Ross Gerber on Elon Musk & Tesla’s 16% Revenue Growth, Google Cloud Next Conference, AI Insurance

欢迎各位收看《The Information》的TITV节目。

Welcome everyone to The Information's TITV.

我是阿卡什·帕什里卡。

My name is Akash Pasricha.

今天是4月23日，星期四。

It is Thursday, April 23.

今天第一个话题是特斯拉发布了第一季度财报，首席执行官埃隆·马斯克的语气比平时更加谨慎。

First up today, Tesla reported first quarter earnings, and CEO Elon Musk struck a more cautious tone than usual.

接下来我们会连线格伯·卡瓦苏奇的总裁兼首席执行官罗斯·格伯，听听他的看法。

We'll get a reaction from Ross Gerber, president and CEO of Gerber Kawasaki.

我们的谷歌记者正在拉斯维加斯的谷歌云峰会现场。

Our Google reporter is on the ground in Vegas at Google's Cloud Conference.

她稍后会加入节目，分享她了解到的信息，最后我们还会回顾ServiceNow的财报表现。

She will join the show to discuss what she has learned, and we'll wrap with a look at ServiceNow earnings.

财报发布后，股价出现了大幅下跌。

The stock is taking a big dip in the aftermath.

我们的一位AI记者还将深入报道她发表的一篇独家新闻，探讨大型保险公司如何应对与人工智能相关的损失。

One of our AI reporters will also dig into exclusive reporting that she published about how big insurers are reacting to AI related damages.

这期节目会很有趣，咱们这就开始吧。

It's gonna be a fun show, so let's get right on into it.

特斯拉公布了第一季度的业绩。

Tesla reported its first quarter results.

营收增长了16%。

Revenue jumped 16%.

公司在电话会议中谈到了许多长期目标。

The company talked about many of its long term ambitions on the call.

为了深入分析这些业绩数据，我想请出Gerber Kawasaki财富与投资管理的总裁兼首席执行官Ross Gerber。

To break down the results, I want to bring on Ross Gerber, president and CEO of Gerber Kawasaki Wealth and Investment Management.

Ross，早上好。

Ross, good morning.

欢迎做客。

Welcome to you.

你对昨天公布的财报怎么看？

What did you make of the results yesterday?

嗯，信息量挺大的，得慢慢理清楚。

Well, was kind of a lot to unpack.

我觉得光看数字还不够，投资者得仔细分析背后的东西，因为财报里还有很多调整项，我得等看到SEC的文件，才能明白他们是怎么得出这些比预期还好的数据的。

You know, I think there was a lot beyond the numbers that I think investors have to parse out more than just the numbers because there's still a lot of adjustments in the actual numbers that I wanna see the SEC documents to see how they got what I considered better than expected numbers than what I expected.

不过这些数字里头，好像有些地方不太对劲，咱们待会儿再细说。

But it seems like there's a little bit of funny business in the numbers that we'll get to.

但不管怎么说，特斯拉卖电动车的业绩还算不错。

But either way, Tesla had a decent quirrel selling EVs.

不过在电话会议里，各种问题一股脑儿全抛出来了，却没见任何实质性的解决方案。

But during the conference call, was kind of like a calamity of different issues being brought up and no real resolution being discussed.\

你具体是觉得哪儿不对劲？

What funny business are you concerned about here?

你看他们的利润率一直在下滑，季度末的库存又那么高，我们本来还以为他们的数据应该更差才对。

Well, know, when you look at their margins, which continue to go down and the amount of inventory that they had at the end of the quarter, You know, we suspected their numbers should be actually worse.

不过有一些汇率方面的因素，还有一些跟关税有关的事情。

But there was, like, FX things with, you know and then, like, tariff related stuff.

所以就像我刚才说的，我得看SEC的正式文件，不能光看随便哪张纸。

So I just, like as I said, I have to see SEC documents, not just some, you know, piece of paper.

但实际上账上的现金只增加了大概7亿美元，而他们报告的现金流是14亿美元。

But actual cash on the books only increased by about $700,000,000 versus $1,400,000,000 of cash flow that they reported.

所以我们得看看到底有多少钱真正计入了最终利润。

So we kind of look at like what's the real amount of money that hit the bottom line.

所以这些数字并不算大。

So these are not big numbers.

然后当你使用GAAP（公认会计准则）数据时，就像我们对特斯拉那样，因为埃隆的薪酬方案，你没法把这部分成本从收益里剔除掉。

And then when you use GAAP numbers, which we do with Tesla because of Elon's comp package, you can't really discount the cost of his comp package in earnings.

你只能把它算进去。

You just have to count it.

数额实在太庞大了。

It's so massive.

所以现在GAAP盈利只剩下大约13美分，特斯拉基本上算不上盈利，你知道的，大部分时间都是这样。

So now you're down to about 13¢ of GAAP earnings, and Tesla basically is not profitable, you know, for the most part.

你刚才提到SEC的文件。

Now you talked about the SEC filings.

今天早上我们在新闻室看那些申报材料时，发现他们悄悄加了一条关于20亿美元与一家AI硬件公司交易的备注。

We were were looking at the queue this morning in the newsroom, and they kind of snuck in this note about this $2,000,000,000 deal with an AI hardware company.

昨晚的所有演示里，他们根本没提这件事。

They didn't say anything about it in the well, in any of the presentations last night.

你有没有什么线索，知道这可能跟什么有关？

Do do you have any sense, you know, what this might be related to at all?

他们也没说对方是谁。

They didn't say who it was.

我刚才说的，你又看到了吧。

You know, once again, this is what I'm saying.

他们的财务数据里这类事情太多了，我实在没法把所有线索串起来。

There's a lot of this kind of stuff in their numbers and I just can't put it all together.

所以啊，SEC的文件里头，你没法糊弄人。

So it's like, you know, with the SEC docs, you can't really BS people.

我的意思是，没错，但这可是违法的。

So, I mean, yeah, but it's illegal.

我本来想一行一行仔细看过去，但说实话，我觉得也没那么重要。

So you know, I like to go line by line and get through this thing, but I don't think it matters that much.

我觉得趋势已经很清楚了，公司往哪个方向走，马斯克也说得明明白白：未来很长一段时间，这都将是亏损、现金流为负的业务，因为他们要押注他眼中的未来。

I think the trend is very clear where the business is going and Elon has made it very clear that this will be an unprofitable, cash flow negative business for a long time to come as they invest in what he perceives the future to be.

现在，数字就是这些数字。

Now, so the numbers are what they are.

那他在电话会议里的语气呢？那种感觉和氛围呢？

What about his tone on the call, the color that he gave?

他这次的语气，跟以前其他电话会议相比，有什么不一样？

How did that contrast with the tone that he's taken previously in other calls?

以前我们就有过这种问题，特斯拉的电话会，像昨天那种，基本都是泼冷水的类型。

Well, we used to have this problem in the old days with Tesla where Elon would do these conference calls like he did yesterday that are basically downers.

你知道吧，他做电话会议时的状态不同，语气也会不一样。

You know, it just kind of depends on the mood he's in when he's doing the calls.

2018年和2019年他老这么搞的时候，我经常发邮件问他：你到底在干什么？

And back in 2018 and '19 when he used to do that, I used to send him emails saying, what the heck are you doing?

这简直是史上最丧的一次电话会。

You know, this is, the most depressing call.

你能不能表现得积极点？别老提高管离职这种事。

Can you sound like things are, you know, going well and, you know, stop talking about executive departures and things like that?

他有回应吗？

Did he respond?

他有回应吗？

Did he respond?

有啊。

Yeah.

他确实回了。

He did.

他确实改变了财报电话会的语气。

And he did change the tone of his conference calls.

你记得吧？大概五年前，特斯拉股价开始最后那波暴涨，整个2019到2021年一路狂飙，最高涨到每股400美元，而五年前它才刚起步。

If you remember, it was about five years ago to the moment where Tesla stock made its final run up, and and it was really just an amazing run through 2019, 2020, and 2021 where it peaked at about $400 a share where it is today five years ago.

所以，回过头看那段日子，那真是特别乐观的时期。

So, you know, if you go back to that period of time, it was a very optimistic period.

要说埃隆人生中最顺的阶段，那肯定是2021年。

It was probably the best period of time in Elon's life was 2021.

可后来他觉得日子太顺了，干脆买下了推特。

And then he decided it was too good, and he decided to buy Twitter.

明白了。

Okay.

那他现在怎么又回到那种悲观消极的状态了？

So so why is he back now to the to the doom and gloomy version?

是的。

Yeah.

我觉得吧，这个人给自己施加的压力太大了。

Well, I think, a, this guy puts so much pressure on himself.

你知道的，他手里同时握着好多盘棋，那么多庞大的项目都在推进，换作普通人，根本没法健康地承受这种级别的压力，咱们就姑且这么说吧。

You know, you're talking about somebody who has so many irons in the fire, so many massive projects that he's working on that no human can absorb that level of stress healthily, let's just say that.

我敢肯定，不管你经营自己的公司，还是像我一样经营我的公司，作为一个人，能承受的 stress 是有上限的。

I'm sure running your company, just like I run mine, there's just a certain amount of stress you can deal with as a person.

显然，埃隆可能有点超人的能力，但他毕竟还是个人。

And obviously Elon has some sort of superhuman powers, but he's still human.

所以我觉得这些事情全都堆到他头上了。

And so I think you're seeing this all kind of pile up on him.

你懂的。

You know?

比如，他现在又要建什么关税工厂，对吧？这还是在原本就乐观的工厂基础上加的。而且他把Model S生产线都拆了，还有那些Robotaxi到现在还跑不起来。

Like, now he's gonna build a tariff fab factory, you know, like, on top of, an optimist factory, and he's taken apart the model s line, and he's and he's got robotaxis that don't work yet.

他一边在扩大规模，一边搞半挂车，还有各种乱七八糟的项目。

And he's putting he's scaling those and semi and this and that.

还有啊，顺便说一句，SpaceX要上市了，我们刚花了二十亿美金买下Cursor，又刚完成了XA的合并。

And then, oh, by the way, we're taking SpaceX public, and we just bought Cursor for 20 gazillion dollars, and we just got, you know, merged XA.

当然，他也不是那种...

And I'm like granted, he's not like Mr.

顾家的好男人。

Family Man.

他没什么社交生活，但这也太离谱了。

He doesn't have a social life, but still, this is insane.

要是他明天突然走了，留下那么多没做完的事，简直没法收场。

And, I mean, if he passed away, like, tomorrow, there would be so much unfinished business.

这些项目，别人得花五十年才能慢慢搞完，你知道吧？

It would take fifty years for other people to build these projects, you know?

所以我总觉得，他手里的事太多了，根本没法集中精力。

And so I just think that there's too many things going on and there's no focus.

明白了。

Okay.

所以我想要深入聊聊你刚才提到的那几件事。

So I to double click on a couple of those things you mentioned.

比如关税和制造，昨晚通话中他特别强调了这一点。

So, tariff fab is one part of the equation that he stressed on the call last night quite a bit.

作为投资者，你怎么看这个问题？它对SpaceX和特斯拉的成功，以及整体长期目标有多重要？

How are you thinking about that as an investor in terms of how important it is to the success of SpaceX and Tesla and the long term ambitions as a whole?

如果这事没做好，会不会影响到所有这些计划的成败？

Does it does it does he need to get this right for any of this to work?

因为我关注埃隆很久了，所以我能理解他。

So because I've been doing the Elon thing for a long time, I get him.

明白吗？

Okay?

他在想的是十年后自己会面临什么问题。

So he's thinking about what his problem is going to be ten years from now.

他的眼光总是比普通人，比如我，要远得多。

His vision is always well far forward of what, you know, mine or normal people's would be.

所以他在想五到十年后他会遇到什么问题，那就是如果他把所有这些产品同时扩大规模，我们都知道算力不够。

So he's thinking about the problem he's going to have in five or ten years, which is if he scales all these products at the same time, we all know there's not enough compute.

芯片根本不够。

There's just not enough chips.

而且他们现在已经面临这个问题了。

And they already have this problem.

所以获取内存芯片已经是个难题了。

So it's already a problem getting memory chips.

现在获取计算芯片或者推理芯片就已经是个难题了。

It's already a problem getting compute chips or inference chips.

所以说，建这么个工厂，对特斯拉和SpaceX来说，是非常合理的。

So so building this factory, you know, makes a ton of sense for Tesla and SpaceX.

毕竟，芯片就像是未来的大脑细胞，他们肯定需要这些。

Like, they are gonna need these like, chips are like the brain cells of the future.

对吧？

Right?

问题是，这需要大量资本投入，而且是高度专业化的技能，还得用上一大堆极其昂贵的设备，比如ASML的那些。

The problem is this is a very, like, capital intensive specialized skill set that involves tons of really expensive equipment, like, from ASML.

你想想，单是搞懂并运营这么一个业务，就已经是天大的事了。

You know, we're talking about, like, that alone is a huge business to try to build and run.

对吧？

You know?

所以我祝他成功，毕竟我们确实需要这些东西，而且这是核心环节。

And so I wish him luck with it because we do need all this, and that is a core component of it.

我质疑的是，如果我们只盯着这一块，那当然很好。

Where I question is, well, if that was all we were focused on, that'd be great.

但说实话，跟其他所有需求比起来，这点产能根本就是杯水车薪。

But, you know, it's just like a drop in the bucket with everything else.

对特斯拉来说，在利润率这么低的情况下，投这么多资本进去，其实不太划算。

And the CapEx for these things in the margin doesn't make a ton of sense for Tesla.

你刚才提到SpaceX的光刻机订单，这个我也想听听你的看法。

Now, you talked about the SpaceX cursor deal, which is another thing that I wanted to get your take on.

还有另一件事。

Another thing.

没错。

Right.

你对这笔交易怎么看？

What did you make of that deal?

你觉得这反映出 xAI 的运营情况如何？

What do you think it says about xAI and how that business is operating?

我们已经看到 xAI 的高层因为各种原因纷纷离职。

Well, we've seen the massive departures at xAI at the top for various reasons.

我觉得很大一部分原因是 Grok 并不像 Quad 那样被人广泛使用。

I think a lot of it is that Grok is not a model that people use like Quad or something else.

而且我觉得，AI 在编程方面确实挺厉害的。

And then I think, know, AI is good at computer coding.

结果发现，它在语言方面也很出色。

It's good at languages, it turns out.

你知道的，它在语言理解和模式学习方面真的非常擅长。

You know, it's like really good at language and learning patterns.

所以，编程是一个变化非常快的领域。

And so, like coding is something that is rapidly changing.

我们在ServiceNow这样的软件股身上看到了这一点。

We're seeing this in the software stocks like ServiceNow.

我们正在经历这种巨大的变化。

We're seeing this massive change.

所以，就像我们公司内部在云端开发了一套软件，目前正在测试中。

So, like, we've built software on cloud internally at our firm that we're testing right now.

你知道的，每隔几天我们就特别抓狂，觉得这玩意儿简直没法用，想直接扔了，但没过几天他们又升级一下，效果就好多了。

And, you know, every few days, we get super frustrated, thrown in the trash, and then a few days later, they upgrade stuff and it works better.

对吧？

You know?

而且我们也在实时摸索它到底能做到什么程度，结果发现它能做的事情真的让人惊叹。

And and we're kinda trying in real time to see what the capabilities of this is, and it's amazing what you can do.

所以我认为他们的意思是，如果我们打算建数据中心、打造AI基础，那未来就得有足够强的编程能力，快速开发出高质量的软件。

So I I think what they're saying is if we're gonna build data centers and we're gonna have an AI base, We want the coding chops to build super fast, super good software in the future.

我觉得这更多是关于内部使用，而不是做产品去卖，关键是怎么用好这些编程工具。

So I think it's more about internal than a product to sell, but more about how do we use this coding software?

我们该怎么升级现有的工作方式，充分利用手头的服务器，这些服务器

How do we upgrade what we're doing and use the servers we have, which are

多得不得了，而且得真正用上所有这些计算资源，毕竟他们投了这么多钱在上面。

tons And of actually make use of all I mean, all the all this compute that they're

他们正在投入的。

that they're

投资的计算资源。

investing in.

我们得有应用程序来用起来，而且得

Like, we have to we need applications to use And

SpaceX 就写了大量的代码。

So so SpaceX writes tons of code.

从这个角度看，这确实非常说得通。

So this just, like, makes a ton of sense from that perspective.

不过，花六百亿美元买下它，这价格实在太高了。

Now once again, paying $60,000,000,000 for this is a huge price to pay.

但话说回来，当你手握两万亿美元的估值，随便发点股票，对他们来说就跟不要钱似的。

But, you know, when you're throwing stock around at a $2,000,000,000,000 valuation, it's it's almost like free to them, you know?

他就像个糖果机，把糖果往四面八方撒，因为他想为未来十年的蓝图把所有零件都凑齐。

So he's got this like candy machine and he's throwing the candy at everybody because he's trying to put all the pieces in place for what he perceives the future to be in about ten years.

我不会说他错了。

And I'm not gonna say he's wrong.

我只是觉得，这跟他当年做特斯拉差不多，把所有资源都押上去，拼命冲向下一个目标。要是失败了，那就跟泰坦尼克号一样。

I I just think it's like it's like what he did with Tesla where you leverage everything to try and But get to that next if he doesn't succeed, it's like the Titanic, you know?

那我有个问题，SpaceX和特斯拉合并，你觉得这主意怎么样？

So my question is, I mean, a SpaceX and a Tesla merger, I mean, do you think that's a good idea?

我觉得这迟早会发生，因为投资者真正想要的是同时持有这两家公司的股票。很多特斯拉的股东其实想买SpaceX，所以他们会卖掉一部分特斯拉股票来换SpaceX的。

Well, I think it's an inevitability because when you actually look at what investors want is they wanna own both stocks, but many of the Tesla investors wanna own SpaceX, and they're gonna sell their Tesla stock or some of their Tesla stock to buy SpaceX.

所以这会给特斯拉带来抛售压力，再加上同时运营两家上市公司的利益冲突，简直让人面临无数法律纠纷和利益冲突的风险。

So it creates selling pressure on Tesla along with the conflicts of interest of basically running two public companies, and it just exposes you to so many, conflict of law you know, interest lawsuits.

而且因为这两家公司的业务高度交织——SpaceX还掌握着人形机器人的核心技术——合并起来反而是最自然的选择。

And and so, ultimately, because they're so intertwined, the businesses and SpaceX owns the brains to the optimist robot and all this, that it just makes the most sense for them to come together.

一旦合并，投资就变得非常二元化：你要么相信埃隆，押注他，要么就不信。

And if they come together, you've got a very binary investment of, like, I'm either investing with Elon and I'd like it or I don't.

所有这些业务整合在一起，就形成了一家类似伯克希尔·哈撒韦的AI巨头，如果你想真正押注未来，这几乎是纯粹的标的。

And you've got sort of all these businesses together that create, like, a Berkshire Hathaway of AI and something that if people really wanna bet on the future, this is a very pure play on that.

所以但是

So But

但你不觉得他可能想稍微对冲一下风险吗？

but you you don't think maybe he'd wanna hedge his bets a bit?

你知道的。

You know?

如果SpaceX没达到他预期的效果，他或许还想保留特斯拉，让这个乐观的赌注保持相对独立。

Like, if SpaceX doesn't doesn't do what he hopes it to do, maybe he wants to keep Tesla sort of, you know, the the optimist bet somewhat sep separate.

但你知道，埃隆根本不是这么想的，他从不搞对冲。

Well, you know, that's just not the way that Elon thinks he's not a hedger.

他是个全力以赴的人。

He's an all in guy.

恰恰相反，他就是这样。

So it's exactly the opposite.

我觉得他是打算把整个价值上万亿美元的商业帝国押上去，想把事情放大十倍，做成一个价值十万亿甚至更多的巨头，彻底统治科技和航天领域。

I think it's more that he's about to leverage his entire, you know, multi trillion dollar empire to try to do something to 10 x it, you know, into something that looks like 10,000,000,000,000 plus dollars that's just dominating every area of technology and space.

所以，这事到底能不能成呢？

So, you know, will that happen or not?

我从不跟埃隆对着干。

I'm not somebody who bets against Elon.

你知道的。

You know?

正因如此，我们才持有这支股票。

I and that's why we do own the stock.

但我确实认为这个愿景存在巨大的风险，而且很多风险是普通投资者可能没考虑到的，比如他的健康状况，以及他能否在未来十年里坚持实现这个愿景。

But I I do think that there's a tremendous amount of risk in the vision, and many of those risks are things that many investors might not consider, like just his health and his ability to see through this vision over the next decade.

对了，罗斯，我快速问一下，你一开始提到，在2018年和2019年你给他发邮件的时候，我问过他有没有回复。

Now I I just very quickly, Ross, you said at the very start, when you you sent in emails, 2018, 2019, I asked you, did he respond?

你说他回了。

You said he came back.

你现在还在跟他发邮件吗？

Is he are you emailing with him?

他有没有在回复

Is he is he responding

你的邮件？

to your emails?

我以前跟他们团队有邮件往来。

I used to email with their team.

你知道的，并不是那种‘哦，埃隆，这个那个’的闲聊。

You know, it wasn't like, oh, Elon, this and that.

他们那时候有个团队，但后来把很多人都裁了。

They have a team well, back then, they they kinda got rid of all these people.

所以当时负责投资者关系和沟通的人，其实都是临时兼着做的，因为他们根本没设专门的公关团队。

And so, you know, there were people that would, like, handle the investor relations and sort of like the communications because they actually didn't have a PR communications team.

结果我就这么被推上了前台，因为老上电视谈特斯拉，不知不觉就变成了他们的公关代表。

And so what happened was I became the PR communications team just by being on TV so much talking about Tesla.

我那时候还挺烦的，老说：你们怎么就不找个发言人呢？

And I used to get annoyed and say, you know, why don't you guys have some spokesperson or whatever?

他们就回我：对不起，罗斯。

And they were like, sorry, Ross.

你做得实在太好了。

You know, you do such a good job.

我们真觉得没必要再另找人了。

We'll just like, we don't need it.

你懂的。

You know?

所以我这么干了好几年，就在各大电视台做特斯拉的专访。

And so I did this for, like, years, you know, just doing Tesla spots on on major networks.

但每次他开那些悲观的电话会议，我都特别来气。

And and but I would get super annoyed when he would give these, like, downbeat conference calls.

可一旦他换种语气，变得特别乐观，伊隆这个人其实挺有魅力的，也很讨人喜欢。

And then, you know, when he would change his tone and and be really optimistic, Elon is a very, like, engaging and likable guy when he wants to be.

他一开积极的电话会议，股价就立马往上冲。

And, you know, the stock would rally when he would do a positive conference call.

现在他没人帮衬了。

So, you know, he doesn't have that help anymore.

他把当年那些对他帮助很大的人全给撤了，现在的团队跟以前完全不一样。

He got rid of a lot of those people that I think were really beneficial to him from those days, and and it's a whole different team today.

所以他现在想怎么管就怎么管了。

And and so he's kinda running running things the way he wants to be running them now.

可惜的是，他们现在压力特别大。

And unfortunately, they've got a lot of lot of pressure on them right now.

没错。

Right.

好的，罗斯，非常感谢你来到节目。

Well, Ross, I wanna thank you for coming on.

一如既往，这里是Gerber Kawasaki的总裁兼首席执行官罗斯·格伯，欢迎收看TITV。

As always, that is Ross Gerber, president and CEO of Gerber Kawasaki here on TITV.

谷歌本周在拉斯维加斯举办其旗舰云会议。

Google is hosting its flagship cloud conference in Las Vegas this week.

我的同事 Erin Woo 在现场，并在今天的《应用人工智能》通讯中对此进行了报道。

My colleague Erin Woo is on the ground and wrote about it in today's Applied AI newsletter.

她一直在关注这家公司发布的一系列重大公告。

She was looking at the big announcements that the company has made.

我想请她来详细聊聊这些内容。

I want to bring her on to talk all about it.

Erin，欢迎来到拉斯维加斯。

Erin, welcome to you from Las Vegas.

能跟我们说说吗？谷歌这次云峰会发布了哪些新消息？

Tell us, what did Google announce this week at its cloud conference?

谷歌简要介绍了多种帮助企业搭建和管理AI代理的方法。

Google announced in brief a lot of different ways for companies to get set up and manage AI agents.

今年峰会的核心主题就是AI代理。

And so that was really the big theme of this year's conference was AI agents.

他们还推出了新一代的芯片——张量处理单元。

They also announced a new generation of their chips, tensor processing units.

这是第一次，他们分别推出了用于训练AI模型的芯片和用于运行模型的推理芯片。

For the first time, there are two separate chips for training, the training of AI models, and inference, which is the process of running processes on those models.

新代理，新芯片。

So new agents, new chips.

这一周我们都会持续跟进报道这些内容。

I mean, we will be writing about this in our coverage all week.

大家可以关注一下。

People can check it out.

我特别想从你这里得到的一个更关键的问题是，你参加了这次大会。

The bigger question that I want to get an answer from you, Erin, is I mean, you went into this conference.

你和谷歌云的那些客户聊了很多，你带着哪些主要问题去的呢？

You're talking to all these customers of Google Cloud What on the were the big questions that you went into this event with?

然后我们再聊聊你从这些对话里学到了什么。

And then we'll talk about what you learned from all these conversations.

我

I

说实话，我最大的疑问是，企业到底是怎么用AI的？

mean, really, like, my biggest question was how are companies using AI?

它真的管用吗？

Is it working?

他们在这上面花了多少钱？

How much are they spending on it?

他们觉得这笔投资有回报吗？

Do they feel like they're getting a return on that investment?

所以我来参加这次会议，主要是想和大量谷歌云的客户聊聊。

And so I really came into this conference wanting to talk to a ton of customers of Google Cloud.

明白了。

Okay.

那他们跟你说了些什么？

And what did they tell you?

有一个主题我本来没料到，但昨天和所有人的交流中都自然地冒了出来：现在想真正用上AI，还是特别难。

So one theme that I didn't quite expect, but that came up quite organically in all my conversations yesterday, was it's really hard still to get started using AI.

公司和谷歌云的合作伙伴都告诉我，从零开始搭建第一个智能代理就有很多麻烦，比如确保数据和流程都对得上。

And so companies and Google Cloud resellers were telling me that there's difficulties like setting up your first agent, like making sure like your data and your processes are correct.

而对于那些已经用了智能代理的公司，我接触的一位人士把这叫作‘代理泛滥’。

And then for companies that have agents, one source that I talk to described this as agent sprawl.

各个员工各自搭建了不同的代理，干着五花八门的事，结果现在根本管不过来。

And so there's just all of these different agents set up by different employees doing different things, and so it's getting hard to manage all of this.

我觉得你在通讯里有一句话说得特别精辟：谷歌和其他前沿实验室推出的工具能力，已经远远超过了大多数公司理解和落地它们的能力，尤其是当智能代理和其他AI产品变得越来越复杂的时候。

I thought one line that you put very concisely and well put in your newsletter was you said, The abilities of tools from Google and other frontier labs are outpacing the abilities of many companies to understand and implement them, particularly as agents and other AI offerings get more complex.

所以意思就是，实验室那边创新不断，模型越做越好，越来越强。

So the idea here is that, I mean, the labs have all this innovation, the models are getting better and better and better.

但与此同时，像家得宝这样的客户却说，别减速，但我们真需要点帮助，搞清楚怎么实际落地应用。

And meanwhile, you have customers like, you talk to somebody from The Home Depot, for example, who are basically saying, Not slow down, but we need a little bit of help figuring out how to actually implement them.

是的。

Yeah.

昨天谷歌宣布了一项7.15亿美元的基金，专门支持咨询公司帮助客户接入和启动AI项目。

So one of Google's announcements yesterday was a $715,000,000 fund for consulting firms to help their customers get adopted and get started with AI.

这跟OpenAI之前跟咨询公司做的合作其实很相似。

And so this mirrors things that, for example, OpenAI has also done similar deals with consulting firms.

尽管AI公司总说要简化流程、淘汰中间环节，可到头来还得靠这些咨询公司来教大家怎么用。

And so even though AI companies are saying like, Oh, we're going to cut out a lot of these processes, they're also needing these exact same consulting firms to teach people how to use it.

我跟家得宝的执行副总裁乔丹·布拉吉聊过。

And so I was chatting with Jordan Bragi, who's an Executive Vice President at The Home Depot.

他说了个特别有意思的说法，把这事儿比作工厂的生产线。

And he said this thing that was really interesting was that he was comparing it kind of to a factory floor.

所以，从头到尾把一个产品做出来，这个想法是关键。

And so the idea of getting a product from beginning to the end.

而像家得宝这样的公司，目前正处于自动化部分流程的阶段。

And at this point where he said that The Home Depot and other companies are at is they are automating some steps of the process.

他们已经给所有企业员工都配备了Gemini，也给团队提供了Copilot。

And so they've given all of their corporate employees, for example, Gemini, they've given them Copilot.

软件开发团队在用OpenAI、Anthropic和Gemini的各种工具，来回切换着试。

The software development teams are using things from OpenAI and Anthropic and Gemini and like switching in between those.

但他主要想说的是，他们在生产线上已经能自动化某些环节了。

But what he's saying basically is that they have figured out how to automate certain steps on the factory floor.

这些环节确实变快了，但一碰到下一个步骤，又卡回原来的老问题，流程还是老样子。

And so those steps are getting so much faster, but then it only really speeds up until you get to the next step where then you're just back to kind of the same bottlenecks and the same workflows.

他提到，家得宝和其他公司都看到了单个环节的进步，但没人真正彻底重构过整个流程。

And so he was saying that The Home Depot and other companies have seen advances in individual steps, but no one's really been able to overhaul the whole workflow.

只有做到这一步，才可能带来真正的飞跃。

That's what would need to happen for a really drastic step change.

为了明确一下，你刚才说的是那位跟你聊过的高管，他负责家得宝平台网站的软件部分。

And just to be clear here, you're talking about this executive you talked to, they're in charge of the software side of Home Depot's platform website.

能详细说说他们的工作吗？

Walk us through their

他主要说他是在线业务的副总裁。

They main said they're vice president of online.

对。

Yeah.

就是在线业务这块的工作。

So like work.

在线业务。

Online.

明白了。

Okay.

他们说，我们有一群专注于产品的人员，比如软件工程师可以开发出更好的功能或优化某些功能，但要把这些更新接入网站，再调整商业模式，这些环节依然是瓶颈，根本提速不了。

And so they're saying that I have these product focused people, for example, that maybe the software engineers can code up something better or get a feature better, but then plugging it into the website, figuring out how the business models will change, that is still a bottleneck that doesn't get any faster.

是的。

Yeah.

他跟我讲的是，我可以把写代码的时间砍掉一半，或者类似这样。

So what he was saying to me was basically like, Okay, so I can take the amount of time that it takes to write a piece of code and I can cut that in half or whatever.

但讨论同事、做市场调研、跑测试、决定X、Y、Z这些事呢？

But what about the process of having discussions with colleagues, of doing market research, of like running tests, of deciding like X and Y and Z.

做生意真的太复杂了，这是我在这次演讲大会上学到的。

Like there's so many like it's so complicated to run a business is the thing I'm learning at the speech conference.

所以像那些被吹得天花乱坠的编码工具，比如Cloud Code，或者谷歌那一堆编程工具。

And so like there's all these other steps that, for example, like these highly touted coding tools like Cloud Code or Google's got like a million coding tools.

这些工具确实能加快流程中的某些环节，但整个大流程还是没变。

And so all of these things can speed up steps in the process, but there's still this whole big process.

那你离开第一天的会议时，有没有觉得人类在AI时代还是有位置的？

So did you walk well, you haven't walked away yet, but did you walk away from the first day of the conference feeling a little bit more optimistic that humans still have a role in AI?

嗯，有吧。

I mean, yes.

人类在人工智能领域 definitely 依然扮演着重要角色。

Humans definitely still have a role at AI.

当然了。

Absolutely.

所有这些复杂的公司流程，根本没法完全自动化。

All of these very complex corporate processes cannot be automated.

但与此同时，有个客户圆桌讨论，客户们在使用 Gemini Enterprise，也提到了一些具体案例，比如用它来做客户服务。

But at the same time, there is a customer panel with customers using Gemini Enterprise and there is a couple use cases around like, oh, we're using this for customer service.

用它来自动化客服机器人，或者增强客服支持能力。

We're using this to automate customer service agent bots or have more support for that.

所以没人会明说‘我们要因为这个裁员’，毕竟大家都明白现在的舆论环境有多敏感。

And so nobody will come outright and say that, oh, we're going to lay employees off because of this, because everyone kind of like understands the press situation that they're in.

但有人提到，比如花旗财富管理的一位高管就说，要是想把财富管理业务扩大到我们想要的规模，原本得招几千人。

But like people are saying like, oh, like an executive from Citi Wealth was saying that, well, to expand the wealth management business as much as we want, like we'd have to hire like thousands of people.

但现在有了这个 AI 代理，我们用更少的人就能做到。

But now we have this AI agent that lets us do it with fewer employees.

比如家得宝也提到，员工离开客服中心后，我们其实没必要像以前那样招那么多人了。

Or like The Home Depot was also saying like, well, like we don't necessarily, like when people leave call centers, we don't necessarily have to hire as many people as we like maybe would have used to.

确实有一些工作正在被自动化取代。

And so like there are certainly jobs that are being automated.

这正在影响企业未来的招聘计划。

This is impacting the amount that companies are planning to hire.

最后一个问题。

Last question for you.

我知道这是谷歌云大会，但客户在交流中有没有提到其他AI公司，比如Anthropic或者OpenAI？

I know this is the Google Cloud Conference, but did any of the other AI labs come up in conversations with customers, the Anthropix and OpenAI's of the world?

现在大家是怎么看待谷歌在竞争格局中的位置的？

What are people thinking about Google's place in the competitive landscape right now?

对我来说，有意思的是Copilot总被提到，不过说句对不起微软的话，在旧金山跟人聊天，特别是跟FrontierLab的人聊时，根本没人谈Copilot。

I mean, one of the interesting things to me was like how often Copilot came up because with apologies to Microsoft, like I feel like when I'm talking to like people in San Francisco, when I'm talking to like FrontierLab employees, like nobody is really talking about Copilot.

大家觉得这是一款人人都在用的重要AI工具。

It's like, oh, this is like a major AI tool that everyone's using.

但Copilot在对话中频繁出现，这说明了微软Office套件的强大影响力，以及它已经深深融入日常工作的程度——既然我们本来就在用Office，那用Copilot也就顺理成章了。

But like CoPilot came up a lot in conversations, which is a testament to the power of the Microsoft Office suite and just kind of like how ingrained this is and how it's like, well, we're already using Microsoft Office so now we can use CoPilot.

当然，Anthropic也经常被提到。

And of course, Anthropic also came up a lot.

很多关于编程的讨论都提到，大家对Anthropic新发布的模型非常期待。

A lot of the conversations around coding were talking about like, well, we're really excited about Anthropic's new model releases.

你有没有感觉到，人们喜欢Copilot，是因为它容易上手，毕竟他们早就用了这些微软软件？

Did you get the sense that people were happy with Copilot or just that it was the easiest thing to implement because they already have all of this software

是的。

Yeah.

我的意思是，人们并没有特别推崇Copilot。

With mean, people were not singing Copilot's praises necessarily.

他们也没抱怨Copilot，但也没觉得这是他们用过最棒的AI产品。

People were not like complaining about Copilot, but they were not like, wow, this is the best AI product we've ever had.

有个人告诉我，他们才刚开始试用Gemini，目的是想分散一下风险。

And like one person was telling me like, yeah, like we also like we're just getting started on Gemini because we want to diversify.

所以说实话，大家根本不会再单纯靠谷歌搜索了。

So I mean, if anything, like the theme is that no one is just going to Google for stuff.

每个人都试着用不同厂商的产品，来回切换着用。

Everyone is trying out products from different model developers and they are switching back and forth between them.

我觉得有点意思的是，连谷歌大会上的Copilot用户，都在找别的替代方案。

And I think perhaps it's telling in some ways that you have co pilot customers at the Google conference maybe looking for an alternative.

说到底，这类大会的核心目的就是让人们对新东西有所了解。

I mean, that's kind of the whole point of these conferences is to educate people.

很多有趣的观察发现。

A lot of interesting observations.

埃琳，谢谢你来参加节目。

Erin, I want to thank you for coming on.

这位是埃琳·吴，我们派驻在拉斯维加斯的谷歌记者，来自TITV。

That is Erin Woo, our Google reporter on the ground in Las Vegas here on TITV.

ServiceNow第一季度营收增长了22%，但投资者更关注接下来的发展。

ServiceNow posted 22% revenue growth in the first quarter, but investors are focusing on what comes next.

公司表示，最近完成的对网络安全公司Armis的收购拖累了利润。

The company says its recently completed acquisition of cybersecurity firm Armis is weighing on profits.

我的同事劳拉·布拉顿正在跟进这个报道。

My colleague Laura Bratton is covering the story.

她现在加入我们，带来更深入的分析。

She joins me now with more analysis.

劳拉，欢迎再次做客节目。

Laura, welcome back to the show.

很高兴你今天能来。

It's great to have you here.

我们来聊聊ServiceNow。

Let's talk about ServiceNow.

昨晚公布的财报中，有什么特别让你注意的地方？

What stood out to you from the results last night?

是的。

Yeah.

让我印象最深的是，你提到ServiceNow说Armis这笔收购会略微拖累利润，他们下调了全年盈利预期，却同时上调了收入预期。

What stood out to me was what you talked about the fact that ServiceNow said the ARMS deal is going to dig into profit slightly, and they lowered their guidance for profitability metrics for the full year while also raising their guidance for revenue.

我觉得投资者还向我指出一点：虽然订阅收入增长超过了ServiceNow之前预测的水平，但增速却低于上个季度。

I think another notable point that investors pointed out to me was that while subscription revenues grew more than ServiceNow had previously said, It grew less than the previous quarter.

这在一定程度上是因为中东地区因地缘政治冲突，导致一些交易被推迟了。

And that was partly because of some delayed deals in The Middle East because of geopolitical conflicts there.

所以投资者说，这些才是他们最关心的问题。

So investors said that those were the things that were on their mind.

我现在正在看股价。

So I'm looking at shares right now.

今天早上股价已经跌了超过百分之十五。

They're down more than 15% this morning.

你接触的那些人，觉得这是不是反应过度了？

The people you're talking to, is this an overreaction?

我的意思是，这家公司增速依然超过百分之二十。

I mean, here's a company that is still growing faster than 20%.

我明白毛利率的问题。

I get the gross margin issue.

我也理解那些收购案。

I get the acquisitions.

我们之前在节目里也说过这一点。

Again, we've said this on the show.

这家公司依然在健康增长。

This is a company that is still growing healthily.

它还在持续产生自由现金流。

It's still generating free cash flow.

你怎么看股价的下跌？

What do you make of the stock drop?

是的，我觉得这被夸大了。

Yeah, I think it's really overblown.

就像你提到的，ServiceNow 在其领域里依然是顶尖表现者。

I mean, ServiceNow, as you mentioned, it's a top performer in its category.

它为客户提供IT服务和IT服务管理软件。

Provides IT services, IT service management for its customers and software for that.

它在这一领域绝对是领头羊。

And it's really a leader in its category.

它的增长速度远超其他软件公司，比如Adobe、Salesforce、HubSpot和Atlassian。

It's also growing a lot faster than other software companies, know, such as Adobe, Salesforce, HubSpot, Atlassian.

所以投资者也觉得，正如你所说，这次股价下跌被夸大了。

And so investors really think that this reaction is overblown, as you said.

我还会想，什么样的公司适合做收购方，什么样的公司适合被收购？

And I also think about it a little bit in the way that what company would be a candidate to be acquired and what company would be a candidate to be an acquirer?

显然，ServiceNow已经表明它愿意进行收购。

Obviously, ServiceNow has shown that it is willing to make acquisitions.

收购当然要花成本，但说实话，我也不太确定。

Obviously, it will take a cost to make these acquisitions, but, you know, I I don't know.

我还是从AI的角度来看，觉得它站在了最前沿。

I I still look at it from the AI story as, like, it's on the forefront.

但你知道，这种事情每季度都会发生。

And so but, you know, this happens every quarter.

对吧？

Right?

他们总是搞这种超预期和赶超的把戏，SaaS的末日论还是老样子。

Like, they they do the beat and race thing, and it's SaaS apocalypse still.

我觉得这事儿也被夸大了。

And it's, I don't know, I think it's overblown as well.

是的。

Yeah.

就像我接触的一些投资者说的，现在市场太敏感了，只要有一点不完美，这些软件股立马就会引发抛售。

And I mean, as investors I talked to said, you know, the market is so sensitive right now, and any anything that's not perfect is gonna send, you know, the market into sell off mode for these software stocks.

投资者对ServiceNow的战略很有信心。

And investors are confident in in ServiceNow strategy.

ServiceNow真正想做的，是从一家传统软件公司转型成管理AI代理的中央仪表盘，帮企业监控风险、保障这些AI代理的安全。

ServiceNow really wants to go from being a traditional software company to being kind of a dashboard for managing AI agents and helping companies monitor risks and safeguard those AI agents.

投资者觉得，这个战略说不定真能见效。

And investors think that, you know, that strategy could pay off.

但即使不成功，ServiceNow 依然是个业绩出色的 IT 服务管理软件公司，在这个领域处于领先地位，增长速度也远超其他软件公司。

But also even if it doesn't, ServiceNow is still a high performing IT service management software company that's a leader in this category that's growing faster than other software companies broadly.

所以目前来看，其实没什么太大下行风险。

And so, you know, there's there's really not a lot of downside at the moment.

除非你特别特别担心 AI 会一夜之间彻底颠覆整个软件行业——有些投资者确实这么想。

If you're not, you know, super, super worried about AI completely disrupting and eating the software industry overnight as some investors clearly are.

我们来聊聊你今天发表的那篇报道。

Let's pivot to a story that you published today.

你写了一篇独家报道，讲保险公司是怎么看待 AI 的。

You wrote an exclusive story about the insurance sector and how those companies are looking at AI.

跟我们说说，你采访中发现了什么？

Tell us a little bit about what you found in your reporting.

是的。

Yeah.