ways.

因此，我认为会有四类市场参与者。

So I would say to me, there will be four market participants.

首先是天使投资人。

So there are angel investors.

他们非常重要，因为过去如果你是创业者或创始人，你会认真构建你的董事会，希望从这家机构获得一位市场拓展专家。

And these are really, really important because it used to be if you were an entrepreneur or a founder, you would really thoughtfully construct your board and you'd want a go to market expert from this firm.

然后你希望从另一家机构获得一位技术专家，再从另一家机构获得一位运营专家。

And then you'd want a technology expert from this firm and then maybe an ops expert from this firm.

现在你只需找三位天使投资人，让他们为你完成所有这些工作，并在一开始就介入。

Now you just get three angels, do all that for you, and you put them in at the very beginning.

我曾与一位进行A轮融资的人进行了一次令人惊叹的对话，我当时想，嗯，这确实有道理。

I had a stunning conversation with Series A raising money, and I was like, Hey, I do think it makes sense.

就像许多种子轮、A轮和B轮投资者一样，他们拥有我永远不具备的神奇技能，能帮助你打造公司。

Like a lot of seed and A and Series B investors, they have magical skills that I will never have around helping you build your company.

他们确实带来了巨大的价值。

And wow, they add a lot of value.

他回答说：‘没错，但我可以从天使投资人那里获得这些。’

He's like, Yeah, but I can get that from angels.

换句话说，他的观点是：我可以从任何我想找的人那里拿钱。

However much you get In other words, his point was, I can take money from whoever I want.

我不需要一个能提供增值服务的投资人，因为我已经从我的天使投资人那里获得了价值。

I do not need a value add investor because I get value from my angels.

他们非常认真地考虑过这些天使投资人。

They thought about them very carefully.

这真的非常有趣。

So that was really, really interesting.

天使投资人是这一领域非常重要且将持续存在的部分。

Angels are a big, big part of this and here to stay.

然后我认为还有专注于种子轮到B轮的专业机构。

And then I think there's the seed through B specialists.

我认为，这些是许多优秀的传统风投公司。

These are, I think, a lot of the great traditional firms.

在这里参与，你真的必须提供价值。

And here, man, to play here, you really do have to add value.

你必须提供价值，才能在这里建立品牌。

You must add value to have a franchise here.

但在我眼中，那些种子轮到B轮的机构只是认证机构。

But those seed through B firms, in my mind, they're just credentializing agents.

他们提供价值，但也为公司和创始人向后续投资者提供背书。

They add value, but they also credentialize the company and the founder for later investors.

他们就像哈佛大学。

They're like Harvard.

天啊，这真是个绝佳的生意。

And, man, that's a great business.

做哈佛真好。

It's good to be Harvard.

成为这些真正主导早期阶段的投资机构非常好。

It's very good to be one of these really, really dominant early stage investors.

然后我认为，世界其他地方将是这些大型多阶段投资机构。

Then I think the rest of the world is going to be these really big multi stage investors.

我认为最终每个人都会成为跨阶段投资者。

I think ultimately everybody's going to be a crossover investor.

要与跨阶段投资者竞争，你就必须成为跨阶段投资者。

To compete with crossover investors, you must be a crossover investor.

否则，他们拥有更优越的捕鼠器。

Otherwise, they have a superior mousetrap.

对创始人来说，说‘我们永远不会退出’是一种极其强大的价值主张。

It's such a powerful value proposition to a founder to say, Hey, we're never going to sell.

我们会在你的IPO时继续买入，可能持有你二十年。

We're going to buy more on your IPO, we could own you for twenty years.

这太有力量了。

That's so powerful.

我认为，正是这种竞争压力让红杉资本——无论他们是否是最好的，至少是三大顶尖风投之一——能够保持现状。

I think this competitive pressure is why Sequoia, arguably, whether they're the best, they're one of the three best venture investors, is okay.

他们实际上转型为了一家交叉型投资者。

They convert it to effectively being a crossover investor.

然后我认为，如果你不属于这三大类别之一，将会变得极其、极其、极其困难。

And then I think if you're not one of those three buckets, it's going to be very, very, very hard.

我觉得这很酷，因为这也会为有限合伙人（LPs）带来更好的结构。

And I think this is cool because it is going to put By the way, I think that this is a better setup for LPs too.

我应该说明，我接下来要说的所有内容都来自我的朋友和导师安东尼奥·格拉西亚斯，我真心希望他能尽快来参加我们的播客。

And I should just say everything I'm about to say comes from my friend and mentor, Antonio Gracias, who I really hope goes on the podcast sometime soon.

我们已经安排好了。

And we're scheduled.

谢谢。

Thank you.

是的。

Yes.

但安东尼奥是我最重要的导师之一。

But Antonio is one of my most important mentors.

我从他身上学到了很多。

I've learned so much from him.

任何与他打过交道的人都会说，世界上没有任何一位风险投资人能像他和他的公司那样为投资的公司创造更多价值。

Nobody who has dealt with him would say that there is a venture investor in the world who is more value added than what he and his firm do for their companies.

我现在参与的公司与他们有合作，也因此逐渐了解了他。

I've been involved in companies with them now and it's gotten to know him.

确实如此。

It's true.

也许其他人也能带来同等的价值。

Maybe people add as much value.

没有人能带来更多的价值。

No one adds more value.

是的。

Yeah.

但他的整个风险投资策略——我毫不客气地借鉴了其中很多——都围绕着多阶段投资带来的深远优势。

But his whole strategy, a lot of his strategy for venture, which I've ruthlessly copied, is around the idea of their profound advantages towards being multistage.

我的一个原则是，在我了解一家公司六个月之前，我不会做出重大投资。

One of my rules is I don't make a significant investment until I've known a company for six months.

为什么呢？

Why would you?

我是一个跨阶段投资者，所以我关注的是内部收益率，而不是摩尔定律。

I'm a crossover investor, so I'm IRR focused, not Moik focused.

这是一大区别和优势，能够逐项与摩尔定律进行对比。

And that's a really big distinction and advantage being able to R by R against Moik.

因此，我认为这些多阶段基金的策略，以及你在市场上看到的，我觉得非常合乎逻辑，就是：嘿。

And so I do think a lot of these multistage firms, the strategy they have and you're seeing in the market, which I think is imminently logical, is, hey.

你早期投入小额资金。

You write these small checks early.

顺便说一下，帕特里克，这正是我们多次与你共同投资的原因。

By the way, Patrick, it's a reason we've co invested with you several times.

我们一直在观察这家公司。

We watch the company.

如果公司表现良好，我认为对于这种多阶段交叉型基金来说，一个成功标准是：任何进入你投资范围且表现良好的公司，都不应再进入公开市场。

And if they execute, I do think a success metric for kind of a multistage crossover firm becomes no company that is in your portfolio that is performing well should ever go out to the market once they come into your strike zone.

你掌握了他们的数据。

You're getting their metrics.

你对他们了如指掌。

You have perfect information on them.

你直接与首席执行官沟通。

You're talking to the CEO.

你想要什么指标都能拿到。

You're getting whatever metrics you want.

然后你就说，哇。

And then it's like, oh, wow.

我们可以看出他们还剩下十二到十五个月的现金。

We could tell they've got twelve to fifteen months of cash left.

是的。

Yeah.

现在是提前介入的好时机。

This is a good time to preempt.

砰。

Boom.

我认为这就是世界正在演变的方式。

And I think that is the way the world is devolving.

所以天使投资人、种子轮到B轮的专业投资者能提供大量价值，一些运营增长型公司也能像早期阶段公司那样提供同等价值，而这些大型多阶段全能型公司和一些传统风投机构正在成为这样的角色。

So angels, seed through b specialists who add a lot of value, some operational growth firms that add as much value as those early stage firms, and then these big multi stage bulge bracket firms and some traditional venture firms are becoming those.