英伟达200亿美元收购Groq、Waymo估值1000亿美元与预测市场的风险 | 2025年12月29日

欢迎各位来到Informations TITV。

Welcome everyone to the Informations TITV.

我的名字是阿卡什·帕什里查。

My name is Akash Pashricha.

今天是12月29日，星期一。

It is Monday, December 29.

希望你们度过了一个愉快的假期，并对即将到来的2026年充满期待——距离新年只剩三天了，但在那之前我们还有几期节目要与大家分享。

Hope you all had a wonderful holiday and are excited for 2026, which is just three days away, but we have got a few more shows for you before then.

今天，我们将深入解析圣诞节前夕公布的英伟达以200亿美元收购初创芯片公司Grok的交易。

Today, we are breaking down the $20,000,000,000 NVIDIA Grok deal announced late on Christmas Eve.

随后，我们将深入分析自动驾驶汽车领域，探讨Waymo在寻求估值高达1000亿美元的融资轮次时的战略布局；最后，我们将回顾预测市场取得突破性进展的一年，并探讨该领域在新的一年可能面临的挑战。

We'll then dig into some new analysis on autonomous vehicles and how Waymo is positioning itself as it explores a funding round valuing the company at $100,000,000,000 and we will wrap with a look at the breakout year that prediction markets have had and the challenges that the sector could now face going into the new year.

这将是一期精彩的节目，让我们马上开始。

It's going to be a fun show, so let's get right on into things.

我们仍在梳理英伟达以200亿美元向初创芯片公司Grok授权的技术交易，对整个芯片行业意味着什么。

We are still piecing together what NVIDIA's $20,000,000,000 licensing deal with chip startup Grok means for the broader chip industry at large.

如果你错过了，这是一笔在圣诞节前夜假期前夕宣布的交易。

In case you missed it, this was a deal that was announced late on Christmas Eve right before the holiday.

为了全面解析这笔交易，我想邀请我们的财经副局长科里·温伯格。

And to break it all down, I want to bring on Corey Weinberg, our Deputy Bureau Chief of Finance.

科里，欢迎回到节目。

Corey, welcome back to the show.

很高兴你在这里。

It's great to have you here.

嘿，早上好，阿卡什。

Hey, good morning, Akash.

颜色匹配。

Matching colors.

我们还沉浸在节日气氛中，我的朋友。

We are still in the holiday spirit, my friend.

我觉得我们看起来不错。

I think we're looking good.

我们必须保持这种风格一直到2026年。

We got to keep up this look going into 2026.

我喜欢这个。

I like it.

我喜欢这个。

I like it.

你看，詹森·黄显然一直工作到假期前夕。

Well, so look, Jensen Huang was clearly working right up until the holiday.

就在圣诞节前夕，他们宣布了与Grok的这笔交易。

There was this deal with Grok that was announced right on Christmas Eve.

我想听听你的分析，但先给我们简单回顾一下Grok的故事，比如这家公司的历史，以及它之前取得了多少进展。

I want to get your analysis of it, but just remind us a little bit about the story of Grok, you know, the company's history, how much traction the business was getting.

是的。

Yeah.

首先，阿卡什，这笔交易我觉得我们在圣诞节前夕的报道中已经提过了。

Well, first of all, Akashi, this deal was I think we put it in our story on Christmas Eve.

这真是个惊人的消息。

It was a stunner.

没人料到会发生这种情况。

No one really saw this coming.

我2025年的 bingo 卡上可没这一项。

I did not have this on my twenty twenty five bingo card.

关于Grok的问题，是的，让我们回溯一下，这家公司自2016年就已存在。

The question around Grok and and, yes, to take us back, this company has been around since 2016.

其联合创始人之一是乔纳森·罗斯，他在参与谷歌TPU（张量处理单元）团队后创立了Grok。

Among its cofounders were Jonathan Ross, who started Grok after being part of the team at Google to create the TPU, the tensor processing unit.

早在生成式AI热潮和ChatGPT出现之前，Grok的构想就是打造一款超高速芯片，最终超越英伟达。

And the idea with Grok even back before the Gen AI craze and, you know, before the advent of ChatGPT was they were going to build a chip that was going to be super fast and was going to beat NVIDIA eventually at its own game.

一旦ChatGPT兴起，这家公司就开始向外界传达一个信息：它的芯片在推理计算方面表现卓越。

And once ChatGPT took hold, the company really was sort of taking to the world this message that its chip was really good at inference compute.

它在运行AI模型方面非常出色。

It was really good at running the AI models.

如果NVIDIA的GPU在训练和推理方面都很出色，那么Grok则专精于推理。

If NVIDIA's GPU was good at both training and inference, well, Grok was gonna be specialized for just inference.

因此，它为这一愿景筹集了大量风险资本，但这笔交易令人震惊，因为没人真觉得Grok在与NVIDIA的竞争中取得了显著进展。

And so it raised a bunch of venture capital for this vision, but this deal was a stunner because no one really thought that Grok was really gaining that much traction against NVIDIA.

它一直是个边缘的新兴竞争者，努力追赶，但从未真正构成威胁。

Like, it was always this sort of upstart competitor off to the side, giving it giving it a good go.

我们谁都没料到它会迎来一笔200亿美元的交易。

None of us saw a $20,000,000,000 deal coming towards it.

但他们确实已经在市场上推出了芯片，并且至少获得了一些业务，我记得我们曾写过关于它在中东地区从当地客户那里获得收入的故事。

But they did have a chip out in the market that was getting at least some business from I I remember we were writing stories about the revenue it was generating in The Middle East and stuff like that from customers there.

是的，它最大的一笔交易发生在2025年初。

Yeah, its biggest deal came back at the start of 2025.

它是在今年启动的。

It started this year.

它获得了沙特阿拉伯王国15亿美元的承诺投资，这正是当时‘主权AI’成为最热门术语之一的时候，而这一合作至今仍在继续。

It booked a $1,500,000,000 commitment from The Kingdom Of Saudi Arabia to essentially this was when, and it's still going, you know, the sovereign AI was like the hottest one one of the hottest phrases.

各国将建设自己的集群来运行AI模型。

Countries are going to be building their own clusters to run to run AI models.

而Grok确实在那里站稳了脚跟。

And Grok really got, you know, a toehold there.

这是它迄今为止最大的一笔交易。

This was its biggest deal by far.

但当你在与NVIDIA这样的巨头竞争时，相比之下，它就像海洋中的一条小鱼。

But when you're talking about going against a giant like NVIDIA, I mean, is, I mean, it's a tiny minnow in an ocean compared to NVIDIA's mark.

对。

Right.

所以好吧。

So okay.

那我们开始吧。

Let's get to it then.

你认为NVIDIA为什么要做这笔交易？

Why do you think NVIDIA did the deal?

这就是所有人仍在努力维持的局面。

That's what everyone's still trying to keep together.

我的意思是，我们在报道中发现的一些事情，以及其他一些事情，让我对芯片上的内存有了比以往更多的了解。

I mean, some things that have come up on in our reporting so far as well as some others, you know, is I'm learning more about on chip memory than I really ever have in the past.

这些Grok芯片的设计理念是，它们之所以更快，部分原因在于它们在同一芯片上同时处理内存和计算。

The idea with these Grok chips are they're supposed to be faster in part because they're handling both memory and compute on the same chip.

内存通常是英伟达看到价格飙升的商品，整个市场也是如此。

Memory is often a commodity that Nvidia has seen spiking prices for, and so is the rest of the market.

因此，这里的思路是，如果将Grok TPU放在GPU旁边，它在扩展时，尤其是在推理市场中，就能获得某种内存优势。

And so the idea here is that if it puts the GROP TPU next to the GPU, it can have, you know, sort of a memory advantage as it looks to expand, particularly the inference market.

这里更广泛的结论是，过去一年我们听到了很多关于市场从‘这全都是关于训练’转向‘现在我们必须投入大量资源来运行AI，因为越来越多的人在使用它’的说法。

That's kind of the broader takeaway here is that we've heard a lot over the past year about how the market for for Sambeys and chips are are gonna go from, you know, this is all about training.

这全都是为了构建这些集群来训练这些大语言模型，而现在我们必须投入大量资源来运行AI，因为越来越多的人在使用它。

This is all about building these clusters to train these these LLMs to now we got to spend a lot to run this AI as more and more people use it.

所以归根结底，人们将这笔交易解读为：英伟达希望加强其在推理市场的立足点，并提升该领域的技术。

And so, you know, at the end of the day, people are talking about this deal as Nvidia wants to strengthen its toehold in sort of the inference market and improve its technology there.

而且，Grok 也能帮助应对内存价格的上涨。

And also it's seeing the spike in memory prices that Grok could help with.

所以这笔交易的结构是我们以前见过的。

So the structure of this deal is one that we've seen before.

这是一种奇怪的收购人才型交易，一再出现。

It's kind of this funny acquihire structure that has come up time and time again.

不过，我认为这可能是迄今为止最大的一笔，我不知道具体大了多少倍，但据我回忆，像 Character AI 这样的交易，都是几十亿美元级别的。

Although this is, I think, the biggest by, you know, I don't know how many factors, but you know, those deals, from what I recall, I mean, the Character AI deal, these were all a few billion dollars.

我不知道你是否认为，随着这些交易变得越来越大，监管机构会不会开始关注它们？

You know, I wonder if you think there's a chance that regulators at all start to look at this anymore now that they're becoming so big.

还是说，目前监管机构的态度就是：随便吧，我们就让这些交易继续发生。

Or is it just the way that regulators are right now that it's like whatever, you know, we'll just let them keep happening.

我个人不认为会这样。

I personally don't expect it.

我的意思是，到目前为止，我还没有就反垄断方面做太多额外的报道。

I mean, I haven't done a ton of extra reporting on the antitrust side as of yet.

但我的意思是，如果监管机构真想更仔细地审查这种结构，到目前为止他们已经有大量案例可以着手了。

But, I mean, I think if regulators had wanted to look more closely at this structure, they would have had plenty of examples to do so thus far.

而且我们迄今为止几乎没有听到任何风声，尤其是在特朗普政府时期。

And we really haven't gotten much of a whiff of it, especially in the Trump administration.

在拜登政府时期，微软曾进行过一笔这样的交易，当时曾有一些关于此类交易的议论。

There were some rumblings about the kind of deal that Microsoft had done under the Biden administration where you saw some rumblings about this.

但你看，我认为英伟达、黄仁勋，以及我们应当指出，Grok的一些投资者往往在特朗普政府时期达成了他们的目标，尤其是围绕着‘美国在人工智能领域的优势具有国家利益’这一理念，而英伟达正是其中的关键部分。

But look, I think Nvidia, Jensen Huang, and we should note some of Grok's own investors have gotten have often gotten what they wanted out of the Trump administration, particularly around this idea of there is a sovereign interest in The US's AI advantage, and NVIDIA is a huge part of that.

所以，我相信他们的律师在设计这笔交易时，已经考虑到至少能让监管机构在采取行动前有所犹豫。

So, you know, I'm sure their lawyers structured this deal in a way that would at least give regulators pause before taking action.

我想问你一个关于这些AI芯片初创公司的问题。

Let me ask you a question about these AI chip startups broadly.

我的意思是，我们在《The Information》有一个专门追踪这些AI芯片初创公司的数据库，我们跟踪了将近两打这样的公司。

I mean, we have an entire database dedicated to these AI chip startups here at The Information, and we track, I think it's nearly two dozen of them.

我们还听到了一些风声，称英特尔正在考虑收购另一家芯片公司SambaNova。

And we've seen rumblings of Intel looking at another chip company, SambaNova.

我们显然对Cerebras公司进行了非常密切的报道。

We obviously you have reported very closely on Cerabros' or Cerebras.

我总是不确定该怎样正确发音‘mystery’。

I never know what the right way to pronounce mystery.

Cerebris，cerebris。

Cerebris, cerebris.

我的意思是，他们走向公开市场的历程。

I mean, you know, their journey to the public markets.

科里，我更想问你一个更广泛的问题：当这些芯片初创公司开始被收购或被吸收时，这究竟是这些AI芯片初创公司的终点，还是它们与AMD、NVIDIA以及谷歌的TPU竞争的终结？

I mean, the broader question I have for you, Corey, is as these chip startups appear to start to get acquired or aqua hired, I mean, is this the end or the beginning of the end of these AI chip startups broadly and competing with the AMD's and the Nvidia's and the Googles with the TPUs?

我的意思是，它们从来就不仅仅是一个希望。

I mean, they never really were a just a hope.

这只是一个希望。

It's just a hope.

但是，

But,

我的意思是，如果你是个风险投资人，你并不需要这家公司成为下一个英伟达。

I mean, if you're a if you're a venture capitalist, you don't need this company to become the next NVIDIA.

这几乎更好，或者说，通过被英伟达、AMD或阿尔法贝塔收购，能更快地套现。

It's almost better, or at least it's faster money to get paid out pretty quickly in a in an acquisition by an NVIDIA or an AMD or an Alphabet.

别忘了，OpenAI正在试图自研芯片，ARM和软银也在努力，有很多人想进入这个领域，希望通过自研芯片或建立其他合作关系来减少对英伟达的依赖，从而削弱其主导地位。

And, like, let's not forget OpenAI trying to build their own chip, ARM and SoftBank trying to build like, there are a lot of people trying to get in the game of we want to reduce our reliance on NVIDIA by either building our own chip or coming up with, you know, sort of another partnership that, you know, where we can offset that dominance.

所以，在某种程度上，我不希望在2026年还留在AI基础设施行业。

And so in some ways, I wouldn't want to be in the AI infrastructure business still, 2026.

但这里确实是个好地方。

It's a good place to be.

另一个最大的收购兼雇佣型交易是Scale AI。

The other sort of biggest of these big acquihire type deals was Scale AI.

它也是AI基础设施和数据标注领域的公司。

That's also AI infrastructure and data labeling.

你知道，这里确实是个好地方。

You know, it's a good place to be.

人们仍在大量投资于此。

People are still spending a ton of money on this.

无论是在SambaNova、Etched还是Cerebris这些公司，至少在短期内，它们很可能都会取得不错的成果。

And whether it's, you know, SambaNova or Etched or, you know, Cerebris, like, you know, these guys are probably going to have, at least in the short term, pretty good outcomes.

长期来看，谁知道呢？

Long term, know, who's to say?

很好。

Great.

科里，感谢你今天来做客。

Well, Corey, I want to thank you for coming on.

这位是科里·温伯格，我们《信息》报的财经副总监。

That is Corey Weinberg, our Deputy Bureau Chief of Finance here at The information.

好的。

Okay.

如需更多关于英伟达、Grok以及Waymo的报道，我请来密切关注这一领域的保罗·米克斯。

For more coverage on all things NVIDIA and Grok, and also Waymo, I want to bring on Paul Meeks, who follows this space very closely.

保罗是自由资本市场的科技研究主管。

Paul is the head of technology research at Freedom Capital Markets.

保罗，早上好。

Paul, good morning to you.

很高兴见到你。

It's great to see you.

节日快乐，我的朋友。

Happy holidays, my friend.

关于这个NVIDIA与Grok的交易，我想听听你的看法。

So this Nvidia Grok deal, I mean, I want to get your thoughts on it.

我想先问你一个更广泛的问题：Grok这是认输了吗？是不是觉得根本不可能在很大程度上与NVIDIA竞争？

The broader question I want to ask you to start with is, is this Grok throwing in the towel and saying, Hey, there was no way that we were ever going to compete with Nvidia in a large part.

不如直接加入他们。

Let's just join them.

你是这么看待这笔交易的吗？

Is that how you see this deal?

我其实认为他们确实有真本事，因为英伟达的那些人，虽然这话不针对你，但他们两个人加起来都比不上那些人聪明。

I actually think that they really have some game, because the folks at Nvidia, no offense to you, but they're smarter than both of us put together.

非常精明，运营能力很强。

Super savvy, good operators.

如果他们愿意花200亿美元购买一项非独家许可，而不是独家许可，再加上部分员工，而不是全部员工，我认为他们确实认为Grok在推理领域可能对他们的业务构成威胁。

If they're going to pay $20,000,000,000 for a non exclusive license, not an exclusive license, and some personnel, not all personnel, I think that they really thought that Grok could be a threat in inference to their franchise.

布莱恩，你今天早上发布了这份报告，你觉得英伟达是趁机行事，还是只是走投无路？

Brian, you put out this note this morning, is NVIDIA being opportunistic or just desperate?

你对这个问题的答案怎么看？

What do you think the answer is to that question?

我认为他们是在趁机行事。

I think they're being opportunistic.

他们有这个灵活性，对吧？

They have the flexibility, right?

这家公司在本财年的过去九个月里，产生了600亿美元的自由现金流，而这还不包括他们资产负债表上原本就有的500亿美元。

This is a company that in the last nine months of this fiscal year has generated $60,000,000,000 in free cash flow, and that's on top of $50,000,000,000 that they already have on their balance sheet.

因此，如果他们相信——我也确实认为这是事实——我们现在正迅速从训练转向推理，而且我们还面临着电力方面的真正瓶颈。

And so if they believe, and I do believe that this is true, that we have a pretty rapid transition at this point from training to inference, and we also, of course, have a real bottleneck with power.

如果Grok的这些LPU真的能提升速度——我认为这只是较小的部分——但如果它们能在这片市场中实现十倍的能效提升，那么这将为我们进入2026年解决一些真正重要的问题。

And if these LPUs from Grok can actually increase speed, which I think is the smaller part of it, but if they can really bring 10 times the power efficiency in this market, That solves some real important questions as we go into 2026.

所以我喜欢这个。

So I like it.

对。

Right.

我想和你聊聊我们上一个环节结束时提到的那个问题，也就是更广泛的AI芯片初创公司整体格局。

I want to talk to you a little bit about the question that we ended the last segment with, which is the broader AI chip startup landscape at large.

我的意思是，这些公司已经涌现了数十家，而且那些获得大量融资的公司尤其如此。

I mean, these companies have popped up in the dozens in terms of the ones that are getting a lot of funding.

它们将继续获得大量融资。

They will continue to get a lot of funding.

你认为这些初创公司的最终命运会如何？

What do you think is the fate of these startups ultimately?

你认为它们最终都会被收购吗？

Do you think they are all destined for acquisition?

你认为它们能在这个市场中为自己争取到一小部分份额吗？

Do you think they will carve out a small part of this market for themselves?

我的意思是，你觉得这种情况会如何发展？

I mean, how do you see that playing out?

非常好的问题。

Excellent question.

我是这样看的，我觉得自己有点反主流观点，我认为AI基础设施支出不仅会在2026年之前持续火热，之后也会如此。

The way I look at this, and I think I'm a bit of a contrarian, I expect AI infrastructure spending to be hot and heavy, not just through 2026.

我认为这一点几乎可以确定。

I think that's pretty much assured.

当我们的公司，尤其是那些占累计基础设施建设支出90%的超大规模云服务商，在公布下一季度财报并给出2026年指引时，我认为它们会再次提高资本支出预算。

And when our companies, particularly the hyperscalers that spend 90% of this cumulative infrastructure build, when they announce their next quarter and give us guidance for 'twenty six, I think they're going to increase their CapEx budgets again.

因此，我认为整个基础设施领域，特别是那些精选的半导体公司，在2026年有着非常光明的前景，我不认为2027年的支出会骤然下滑。

So I think the entire infrastructure space, particularly the supporters, the select semiconductor companies, have very bright outlook for 'twenty six, and I don't expect the spending to fall off a cliff in 'twenty seven.

也许会趋于平稳，也许会稳定下来，但我认为不会大幅下滑。

Maybe it plateaus, maybe it stabilizes, but I don't think it drops much.

因此，我认为在未来两年内，这些公司有着不错的增长空间。

And so I think these companies, the least the next two years, have a nice runway.

但随着时间推移，当然总会出现泡沫，就像互联网泡沫破裂那样。

But over time, of course, there's always a bubble, just like there was in the internet bubble when it popped.

这种情况一定会发生。

That will happen.

我向你保证。

I guarantee it.

这种事情总是会发生的。

It always does.

其中一些公司将会被淘汰。

And some of these companies will be wiped out.

但我认为，至少在未来两年内，前景依然清晰。

But I think that the skating is clear for at least the next two years.

有些人认为这已经接近尾声了。

Some people thinking it's close to the end.

我不这么看。

I don't see that.

所以，即使这些公司没有占据显著的市场份额，你认为市场仍会以足够快的速度增长，使它们依然能够成功吗？

So the market is going to continue to grow fast enough that even if these companies don't have significant market share, you think they will be able to be successful even on their own?

我认为最有前景的公司会被收购，就像Grok那样，从英伟达获得了200亿美元，当然也会有一些公司失败。

I think that the most viable ones will be swept up, like you saw with Grok, dollars 20,000,000,000 from Nvidia, and some will fail for sure.

但我确实认为，对于某些特定公司而言，仍将存在一个持续增长、目前可能尚未盈利但未来有望盈利的利基市场。

But I do think for select companies, there will continue to be a nice, strong growing, and maybe not profitable now, but profitable down the road niche for select companies.

我想问你关于另一类公司的问题，我们一直非常关注这些‘新云’公司，而这些公司的许多创始人或高管，如果你称它们为‘新云’，他们会举手表示：‘我们不称自己为新云，我们称自己为AI专用云。’

I want to ask you about another class of companies that we have been thinking a lot about, these neo clouds, and a lot of the founders or executives of these companies, if you call them a neo cloud, they'll put up their hands and say, Well, we don't call ourselves a Neo Cloud, we call ourselves an AI specialized cloud.

这是AI基础设施故事中另一个让我着迷的领域。

This is another segment of the AI infrastructure story that has fascinated me.

你认为最终那些与超大规模云服务商竞争的云提供商将何去何从？

What do you think ultimately happens to those cloud providers that are competing in some ways with the hyperscalers?

你有像Crusho这样的数据中心公司。

You've got data center companies like Crusho, for example.

它们正试图扩展自己的云业务。

Mean, they're looking to build out their cloud businesses.

这和芯片故事一样吗？即市场增长足够快，还是会看到整合？

Is it the same as the chip story where the market is growing fast enough, or do we see consolidation?

也许未来会出现整合，一些新兴云公司可能会被传统超大规模云服务商甚至英伟达收购。

Maybe consolidation down the road and some of these neo clouds might actually be acquired by traditional hyperscalers or even Nvidia.

例如，我对CoreWeave持有买入评级。

Like for example, I have a buy rating on Coreweave.

CoreWeave是这些公司中最具有代表性的。

CoreWeave is the most emblematic of these companies.

我的目标价是100美元。

I have a $100 price target.

关键在于供应，而不是需求。

The key here is supply, not demand.

我认为，正如我之前所言，未来几年甚至更长时间内，这些公司的需求将非常强劲。

I think demand, echoing my previous comments, is going to be robust for these players for the next couple of years, maybe even beyond.

与此同时，它们面临的问题是供应，而不是需求。

And in the meantime, their issue, their gating factor is supply, not demand.

所以我喜欢这个方向。

And so I like it.

我认为它们被误解了。

I think they're misperceived.

其中一个

One of

CoreWeave的问题在于其资产负债表较弱，必须持续筹集资金，而且它已经负债累累，未来需要逐步降低其加权平均资本成本。

the problems with Corweave, it does have a weak balance sheet, and it is going to have to continue to raise capital, and it's already pretty heavily levered, and it's going to have to drive down its weighted average cost of capital over time.

我认为它们能做到。

I think that they will.

但还有其他公司不像它这样高杠杆，比如Nebius，还有Applied Digital，它实际上把CoreWeave当作大客户。

But then there's other companies that are not necessarily as levered like Nebius and maybe Applied Digital, which actually has CoreWeave as a big customer.

我认为它们的前景相当不错。

I think that they have pretty good prospects.

而目前，市场并没有按照它们拥有良好前景来估值。

And right now, the market is not valuing them like they have pretty good prospects.

而且again，这是一个供应问题，所以它们必须执行到位。

And again, it's a supply issue, so they have to execute.

这不是需求问题。

It's not a demand issue.

我认为我们在这一点上会保持清晰很长一段时间。

I think we're going be clean on that for quite a while.

在你离开之前，我再问你一个问题，关于Waymo。

Let me ask you one more question before you go about Waymo.

Waymo是一家我们报道过正在寻求以1000亿美元估值融资的公司，这是一个惊人的估值，不过，我的意思是，自动驾驶汽车似乎已经到来，至少在我们观众所在的许多主要城市是这样。

Waymo is a company that we have reported as seeking to raise money at a $100,000,000,000 valuation, which is a whopping price tag, although, I mean, autonomous vehicles, they seem to be here, at least in many of the major cities that our audience is in.

你对这个估值怎么看？

What do you think of the price tag?

你知道吗，我希望他们能成功，毕竟我也是股东，当然，从另一端来看，他们有他们的TPU。

You know, I hope they get it, because I'm a shareholder, of course, on the other side of the house, you know, with their TPUs.

除了核心的数字广告业务外，他们还有一个非常出色的自动驾驶系统。

They also have a very nice driver, in addition to the core digital advertising business.

我一直认为，尽管埃隆·马斯克夸夸其谈，但Waymo实际上领先很多。

I have always felt that despite Elon Musk's bluster, that Waymo was significantly ahead.

尽管Waymo最近有一些广为人知的失误，对吧？

And even though Waymo has had some well chronicled slip ups recently, right?

当旧金山停电时，他们的汽车被困在了街上。

Their cars got stuck in the street in San Francisco when the power went out.

这可是个丑闻。

That's a black eye.

但我认为他们还是能拿到一千亿美元的估值，因为特斯拉还落后一截。

But I think they can get the $100,000,000,000 because Tesla is further down the pike.

我不觉得他们的前景有那么光明。

I don't think they have as bright a prospects.

而随着埃隆·马斯克大谈太空，他们的估值将会实现。

And with Elon Musk talking up to space, they'll get their valuation.

所以我认为Waymo会获得它的两倍价值。

So I think Waymo will get its twos.

所以我觉得这很棒。

So I think it's great.

我希望他们能成功。

I hope they get it.

如果他们能以一千亿美元的估值达成交易，那将是自动驾驶领域的一大荣耀。

That would be a nice feather in the cap for autonomous vehicles if they could do a deal with a $100,000,000,000 valuation.

好了，保罗，感谢你前来做客。

Well, Paul, I want to thank you for coming on.

这位是自由资本市场的科技研究主管保罗·米克斯，来自TI TV。

That is Paul Meeks, Head of Technology Research at Freedom Capital Markets here on TI TV.

好的。

Okay.

2026年最大的新闻之一将是自动驾驶汽车，正如我们刚才讨论的，其中一个关键点将是它们能继续获得多少关注和资金。

One of the biggest stories of 2026 is going to be autonomous vehicles, as we just talked about, and one of the big points is going to be how much traction and funding they can continue getting.

在《信息》杂志，我们当然一直在关注这个故事。

Here at The Information, we have, of course, been following that story.

几周前，我们报道了Waymo正在讨论以1000亿美元的估值融资。

We reported a few weeks ago that Waymo was in discussions to raise money at $100,000,000,000 valuation.

我们的金融分析专栏作家阿妮塔·拉马斯瓦米在过去几周里与她的多个消息源进行了大量交流，以了解我们该如何看待这一估值。

Our financial analysis columnist Anita Ramaswamy has been talking a ton to her sources over the past few weeks to understand how we should be thinking about that price tag.

她今天发表了一篇关于此事的精彩文章，我想请她上来为我们详细谈谈她的发现。

She has a good piece on that out today, and I want her to bring on to tell us more about what she found.

阿妮塔，欢迎再次做客我们的节目。

Anita, welcome back to the show.

很高兴你能来到这里。

It's great to have you here.

很高兴回来，阿卡什。

Great to be back, Akash.

今天节目中所有人都穿着红色。

Everyone's wearing red today on the show.

大家还沉浸在节日气氛中。

Everyone is still in the holiday spirit.

顺便说一下，这是红色。

This is red, for the record.

明白吗？

Okay?

我知道它看起来像紫色，但我还是决定称它为红色。

I know it's looking like purple, but I'm choosing to call it red.

好的。

Okay.

请你详细解释一下你对Waymo追求的1000亿美元估值所做的数据分析，并将它与其他网约车公司、自动驾驶车辆和相关企业进行对比。

Walk us through the numbers that you crunched on the $100,000,000,000 figure that Waymo is going after, and I want you to put it into context with all the other ride hailing companies, autonomous vehicles, franchises that are out there.

当然可以。

Yeah, of course.

我们没有Waymo的完整财务数据。

So we don't have full financials for Waymo.

它是Alphabet的一部分。

It's a part of Alphabet.

他们没有单独披露财务结果。

They don't break out results separately.

但根据我们所知以及《The Information》此前的报道，Waymo正在洽谈以约1000亿美元的估值融资。

But based on what we know and what we've reported here at The Information, Waymo is in talks to raise around at $100,000,000,000 valuation.

这比他们上一轮融资的估值高出一倍多。

That is a little over double what they last raised at.

而他们最近的年化收入约为3.5亿美元，这意味着其年化收入倍数超过200倍。

And their annualized revenue as of recently is around $350,000,000 So that amounts to a multiple on their annualized revenue of over 200 times.

与其他网约车初创公司相比，这个倍数非常高。

In comparison to other ride hailing startups, that's really high.

好的。

Okay.

所以200倍的收入是一个很高的倍数。

So 200 times revenue is a big multiple.

优步和Lyft目前的交易估值大约是多少？

Uber and Lyft, what are they trading at around now?

优步和Lyft的交易估值大约是其未来收入的三倍或低于三倍。

Uber and Lyft are trading at around three times or below three times their forward revenue.

好的。

Okay.

那么，我们认为Waymo能实现这一估值吗？

So do we think that Waymo can grow into this value?

这是个不同的业务。

It's a different business.

我明白。

I get it.

好的。

Okay.

不同的技术。

Different technology.

我们处于早期阶段。

We're early.

但我的意思是，你今天早上在我们年终图表中展示了一张显示Waymo获得进展的图表。

But I mean, you put a chart out this morning in our year end charts that showed traction that Waymo has been getting.

那是一张所有不同地点的地图。

It was a map of all the different locations.

你认为它能增长到那个200倍的估值吗？

Do you think it can grow into that 200x valuation?

我认为答案是肯定的，阿卡什，但我认为这需要很长一段时间才能实现。

I think the answer is yes, Akash, but I think the answer is yes over a really long period of time.

而且我认为投资者正在着眼于更长期的未来。

And I think investors are looking to that longer term future.

如果你预测未来几年，考虑到2030年，今年早些时候我们看到摩根士丹利分析师的一些预估，他们预测Waymo可能实现约25亿美元的收入。

If you forecast out a couple of years and you think about 2030, we saw some estimates earlier this year from Morgan Stanley analysts who are projecting that Waymo could make about $2,500,000,000 in revenue.

如果他们能达到这样的收入水平，他们的估值倍数将接近40倍。

And if they were to be making that level of revenue, their multiple would be a little closer to 40 times.

这仍然很高，但与公开市场上的其他高增长软件公司、人工智能公司相比，并不算过分。

That's still high, but compared to other high growth software companies, AI companies in the public markets, it's not as egregious.

我认为关键在于，Waymo正在快速增长，而我们尚未看到他们真正大规模扩张的完整一年。

And I think the key thing here is Waymo is growing really fast and we haven't seen a full year where they've really, really expanded.

2025年是他们首次突破最初三个城市的年份。

2025 was the first year that they broke out of their three initial cities.

他们又增加了两个城市。

They added two more.

他们现在真正开始扩大规模了。

They're really starting to scale now.

我还想强调的是，与Uber和Lyft不同，Waymo实际上拥有多个收入来源，他们一直提及这些来源，而不仅仅依赖于网约车业务。

And I think it's also important to remember that unlike Uber and Lyft, Waymo actually has a couple of different revenue streams that they've been talking about that they could leverage besides just ride hailing.

除了自动驾驶出租车，他们还在某些市场测试食品配送服务，这一点Uber当然也在做。

So there's the robo taxis, but they're also testing out food delivery in certain markets, which Uber does of course.

但他们也在考虑将软件授权给其他汽车制造商，这可能成为一项高利润率、高增长的收入来源，具体取决于这项技术的实际成熟程度。

But they're also thinking about licensing their software to other automotive makers, and that could potentially be a high margin, high growth source of revenue depending on how this technology actually matures.

我之前没意识到这一点。

So I didn't realize this.

Waymo现在已经在配送食物了？他们正在拓展这项业务？

Waymo is delivering food right now, they're building out that business?

他们正在拓展这项业务，并且从十月开始，他们宣布与DoorDash合作，在凤凰城推出这项服务。

They're building it out, and they're starting in October they had an announcement with DoorDash to debut that in Phoenix.

所以他们目前在凤凰城以有限的方式启动这项服务，如果成功， presumably 会扩展到其他城市。

So they're kind of starting it out in a limited way in Phoenix, and if it works, presumably they'll scale to other cities.

对。

Right.

要扩展到其他城市，显然需要获得许可和应对监管问题。

Now part of scaling to the other cities is obviously getting the permits and the regulatory angle to all this.

你在文章中提到，Waymo在监管方面实际上比其他公司领先一步。

You talked a little bit in your piece about how Waymo is actually further ahead than other companies on that regulatory story.

是的，没错。

Yeah, that's right.

我认为这将是Waymo未来巨大的优势。

I think that's going to be the huge advantage for Waymo going forward.

我认为这里的先发优势将会很强。

I think the first mover advantage here will be strong.

以纽约为例，这是一个利润丰厚的市场，也是很难构建自动驾驶技术的地方。

If you take New York, for example, that's a hugely lucrative market and a place where it's pretty hard to build autonomous vehicle technology.

而Waymo是第一家获得许可、能够在当地测试其机器人出租车的公司。

And Waymo is the first company that has secured the permits to actually test over there for their robotaxis.

我们还有其他竞争对手，比如特斯拉、亚马逊、Zoox，但它们都远远落后。

We have other competitors like Tesla, Amazon, Zoox, but they're both far behind.

我的意思是，就连Zoox，如果你看一下，目前还没有任何城市授予其载客收费的资质或许可。

I mean, even just Zoox, if you look at it, doesn't have the capability yet or permits from any city to be able to charge passengers for their rides.

因此，它们仅在加利福尼亚州的某些地方以有限的方式运营，甚至无法收取乘客费用，而Waymo早在2025年就已经在做这件事了。

So they're operating in a limited capacity in certain places in California, but they're not even able to take paid passengers, whereas Waymo has been doing that all of 2025.

那Waymo的业务面临哪些威胁呢？

What about the threats to Waymo's business?

特斯拉现在在哪些方面做得更好？

Is there anything that Tesla is better at right now?

是的。

Yes.

关键在于成本。

And the big thing is the costs.

Waymo的车辆配备了大量激光雷达传感器，这是一种非常复杂的技术，用于尝试绘制它们将要行驶的路线。

So Waymo's vehicles have a bunch of LiDAR sensors on them, and it's a really sophisticated technology that they use to try and map the routes that they're going to take.

而特斯拉和埃隆·马斯克采取了一种他们认为便宜得多的方法。

Whereas Tesla and Elon Musk has taken an approach that they say is a lot cheaper.

因此，他们仅使用普通摄像头，而不是昂贵的激光雷达传感器。

And so they are using just regular cameras instead of LIDAR sensors, which are super expensive.

因此，特斯拉制造这些车辆的成本远低于Waymo。

So Tesla is able to manufacture these vehicles for a lot less than Waymo is.

但关键在于，我们仍然尚未看到特斯拉是否真的能够实现完全自动驾驶。

But the key here is that we still have yet to see whether the Teslas can actually drive fully autonomously.

我的意思是，埃隆·马斯克称其为全自动驾驶，但实际上是需要人类驾驶员在座位上监控的全自动驾驶。

I mean, we have Elon Musk calling it full self driving, but it's full self driving with a human behind the wheel.

安妮塔，对我来说，感觉像是我们已经看到，尤其是在2025年，这么多城市已经开始部署自动驾驶汽车，而且还有更多城市即将加入。

It kind of feels to me, Anita, like given that we've seen, especially in 2025, we've seen so many of these cities go ahead with autonomous vehicles, and there's more cities coming.

对我来说，这感觉就像猫已经出了袋子，自动驾驶汽车正在到来，而目前我们看到的投入成本高达一千亿美元。

It feels a little bit to me like the cat is out of the bag, and they are coming, and it's $100,000,000,000 right now that we're seeing as a price tag.

我的意思是，我觉得这个数字在未来两三年内还会继续增长。

I mean, sort of feel like I mean, this number's going to get bigger, in my view, the next two, three years even.

这就是我的看法。

That's my thought on it.

你怎么看？

What do you think?

是的，我同意这一点。

Yeah, I would agree with that.

我认为这类服务将会迎来快速增长，而有趣的问题在于，传统的优步、Lyft等网约车服务将会发生什么变化。

I think we're going to see pretty rapid growth of these kinds of services, and I think the interesting question here is what's going to happen to ride hailing, like the traditional Ubers, the traditional Lyfts of the world.

关于这一点的一些预测也颇具意思。

And some of the projections around that are interesting too.

我看过一些分析师的预测，称优步在未来七年将增长14%，但如果你知道Waymo的乘车服务最终会蚕食优步的部分收入，那么它的增长率实际上可能降至个位数。

I was looking at projections from analysts that said Uber will grow 14% for the next seven years, but you know, if Waymo rides end up cannibalizing some of the revenue from Uber, then its growth rate will actually be closer to the single digits.

因此，我认为现在自动驾驶领域的分析师、投资者和观察者们正在思考的一个重大且价值数十亿美元的问题是：自动驾驶是否会创造出原本不会使用网约车服务的新乘客群体？

So I think the big multi billion dollar question that analysts, investors, and observers of the autonomous vehicle space are asking themselves right now is really, autonomous create new opportunities of passengers that wouldn't have otherwise taken a ride hailing ride?

还是会完全取代已经存在的网约车需求？

Or is it going to be entirely redundant to the ride hailing demand that already exists?

很好。

Great.

安妮塔，非常感谢你今天来做客。

Well, Anita, I want to thank you for coming on.

这位是安妮塔·拉马斯瓦米，我们《信息》杂志的金融分析专栏作家。

That is Anita Ramaswamy, our financial analysis columnist here at The Information.

谢谢，阿卡什。

Thanks, Akash.

好的。

Okay.

预测市场今年表现强劲。

Prediction markets have had a big year.

在我们今年的图表故事中，今天展示了一张图表，清楚地表明了这一点，同时也展现了Calci和Polymarket之间激烈的竞争。

Today in our Year in Chart story, we have a chart showing exactly that, and it also shows how fierce a rivalry Calci and Polymarket are waging against each other.

但我们的财经编辑肯·布朗发表了一篇专栏文章，认为2026年可能会给这个行业带来一些监管挑战和更广泛的困难。

But our finance editor, Ken Brown, has a column out suggesting that he thinks 2026 could bring some regulatory challenges and some challenges, broadly speaking, for the sector.

我想请肯来告诉我们，他为什么认为情况会是这样。

I want to bring on Ken to tell us more about why he thinks that's the case.

肯，欢迎再次做客我们的节目。

Ken, welcome back to the show.

很高兴见到你。

It's great to see you.

嗨，阿卡什。

Hi Akash.

你好吗？

How are you?

还不错。

Doing well.

我很期待谈谈预测市场。

I'm excited to talk about prediction markets.

你觉得会这么难吗？我的意思是，难道就没有一种可能，预测市场会持续增长，让我们迎来比2025年更盛大的一年？

Do you think it's going to be such a hard I mean, look, isn't there a way that prediction markets could just keep surging and we could have another 2025, but even more?

当然。

Sure.

当然，

Of course,

这有可能

that could be

这种情况。

the case.

预测经常出错。

Predictions are often wrong.

对吧？

Right?

好吧，事情是这样的。

Well, so here here's the here's the thing.

预测市场的增长主要来自体育博彩，而且已经蓬勃发展。

They've moved the growth in prediction markets come from sports betting, and it's boomed.

现在体育博彩的金额已经达到数十亿美元。

It has been it's been now billions of dollars in in sports betting.

问题是，体育博彩由各州监管，而各州从体育博彩中赚取大量收入。

The problem is sports betting is regulated by the states, and this and the and the states make a lot of money from sports betting.

因此，当投注转向预测市场时，他们可能会损失收入。

And so they are losing money potentially as the as the betting moves to the the prediction markets.

因此，这是一个挑战，因为你知道，纽约州每年能获得十亿美元的税收收入。

And so that's one challenge because, you know, like New York State gets a billion dollars in tax revenue.

所以他们不会安静地坐着，看着这些人把生意抢走。

So they are not gonna, you know, sit quietly while these guys, you know, take the business.

但另一点是，它增长得太快了。

But the other thing is it's such it's growing so fast.

出现了大量新的竞争者。

There's a whole bunch of new competition.

Coinbase、Robinhood、Crypto.com 等所有这些公司都将推出预测市场。

Coinbase, Robinhood, all these folks, crypto.com are all gonna start prediction markets.

特朗普家族也将推出一个预测市场。

The Trump family is gonna start a prediction market.

所以你知道，这本质上是一种大宗商品业务。

So, you know, it's it it it is kind of a commodity business.

对吧？

Right?

我的意思是，我会在选举或体育博彩上跟你对赌，不管是什么。

I mean, I bet against you over an election or a sports bet, whatever.

任何人都可以提供这种服务。

Anyone can offer it.

所以这是他们面临的另一个问题。

And so that's another problem that they face.

现在让我问你，这个问题如何与不同政党产生交集。

Now let me ask you about how this issue intersects with, different parties.

这是一个政治问题吗？

Is this a political issue?

它具有党派性吗？

Is it partisan?

双方的看法都一样吗？

Do both sides feel the same way?

人们是怎么看待这个问题的？

How do people think about it?

嗯，预测市场和体育博彩总体上是少有的两党共识领域，他们都基本不喜欢它。

Well, prediction markets and sports betting in general, it's a rare area of bipartisan agreement in which they basically don't like it.

大约47%的美国人不喜欢体育博彩。

About 47% of Americans don't like sports betting.

他们认为这没什么价值。

They don't think it adds anything.

只有大约7%的人喜欢它。

Only about 7% like it.

民主党人和共和党人实际上完全一致。

And Democrats and Republicans actually completely agree.

他们都表示，约47%的人认为体育博彩没什么好处。

They all they both say about 47% of them say sports betting is no good.

所以，他们真正谈论的只是像DraftKings这样的体育博彩公司。

And so that's just the sports betting companies like DraftKings and such that that they were really talking about.

这完全是体育博彩的一个全新领域。

This is a whole new realm of sports betting.