亚马逊潜在的500亿美元OpenAI投资、英伟达令人印象深刻的收益与股价下跌

欢迎各位收看Informations TI TV。

Welcome everyone to the Informations TI TV.

我是阿卡什·瓦斯里查。

My name is Akash Vasricha.

今天是2月26日，星期四。

It is Thursday, February 26.

今天我们节目将讨论英伟达的季度重大业绩。

We are talking about NVIDIA's big quarterly results today on the show.

这家公司的业务正在加速增长，这在其规模下尤为惊人。

The company's business is accelerating, which is remarkable at its scale.

我们还会讨论Salesforce和Snowflake的业绩。

We'll also talk about Salesforce and Snowflake's results.

接着我们会探讨大型科技公司正在筹集的债务，以及这些债务如何改变它们的信用状况。

We'll then talk about the debt that big tech companies are raising and how their credit profiles are changing because of it.

我们还将独家披露一家由Kleiner Perkins支持的国防初创公司融资高达15亿美元的消息，节目最后，我们将与Encord的联合创始人讨论机器人数据，而今天早上我们刚刚独家报道了这家公司的相关新闻。

We'll also get to a scoop about a Kleiner Perkins backed defense startup raising up to $1,500,000,000 and we're going to end the show with a conversation about robotics data with the co founders of Encord, a company that we broke some exclusive news about this morning.

但我想今天从昨晚刚发布的一条重磅新闻开始节目。随着OpenAI完成其高达1000亿美元的融资，一些投资者同意的融资条款细节正逐渐浮出水面。

But I want to start the show today with a big story that we published late last night, As OpenAI brings together its massive $100,000,000,000 funding round, details are starting to emerge about the terms of the funding that some investors have agreed to.

我的同事Shri Mukhedi披露了亚马逊拟对OpenAI投资的一些细节，我想请她来详细谈谈。

My colleague, Shri Mukhedi, revealed some of the details of Amazon's proposed investment in OpenAI, I want to bring her on to talk all about it.

Sri，欢迎再次回到节目。

Sri, welcome back to the show.

很高兴你在这里。

It's great to have you here.

很高兴能来。

Excited to be here.

关于亚马逊的投资，我们了解到什么？

What did we learn about Amazon's investment?

我们了解到，亚马逊计划投资高达500亿美元，但有一个条件。

What we learned is that Amazon is investing up to $50,000,000,000 but there is a condition.

第一笔资金将为150亿美元，作为OpenAI当前融资轮的一部分。

So the first tranche will be 15,000,000,000 as part of the round that OpenAI is raising now.

而第二笔350亿美元的付款则取决于OpenAI是否实现通用人工智能（AGI），或者该公司是否上市。

And then the 35,000,000,000 second installment is actually conditioned on whether OpenAI achieves AGI or artificial general intelligence or if the company goes public.

这一点与我们之前所知的情况大不相同，相当新颖。

So that piece is quite new and different from what we had known before.

亚马逊为何要在这里对冲风险呢？

Why is Amazon hedging its bet here?

我的意思是，它难道不认为OpenAI会上市吗？

I mean, does it not think that it's going to IPO?

这还是个问题吗？

Is that still a question?

我们稍后也会解释一下通用人工智能的定义。

We'll get to the definition of AGI in a second too.

完全正确。

Totally.

据我理解，这实际上是亚马逊的一种机会——如果上述任一条件达成，它就能因微软与OpenAI的独家协议而从这种合作关系中获得更大收益。

So my understanding is that essentially it is a opportunity for Amazon to gain more from its relationship from OpenAI if either of those conditions happens due to the exclusivity clause that Microsoft has with OpenAI.

微软一直是OpenAI的长期支持者，投入了约130亿美元，并拥有独家转售OpenAI模型的权利，还能获得20%的收入分成，这一点我们在以往的报道中都提到过。

So Microsoft has been a long time backer of OpenAI, poured around 13,000,000,000 and has exclusivity rights to resell OpenAI's models, as well as actually take a 20% revenue share, which we've all reported before in our past reporting.

但我认为这些条款本质上允许亚马逊从其与OpenAI的关系中获得更多的收益。

But I think these clauses essentially allow Amazon to potentially gain more from its relationship.

因此，如果OpenAI确实实现了通用人工智能，那么亚马逊就能超越其目前所拥有的能力获得更大收益。

So if, for example, OpenAI does achieve AGI, then it allows Amazon to gain beyond what this existing ability it has now.

这些就是亚马逊能够设置这些条件的部分原因。

Those are some of the reasons why Amazon is able to have these conditions.

那么关于通用人工智能的定义，现在比十二或十八个月前这场讨论刚兴起时更清晰了吗？

Now the definition around AGI, is that any clearer than it was say twelve or eighteen months ago when this whole discussion sort of came into vogue?

对我来说并不完全清楚，但我们在2024年之前的报道中曾提到，通用人工智能基本上是指OpenAI基于我们过去对公司财务状况的分析，实现了1000亿美元的利润，而据我们判断，OpenAI甚至到2030年之后才可能达到这一目标。

It's not exactly clear to me, but what we had previously reported back in 2024 was that AGI was basically if OpenAI generated $100,000,000,000 in profits based on sort of our reporting on the company's financials in the past, that com OpenAI won't actually achieve that even till beyond 2030.

因此，我猜测通用人工智能的定义很可能已经改变了，但我也不确定。

And so I my hunch is that the definition of AGI has likely changed, but I'm not sure.

但我们知道的是，无论OpenAI何时定义自己实现了通用人工智能，都会由一个独立的专家小组进行验证。

But what we do know is that whatever or whenever OpenAI defines that they've achieved AGI, they'll actually be verified by an expert panel separately.

因此，微软与OpenAI协议中的这一条件同样适用于亚马逊与OpenAI的关系。

And so that sort of condition from the micro Microsoft OpenAI agreement still holds for its relationship with Amazon as well.

但我们还不知道这个专家组的成员是谁，对吧？

And we don't know who's on that panel yet, do we?

我认为还不知道。

I don't believe so.

嗯，能坐在这个专家组里一定挺有意思的。

Well, that's gotta be a pretty fun panel to sit on.

我的意思是，我只是在想AGI的定义有多大的争议性。

I mean, this is like I'm just thinking about the definition of AGI and how contested that is.

这些人掌握着巨大的权力，不仅对这一笔交易如此，我相信一旦他们所使用的标准公布出来，一定会受到各方的严密审视。

These people really hold so much power, not just for this one deal, but I'm sure that will get scrutinized by everyone if it ever comes out about the criteria they were using.

这确实是一群有趣的人。

That's kind of an interesting group.

你之前还报道过OpenAI和亚马逊之间的云服务协议。

You also had some reporting about the cloud deals between OpenAI and Amazon.

那里我们了解多少？

What do we know there?

完全正确。

Totally.

亚马逊和OpenAI达成了一项云计算协议，OpenAI将在七年内购买价值约380亿美元的云计算服务。

So Amazon and OpenAI agreed to a cloud computing deal for OpenAI to purchase about $38,000,000,000 worth of cloud computing over the course of seven years.

我们昨晚报道的是，这项云计算协议的规模将远超这个数字。

What we reported last night is that cloud deal is gonna expand much larger than that.

所以我们不知道具体会扩大到什么程度，但肯定会大幅扩展。

And so we don't know the order of magnitude, but it's gonna be a significant expansion.

该协议的其他方面还涉及OpenAI为亚马逊的定制模型提供支持。

Other aspects of the deal is also around Amazon actually I'm sorry, OpenAI powering Amazon's custom models.

因此，这可能也包括亚马逊内部使用的各种产品。

And so that could be for internal products that Amazon uses as well.

好的。

Okay.

最后一个问题给你。

And last question for you.

我的意思是，亚马逊和OpenAI之间的关系，以及这种关系的条款。

I mean, is the relationship between Amazon and OpenAI and the terms of that relationship.

我们是否了解其他超大规模云服务商或其它战略投资者的投资情况，以及他们达成的条款？

Do we know anything about the other hyperscalers investments or the other strategic investors and what terms they've agreed to?

我们有一些关于Nvidia交易具体细节的最新报道。

We do have some new reporting about specifics around, like, the Nvidia deal.

所以，我们之前报道过，Nvidia将投资最多300亿美元。

So the the new piece is that we had reported previously that Nvidia would invest up to 30,000,000,000.

我们现在了解到，这实际上是分三笔支付，每年100亿美元，这与软银的投资非常相似，软银的投资也是300亿美元，分三笔，每年100亿美元。

What we now learned is that it's actually gonna be three installments of 10,000,000,000 over the course of the year, and that's very similar to the SoftBank investment, which is also $30,000,000,000 with $3,000,000,000 or sorry, three installments of $10,000,000,000 over the course of the year as well.

很好。

Great.

Shree，非常感谢你前来做客。

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

这位是谢丽·穆皮蒂，我们《信息》杂志负责OpenAI和Anthropic的记者。

That is Shree Mupiti, our OpenAI and Anthropic Reporter here at The Information.

今天早上最大的新闻是投资者对英伟达财报的反应。

The big news this morning is how investors are reacting to NVIDIA's results.

营收增长了73%，较上一季度显著加速。

Revenue grew 73%, a dramatic acceleration from last quarter.

净利润几乎翻了一番。

Net income nearly doubled.

现在邀请到我身边解读财报的是韦德布什公司股权研究主管马特·布莱森。

Joining me now to break down the results is Matt Bryson, Managing Director of Equity Research at Wedbush.

马特，欢迎来到TI TV。

Matt, welcome to TI TV.

很高兴你来到这里。

It's great to have you here.

是的。

Yeah.

能来这里真是太棒了。

It's awesome to be Eric off.

昨晚的结果中，有什么特别吸引你的地方吗？

What stood out to you from the results last night?

依我看来，全是好消息。

I mean, as far as I could tell, it was it was all good news.

对吧？

Right?

正如你所说，本季度的收入实现了加速增长。

They as you said, revenues accelerated in the quarter.

他们的指引远超任何人预期。

The guide was well above what anyone had on their their bingo card.

你还要记住，对于英伟达来说，当他们指引780亿美元时，实际上是在暗示他们觉得自己能接近800亿美元，因为他们会把一部分收入留到下个季度确认。

And you have to remember also with Nvidia that when they guide for 78,000,000,000, they're really telling you that they think they can do closer to 80,000,000,000 because they they move a little bit of revenue in their back pocket.

所以，从业绩角度来看，一切看起来都非常好。

So, you know, from a results perspective, everything looked great.

但今天早上股价却下跌了。

And yet, chairs are down this morning.

是的。

Yeah.

这真的很奇怪。

It's it's really strange.

我跟多位客户进行了交流，也与其他一些分析师谈过。

Had conversations with with multiple customers, talked with some some other analysts.

我认为，至少从基本面来看，没有人能很好地解释为什么股价会下跌。

I I don't think that anyone has a a good explanation, at least fundamentally, as to why the stock's down.

人们谈论的理由都是在财报发布前就存在的潜在担忧。

The the reasons people talk to were all kinda concerns in the background heading into numbers.

比如，超大规模客户的资本支出在2027年前都会保持在这个水平吗？

So things like, CapEx at the hyperscalers remain at these levels into 2027?

到了明年，这些组件成本会不会出现调整？

Might there be a reset in these component costs when you get to next year?

但这些都不是新问题。

But none none of these are new issues.

对吧？

Right?

这些都是长期存在的担忧。

These are all the kind of concerns that have been lingering for some time.

我们真的要到今年年底才能得到答案。

We really won't have answers till till we get to the end of this year.

那关于客户集中度的担忧呢？

What about the concerns around customer concentration?

你有没有考虑过这个问题？

Is that is that something that you've been thinking about?

没有。

Not really.

我的意思是，他们确实说50%的收入来自大型超大规模客户。

I I I mean, certainly, they say 50% the revenues come from large hyperscale customers.

但这就是我们这个领域中IT支出的现实。

But that's that's the reality of I IT spend in in our world.

无论你是供应内存、硬盘，还是NVIDIA的加速器。

What whether you're supplying memory or hard drives or NVIDIA's case accelerators.

这些客户都是巨大的买家，你无法回避这一点。

Those customers just are huge buyers, and you can't get away from that.

那利润率呢？

What about the margins?

利润率在这里是个挺有意思的话题。

So the margins are kind of interesting story here.

他们的毛利率在百分之七十五左右。

Their gross margins are in the mid mid seventies in terms of percent.

你拿这个和AMD——其他最接近的公司相比，它们的毛利率在百分之五十五左右。

You compare that to AMD, some kind of the next best company, I mean, they're in mid fifties.

你觉得NVIDIA的这些利润率能维持多久？

How sustainable do you think those margins are for NVIDIA?

是的

Yeah.

所以人们说存在NVIDIA溢价，对吧？

So people say there's NVIDIA tax, Right?

这实际上意味着NVIDIA能够比其他公司收取更高的芯片价格。

That that effectively NVIDIA is able to charge more for their chips than than anyone else.

人们担心，如果竞争加剧，NVIDIA的毛利率最终会下降。

And it is a concern that if competition ratchets up, that at some point NVIDIA sees its gross margins come down.

真正的问题是，这种竞争会是什么样子？

The the question really is, so what is what does that competition look like?

如果你想想AMD这样的公司，他们的Helios机架上市时必须具有竞争力，但我们目前还不得而知。

If you think about someone like AMD, Helios racks have to be competitive when they come out, and and we just don't know yet.

对吧？

Right?

在我看来，Helios的前景仍不确定。

Helios is still to be determined in my mind.

如果你考虑的是类似定制ASIC的东西，比如谷歌提供的TPU，或者亚马逊告诉我的Traneum，还有Meta和微软也有自己的芯片，这些产品确实有可能更便宜。

If you're thinking about something more like custom ASICs, so what Google offers with its TPUs or what Amazon told me about Traneum, Meta and and Microsoft have their own parts, Certainly, there there's the potential for for those products to be cheaper.

但与此同时，ASIC的担忧在于，你是为特定工作负载而设计的，而人工智能是一个快速演变的领域。

At the same time, the concern with ASICs is that you are building them for specific workloads, and AI is a rapidly evolving beast.

那么，你是否会选错工作负载呢？

And so do you pick the wrong workloads?

而且，ASIC很难做。

Also, ASICs are hard.

对吧？

Right?

我们有一个公司，也就是谷歌，在这方面做得非常好，而很多其他公司则做得没那么出色。

We've got one company that seems to be doing them very well in terms of Google and a lot of companies that aren't doing as great a job.

所以，会存在竞争吗？

So could there be competition?

也许吧。

Maybe.

考虑到我所认为的非常有利的供应合同，尤其是在全球组件价格以疯狂速度上升的背景下，英伟达明年是否会遭受影响？

Could NVIDIA suffer next year when what I believe are very favorable supply contracts, at least in world where the backdrop is component prices are going up at crazy levels.

当这些合同到期后，它们的利润率是否会受到冲击？

When those contracts run out, could their margins be hit?

可能吧。

Maybe.

但同样，我们一直担心明年或后年会发生什么，而这次并没有什么不同。

But again, we've always had concerns around what happens next year or the year after, and that this is this is no different.

现在谈谈明年或后年会发生什么，感觉人们越来越少讨论折旧问题，以及这些芯片的合理折旧周期，还有客户如何考虑用新芯片替换旧芯片。

Now talking about what happens next year or the year after, it feels like people have been talking less about the question of depreciation and what the right schedule is to depreciate these chips and how customers are thinking about replacing old chips with the new chips.

我的意思是，随着这些芯片不断升级，有没有理由相信客户根本不会用新芯片替换旧芯片？

I mean, as these chips get better, is there any reason to believe that customers won't replace old chips with new chips at all?

我的意思是，你总会看到一些替换发生。

I mean, you always you always see some replacement.

对吧？

Right?

如果你想想超大规模云服务商，他们采用的是五年到六年的折旧周期。

And if you think about the the the hyperscalers, they have five year, six year depreciation schedule.

因此，人们默认服务器的可用寿命是五年或六年。

So there there's an assumption that five year, six years is kind of usable life for servers.

我之前担心的是，英伟达的技术演进太快，导致这些设备的可用寿命会被缩短。

I I I think the concern had been that NVIDIA was evolving so rapidly that that that usable life was gonna be shortened.

我们昨晚通话中听到科莉特说，H200芯片的利用率已经达到了极限。

We we heard Colette on the call last night say h two hundreds are being fully utilized as far as they can tell.

所以，这表明客户在设备的整个使用寿命内，都在充分使用这些产品。

So, you know, that that tells you that the customers are getting full use of these products, at least for their usable lives.

而且我还想说，在我自己的交流中，尤其是较小的客户，仍在大量采购H200，部分原因是它们比MBL 72系统更契合一些老旧数据中心的功耗特性。

And and I would also say that in my own conversations, particularly smaller customers are still buying a lot of h two hundreds, in part because they they fit the power profile of some these older data centers a whole lot better than the MBL 72 systems.

因此，我认为这一代产品仍然有巨大的需求。

And and so I think there's still a whole lot of demand for for that generation.

而且，如果需求依然存在，产品也仍在被广泛使用，那至少目前还不存在折旧问题。

And, you know, if there's still demand and they're still getting used, then there's not a depreciation issue at least yet.

你谈的是H200，不是在中国的销量，而是广义上的H200？

You're you're talking about h two hundreds, not the sales in China, but just h two hundreds broadly?

我谈的是美国和欧洲市场，那里仍然有人在购买H200。

I I'm talking broadly US, Europe that there are still buyers for h two hundreds.

再次说明，因为你在使用NVL72时会遇到一些问题。

Again, because some of the problems you run into NVL with NVL 72.

对吧？

Right?

它耗电量太大，一些老旧的数据中心根本无法支持。

It just has a it draws a ton of power that some older data centers, they don't they they they can't support that.

所以你可以部署H200，或者搭载H200的系统，通过出售代币来赚钱。

And so you can put an h 200 in, or a system with h two hundreds in and make money selling tokens.

因此，确实有一些公司在这么做，这说明H200仍然是一个非常重要的产品。

And so there are there are companies that are doing that, which tells you the h 200 is still a very relevant part.

说到H200，昨晚我们有没有更清楚地了解中国方面的贸易动态以及他们的预期？

Now speaking of the h 200, did we get any clarity last night on the trade dynamics with China and what they're expecting there?

所以英伟达表示，美国政府允许向中国出口少量芯片，或已批准了少量对华出口。

So Nvidia suggested that the US administration has is allowing a a small number of exports to China or has approved a small number of exports to China.

然而，中国尚未允许H200进入该国。

The China, however, is is not yet admitting h two hundreds, into that country.

我认为这仍然是一个政治谈判问题。

I I think it's it's it's still a political negotiation.

在某个时候，假设有某种结果出现的话。

At some point, presuming there there there is, some sort of result here.

如果英伟达能够重新发货，其收入可能会大幅增长。

Nvidia could potentially, see a a pretty big boost to revenues if they could ship again.

话虽如此，他们并未对此做出任何指引，而且我认为目前没有人真正预期会有任何进展，因为政治最终结果实在难以预测。

Having said that, they didn't guide for anything, and I I don't think anyone's necessarily expecting anything yet because it's just really hard to figure out what's what's gonna be the end game politically.

我想快速转个话题。

I wanna pivot very quickly.

我们之前在讨论AMD，也看到了新闻，称AMD正在收购Nutanix价值1.5亿美元的股份。

We were talking about AMD, and we saw the news also that AMD is buying $150,000,000 stake in Nutanix.

你能告诉我们Nutanix是做什么的，以及你对这笔交易的看法吗？

Can you tell us what Nutanix is and what you made of this deal?

是的。

Yeah.

Nutanix是一家超融合基础设施（HCI）提供商。

So Nutanix is an HCI provider.

他们-

They-

HCI是什么？给我们解释一下。

HCI break that down for us.

人们不太懂技术

People aren't as tech

懂行。

savvy.

是的。

Yeah.

不，没问题。

No no problem.

所以这是超融合基础设施。

So there it's converged infrastructure.

通常情况下，你的数据中心是由独立的组件组成的。

And so normally, you've got discrete pieces of your data center.

对吧？

Right?

你有一个存储阵列。

You've got a a storage array.

你有一台服务器。

You've got a server.

你有网络设备。

You you've got networking equipment.

超融合基础设施的作用就是将这些组件整合在一起。

What what HCI does is it is it converges that stack.

所以我使用一套软件，它同时管理存储、服务器以及计算资源。

So I I'm using one set of software, and it's managing both the the the storage in the server as well as the compute.

因此，我不必再去操心管理这些不同的组件，或者说管理数据中心里多个独立的系统。

And so I don't have to think about managing these these different entities, if you will, or manage or having multiple entities in my data center.

从AMD的角度来看，我并没有深入研究这笔交易，但我觉得他们在企业市场中仍然是一个相对较小的玩家，尽管他们在超大规模数据中心领域表现非常出色，但在企业市场的渗透率仍然较低。

I I I think from an AMD perspective, and I didn't look too closely at the at the deal, but that they are still a a relatively smaller player in the enterprise, with their compute solutions, where they've done a great job, penetrating hyperscalers, but but less, they have less penetration in the enterprise still.

这为他们提供了一个强大的合作伙伴，有助于在计算领域获得更多企业客户的认可，而我认为他们在服务器计算方面依然拥有巨大优势，因为他们能够设计和销售这些高核心数的CPU。

This gives them a a powerful partner, to work with in terms of getting some more enterprise traction, on the compute side where I I think they still have a huge advantage, in server compute given their their ability to to build and and sell these these really high core count CPUs.

所以对你来说，这本质上是一种客户拓展策略，旨在帮助AMD触达这些企业客户。

So so this for you is is really a a customer expansion strategy to get to those enterprise customers for AMD.

你就是这么看待这件事的。

That's that's how you're thinking about it.

看到昨天的新闻头条时，我第一反应就是这样。

Seeing the news headlines hit yesterday, that was my first thought.

很好。

Great.

好吧，马特，感谢你前来做客。

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

这位是来自韦德布什的马特·布赖森，正在TI TV上。

That is Matt Bryson from Wedbush here on TI TV.

昨天晚上，Salesforce和Snowflake也公布了季度业绩。

Salesforce and Snowflake also reported quarterly results last night.

Salesforce的收入增长了12%，较上一季度有所加速，不过部分得益于其对Informatica的收购。

Salesforce grew revenue 12%, an acceleration from the last quarter, although that was aided in part by its acquisition of Informatica.

Snowflake的收入增长了30%，较上一季度有所加速。

Snowflake revenue grew 30%, an acceleration from last quarter.

我想邀请我们的金融分析专栏作家阿尼塔·拉马斯瓦米，帮我们解读这些业绩。

I want to bring on our financial analysis columnist Anita Ramaswamy to help us break down those results.

阿尼塔，欢迎再次回到节目。

Anita, welcome back to the show.

很高兴你来到这里。

It's great to have you here.

能来这里总是很好，阿卡什。

Always great to be here, Akash.

另一个充满财报的忙碌早晨。

Another busy morning of earnings.

昨晚我们除了英伟达，还有Snowflake和Salesforce发布了财报。

So we had Snowflake and Salesforce last night in addition to Nvidia.

从Salesforce和Snowflake的财报整体来看，有什么特别吸引你的地方？

What stood out to you from the Salesforce and Snowflake results taken together?

是的，确实是个忙碌的早晨。

Yeah, definitely a busy morning.

尤其是软件投资者，最近有很多关注点。

There's been a lot on software investors' minds in particular.

我认为，这两只股票都因市场对AI可能如何颠覆软件行业的负面情绪而受到重创。

I think that both of these stocks have been hit hard by the sort of negative sentiment around how AI is going to potentially disrupt the software sector.

因此，尽管它们的财报整体表现良好，但在发布后股价还是略有下跌。

And so we saw both of their stocks after they reported earnings, despite earnings overall being good, trade down a little bit.

它们今天早上都开始回调了，但我认为我对这两家公司的看法存在分歧。

They both since started correcting this morning, but I think that there's a divergence in how I see these two companies.

根本原因在于，Snowflake 更多处于基础设施层，它是一家帮助其他公司整理数据以运行 AI 工作负载的公司，而 Salesforce 则更专注于面向客户的应用程序。

And that fundamentally is because Snowflake is playing more at the infrastructure layer where they're a company that helps other companies get their data in order to run AI workloads, whereas Salesforce is more so focused on the customer facing app.

这一点最明显地体现在两家公司的营收增长率上。

And I think that's most obviously reflected in the top line growth rate at the two.

就像你提到的，Akash，Snowflake 的年增长率大约为 30%。

Like you mentioned, Akash Snowflake is growing around 30% year over year.

Salesforce 本季度的增长率接近 12%，这实际上是其过去几年来的最快增速。

Salesforce is growing closer to 12% as of this quarter, which was actually their fastest growth rate in the last several years.

我想问问你，AI 是如何以不同方式影响这两家公司的。

I want to ask you about how AI is affecting both these businesses differently.

你刚才已经稍微提到了，但我还想回顾一下 Martin Pierce 昨晚在简报通讯中的评论。

You kind of hinted at it here, but I'm going back to some comments that Martin Pierce wrote in his briefing newsletter last night.

他提到，AgentForce 的年度经常性收入（ARR）仅占 Salesforce 2027 财年预计总收入约 460 亿美元的 1.7% 左右，所以这里的关键是，AgentForce 虽然在增长，但仍是较小的一部分。

He wrote that AgentForce's ARR, for instance, is just about 1.7% of Salesforce's total fiscal twenty twenty seven projected revenue of about $46,000,000,000 So I mean, the story here, Agent Force is growing, but it's still a small part.

Snowflake，如你所说，它属于基础设施层。

Snowflake, like you said, it's an infrastructure layer.

也许他们从中受益更多一些。

Maybe they're seeing a little bit more of a benefit.

我的意思是，这里的情况有所不同吗？

I mean, is the story different here?

我的意思是，确实有一些差异，但即使是Snowflake，也很难准确判断其收入增长的加速中有多少是源于AI。

I mean, there's some difference, but even with Snowflake, it's hard to tell exactly how much of their acceleration in revenue growth has been as a result of AI.

他们公布了客户数量，看起来增长得非常不错。

They've disclosed the number of customers that they're seeing, and that seems to be growing really nicely.

他们还在本季度表示，签下了有史以来最大的一笔交易，金额约为4亿美元。

They also said this quarter at Snowflake that they signed their largest deal ever, which was around $400,000,000 worth.

他们至少部分将这笔交易归因于他们的AI套件。

They attributed that at least partly to their AI suite.

因此，看来Snowflake正在通过向客户销售AI产品而受益。

And so it seems like Snowflake is benefiting from selling AI products to customers.

我觉得关于Snowflake的一个有趣点是，他们的自由现金流利润率已经从一年前的43%提升到了最新季度的61%。

I thought one interesting point on Snowflake is that their free cash flow margin has actually gone from 43% a year ago to 61% in the latest quarter.

这表明，即使他们在AI上持续投入，仍能以更快的速度产生现金流。

So that kind of shows that even as they invest in AI, they have been able to continue to generate cash at a faster rate.

不过，我认为无论是Snowflake还是Salesforce，它们都是规模庞大的公司。

I think with both Snowflake and Salesforce, however, they're huge companies.

正如你所指出的，AI产品，尤其是在Salesforce，目前仅占其整体收入的一小部分。

And like you pointed out, the AI product, especially at Salesforce, pretty is small portion of their overall revenue.

尽管其年增长率超过100%，但我们仍不清楚这会对未来的收入增长产生多大的实质性推动作用。

Even though it's growing over 100% year over year, we still don't really know how much of a meaningful catalyst that's going to be to top line growth going forward.

另一个值得注意的重要点是，Salesforce在2025年11月完成的对Informatica的收购，对其增长起到了显著的推动作用。

And I think another important point to note is that the acquisition of Informatica for Salesforce, which closed in November 2025, was a meaningful accelerant and driver of their growth.

因此，我认为还有待观察，Akash，这两家公司能否持续保持有机增长。

So I think it remains to be seen, Akash, whether the organic growth for both of these companies can continue.

一些投资者对Snowflake的一个负面看法是，他们公布了2027财年的收入指引，其财年截止于1月。

The one negative point on Snowflake that some investors were considering is that they gave their revenue guidance for their fiscal twenty twenty seven, which their fiscal year ends in January.

Snowflake 暗示其收入增长率将放缓至约 27%。

And Snowflake essentially implied that their revenue growth would decelerate to around 27%.

尽管这一增长率对于其主要收入来源的产品收入来说仍然非常快，但相比之前几个季度已经有所放缓。

Now that's still a really fast growth rate for their product revenue, which is the majority of their top line, but it's slower than it was in previous quarters.

因此，这可能表明 AI 正在以某种更根本的方式对这两家公司的业务构成实质性挑战。

And so that could be a sign that there is something more fundamental about AI that is challenging both of these businesses in a meaningful way.

很好。

Great.

Anita，我想换个话题，谈谈你今天发表的关于其他更大科技公司的文章，这些公司甚至比 Salesforce 和 Snowflake 还要大。

Anita, I want to pivot and talk about a story that you published today on other big tech companies, bigger tech companies, even then the sales forces and the snowflakes.

你写了关于超大规模云服务商的内容，比如像 Meta 这样的公司。

You wrote about the hyperscalers, even the, you know, the metas of the world.

有趣的是，我们通常不把 Meta 称为超大规模云服务商，但我还是忍不住把它和它们归为一类。

It's funny, we don't call meta a hyperscaler, and yet I still seem to conflate it with it.

我知道有人会说别把这两者混为一谈，我也没打算这么做，但要讨论这些大型科技公司及其所承担的债务及其对信用评级的影响，确实很难避开这一点。

I know it's people are gonna say, you know, don't conflate the two, and I'm not suggesting it, but it's kind of hard to talk not about Anyway, that in this you wrote about these big tech companies and the debt that they're taking on, and how it affects their credit profile.

你开始写这个故事时，带着哪些问题？

What were the questions that you went into this story with?

我写这个故事时，想知道这些公司为何在最近几个月发行了这么多债务。

So I went into this story wondering about how all of these companies have issued debt in recent months.

我们在十一月看到了两次债务融资。

We saw two debt raises in November.

我们还看到了一次比那更近期的债务融资。

We saw a debt raise more recently than that.

具体来说，我指的是谷歌、亚马逊和Meta，也许有人会说，如果它们真的大力投资于人工智能计算，未来某天或许会成为超大规模云服务商。

From specifically, I'm talking about Alphabet, Amazon, and Meta, which I guess one should argue if they really invest in meta compute, maybe they will be a hyperscaler someday.

但那是另一个话题了，改天再聊。

But that's another conversation for another day.

我写这篇文章真正想解答的问题是：这些公司要到什么程度才会真正面临资本成本上升？

The question that I was really trying to answer with this piece is what will it take for these companies to actually see an increase in their cost of capital?

它们正在发行越来越多的债务。

They're issuing more and more debt.

看起来它们仍在持续将越来越多的资本支出投入人工智能投资。

It seems like they're continuing to plow more and more CapEx into their AI investments.

这些投资越来越依赖债务市场融资，而不仅仅是股权市场。

Increasingly that's being funded in the debt markets, not just the equity markets.

因此，我想研究一下，这些公司是否可能被信用评级机构下调评级，这会带来什么潜在影响。

And so I wanted to look at what the potential impact could be of, could they get downgraded by the credit rating agencies?

它们离达到那个临界点还有多远？

Are they even anywhere close to reaching that point?

还有没有其他因素可能长期显著提高它们的资本成本？

And are there any other factors that could meaningfully increase their cost of capital for the long term?

那你最终发现了什么？

And so what did you ultimately find?

所以对于这些大型科技公司的投资者来说，好消息是，信用评级下调似乎还很遥远。

So the good news for investors in these big tech companies is that it looks like a credit downgrade is pretty far off.

这些公司的临界点大约在债务与EBITDA比率为1倍左右。

The threshold for these companies is around one times debt to EBITDA.

到2026年，所有这些公司——至少根据我与标普分析师的交流——预计债务与EBITDA比率将降至0.5倍以下，大约在2.3倍左右。

And all of them by the 2026, at least when I talk to S and P analysts are projected to be closer to below 0.5 times, let's say, point 2.3 around there.

因此，它们还有很大的举债空间。

And so they have a lot more capacity.

我的意思是，只要这些公司大致达到分析师预期的EBITDA水平，它们就能再增加约2000亿美元的债务。

I mean, of these companies can take on about $200,000,000,000 in incremental debt, as long as they generally meet the projected EBITDA that analysts expect them to make.

当人们计算这些信用评级时，是否只考虑它们资产负债表上的债务总额？

And when people are calculating these credit ratings, is it only the amount of debt that they have on their balance sheet that they take into consideration?

在得出这些评级时，他们还会考虑哪些其他因素？

What are the other factors that they consider as they're coming up with these ratings?

是的，这其实是我在撰写这篇报道过程中了解到的一个有趣点，当时我与这些信用分析师进行了交流。

Yeah, so this is actually something interesting that I learned as I was reporting out this piece, was talking to these credit analysts.

他们不仅考虑信用指标，比如现金流能否覆盖债务以及未来趋势如何，还会考虑一些与业务相关的战略因素。

They're not only taking into account the credit metrics, can the cash flow cover the debt and what that is going to look like going forward, but also some strategic considerations in terms of the business.

如果看这三家公司，Meta实际上是其中信用评级最低的。

If you take all three of these companies, Meta actually has the lowest credit rating of all of them.

据这些信用分析师告诉我，这在很大程度上是因为他们的业务严重依赖广告收入，而另外两家公司的业务更加多元化。

That is largely a function I've been told by these credit analysts of the fact that their business is really dependent on advertising revenue, whereas the other two companies are better diversified.

因此，这带来了一定程度的额外风险。

And so that sort of comes with a little bit more of incremental risk.

在我的一些访谈中，这些分析师提到，他们设定的降级门槛是杠杆率一倍，但如果这些公司未来出现现金流低迷的情况，他们甚至可能重新评估这一门槛。

Some of my conversations, these analysts were saying that they had this downgrade threshold of one times leverage, but that they could even potentially revisit that if these companies show depressed cash flows in the future.

所以在三家公司中，我认为Meta目前看起来风险最高，主要是基于其信用评级。

And so of all three of the companies, I would say Meta's probably, you know, looks the riskiest right now just based on their credit rating.

这并不仅仅取决于他们当前的现金流和债务状况。

And that's not just a function of their current cash flow and debt situation.

这在很大程度上取决于他们未来业务的前景以及分析师如何看待这一前景。

It is very much a function of what the future of their business could look like and how analysts are perceiving that.

你写的另一个让我感兴趣的部分是，你谈到这些分析师如何看待对这类债务的需求，以及这些公司——我是说，这些是蓝筹科技公司。

Now, the other piece that you wrote about that I found interesting is you talk to these analysts about how much demand there is for this debt and the fact that these companies are I mean, these are blue chip tech companies.

我一直在想，显然，对这类债务的需求来自于其信用评级本身。

And I found myself thinking about, well, clearly the demand for the debt comes from the credit rating itself.

不过我在想，也许是否存在一种情况，即信用评级其实并不重要。

Although I wondered if maybe there was a bit of a dynamic in terms of maybe it doesn't matter what the credit rating is.

只要市场对这家公司有需求，这些公司就可以持续发行债券。

If there's demand for this company at all, then the companies can just keep issuing it.

是的。

Yeah.

我认为信用评级对债券的需求有一定影响，但更多是相关性，而非直接的因果关系。

I think that the credit rating has some bearing on the demand for the debt, but it's more correlated than directly causation linked.

这是因为还有其他因素在起作用。

And that's because there are other factors.

投资者将信用评级视为衡量这些企业财务健康状况的一个指标，但他们也在关注许多其他因素。

Investors are looking at the credit rating as one measure of the financial health of these businesses, but they're also looking at a number of other factors.

这类似于你评估一家公司的股票分析师报告时的做法。

It's similar to how you would evaluate an equity analyst report perhaps on a company.

我认为供需动态在这里是一个非常重要的因素。

And I think that supply and demand dynamics are a really big factor here.

比如，如果这些公司持续发行越来越多的债务，最终导致市场供应过剩，而目前我们看到的是，对这些超大规模云服务商债务的需求非常旺盛。

Like if all these companies continue to issue more and more debt and they eventually flood the market with supply, right now we're seeing that there's a lot of demand for this hyperscaler debt.

所有这些债务发行都出现了超额认购。

Like all these debt offerings have been oversubscribed.

它们成功筹集到了想要的资金，但若它们向市场注入过多债务，这种局面可能会发生变化。

They've been able to raise the amount of money that they want to raise, but it's possible that if they flood the markets with so much debt, that that picture changes.

此外，投资者对这些企业基本面的信用状况的看法，也可能与信用评级机构的评估大相径庭。

And there's also the possibility that investors end up seeing the underlying credit health of these businesses quite differently from how the credit ratings agencies are seeing them.

因此，我认为债务端资本成本上升的潜在影响，可能通过多种途径波及这三家公司中的任何一家。

So I think there's a lot of different ways that a potential cost of capital increase on the debt side could flow through to any of these three companies.

但我认为信用评级是一个非常重要且有意义的指标，投资者会密切关注。

But I think credit ratings are a really big, meaningful indicator that investors are going to be looking at.

很好。

Great.

阿尼塔，非常感谢你前来参与。

Anita, I want to thank you for coming on.

这是我们在《The Information》的金融分析专栏作家阿尼塔·拉马斯瓦米。

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

另一家国防公司正在筹集另一轮巨额融资。

Another defense company is raising another big funding round.

Ceronic公司正在筹集高达15亿美元的资金，估值为75亿美元，本轮融资由凯雷资本领投。

Ceronic is raising up to $1,500,000,000 at a $7,500,000,000 valuation in a funding round that Kleiner Perkins will lead.

这来自我的同事、财经副总监科里·温伯格的独家报道。

That is according to an exclusive report from my colleague, Corey Weinberg, our Deputy Bureau Chief of Finance.

我想请科里来详细谈谈这笔交易。

I want to bring on Corey to talk more about this deal.

科里，欢迎再次回到节目。

Corey, welcome back to the show.

很高兴你来这里。

It's great to have you here.

嘿，科赫。

Hey, Koch.

Cyronic 是什么公司？

What is Cyronic?

我从来没听说过这家公司。

I never heard of this company.

是的。

Yeah.

如果你没有花大量时间关注试图进入美国市场的公司，可能就不会注意到他们在我们的水域里活跃。

Cyronic well, if you haven't spent a ton of time sort of in the trying to to come onto the shores of The United States, then maybe maybe you wouldn't have swimming in our in our waters here.

没有。

No.

我的意思是，Cyronic 是一家海军舰船公司。

I mean, Shronic is a is a naval ship company.

他们正迅速成为国防科技领域融资最多的初创公司之一。

They are quickly becoming one of the best funded startups in defense tech.

他们主要建造军舰和海军舰艇，并试图将它们卖给美国海军。

And essentially, what they do is that they build sort of warships and naval vessels and try to sell them to the US Navy.

好的。

Okay.

但你刚才说这些是无人作战舰艇？

But you said in the story these are autonomous warships?

是的。

Yeah.

它们通常是无人的。

They're they're unmanned, generally.

所以Cyronic将他们的自主软件与舰艇捆绑在一起。

So Saronic is bundling sort of their autonomous software with the ships.

这将是美国海军未来的重要优先事项。

That's a big future priority for the US Navy.

好的。

Okay.

所以美国政府现在是Cyronic的唯一客户吗？还是他们还卖给其他人？

So the US government, are they the only customer for Ceronic right now, or who are they selling to?

据我所知，他们是唯一的重大客户。

They're the only significant customer to my knowledge.

我认为，他们也希望像许多国防科技公司一样，向美国的盟友销售产品。

They, I think, aspire to also sell like a lot of sort of defense tech companies to US allies.

有时候，说起来容易做起来难。

Sometimes that's easier said than done.

你知道，各国通常更倾向于从本国公司购买或购买技术。

You know, countries tend to want to to sell or or excuse me, to buy technology from homegrown companies.

所以这可能比较困难，但他们也目标是向美国盟友销售产品。

So that can be difficult, but they aim to to sell to US allies as well.

而且他们也可能，你知道，未来某天可以将软件商业化销售。

And they also, you know, could have a commercial, you know, sort of could sell their software commercially at some point down the road.

他们可能不得不这么做。

And they might need to.

我的意思是，他们现在估值很高。

I mean, they have a big valuation now.

这可能高达90亿美元。

It's potentially up to $9,000,000,000.

那么，我们对这家公司的收入状况了解多少？

And what do we know about the revenue profile for the company?

我们在报道中提到，他们去年的收入刚刚超过2亿美元。

We reported in the story that they generated just over $200,000,000 in revenue last year.

可以说，投资者期望在未来几年看到显著的增长。

It's fair to say investors are expecting a hefty amount of growth in the coming years.

如果他们未能实现增长，那么Cyronic的估值就会显得严重高估，因为你知道，根据上一年的业绩来看，这个倍数已经相当高了。

And if they don't get it, then Cyronic will look wildly overvalued because, you know, that's that's not a that's that's a pretty significant multiple looking at the previous year.

但话说回来，这些可是巨大的项目。

But, look, these are really big ships.

他们试图以高价出售，而投资者则押注海军会购买它们。

They're trying to sell them for a lot of money, and investors are betting that Navy's going to buy them.

好的。

Okay.

所以，你知道，你提到的一个有趣之处是，克莱纳正在主导这笔交易。

So, you know, one of the interesting parts that you had, Nishwari, is that Kleiner is leading this deal.

克莱纳是大型国防科技投资者吗？

Is Kleiner a big defense tech investor?

这是他们在拓展的一个业务领域吗？

Is this a practice that they're building out?

我的意思是，我们是否正看到风投机构开始建立这类业务板块的趋势？

I mean, is this a trend that we're seeing that VCs are starting to build out these practices?

因为我不难想象，投资者要理解这些内容需要多少专业技术知识。

Because I can only imagine how much technical expertise must be required for an investor to understand any of this.

是的。

Yeah.

这真的很有趣，阿卡什。

It's really interesting, Akash.

克莱纳·伯金斯此前从未参与过重大的国防科技交易。

Kleiner Perkins hasn't done a major defense tech deal before.

这将来自他们的成长基金。

This will be out of their growth fund.

你知道的。

You know?

所以这是针对后期公司的，设计上就是给后期公司用的。

So it's late it's supposed to it's designed for later stage companies.

Kleiner Perkins 投资过 Flock Safety，这家公司显然向警察部门销售车牌识别摄像头，但和国防科技略有不同。

Kleiner Perkins has invested in flock safety, which obviously sells sort of license plate scanning cameras to police departments, but slightly different than defense tech.

没有。

No.

我很惊讶。

I was surprised.

我们至少几周前就听说 Cyronic 在四处寻求投资兴趣。

We had been hearing about this round for, at least for weeks that that Cyronic was out there trying to drum up interest.

我原本以为，会是那些把国防科技作为核心宣传重点的风投公司——比如 Andreessen Horowitz、Founders Fund 或 AVC 之类的——来主导这轮融资。

And I personally expected either one of the VC firms that has really, you know, sort of put defense tech front and center of their whole message to to kind of go out and and try to lead this round, you know, like an Andreessen Horowitz or Founders Fund or AVC or something like that.