Anthropic 正洽谈使用微软 AI 芯片，SpaceX IPO 文件披露重大信息

欢迎各位收看《信息》频道的TITV节目。

Welcome everyone to The Information's TITV.

我是阿卡什·佩里特沙。

My name is Akash Pesritcha.

今天是5月21日，星期四。

It is Thursday, May 21.

今天开场，我们带来一条独家报道：信息平台刚刚披露了Anthropic和微软正在洽谈一项新的芯片合作。

We are kicking off the show today with some exclusive reporting that the information published around talks of a new chip deal between Anthropic and Microsoft.

接着我们会聊聊英伟达那组惊人的财报。

We'll then get to the monster Nvidia results.

这家公司刚刚发布了季度业绩。

The company posted its quarterly results.

它预计下一个季度还会再创佳绩。

It is expecting another big quarter ahead.

稍后我们会请一位分析师来详细解读。

We'll break that down with an analyst very shortly.

SpaceX的S-one终于来了。

The SpaceX S-one is finally here.

数据已经出来了。

We have the numbers.

接下来我们来分析一下。

We are going to get to the analysis.

我们马上请来罗斯·格伯，节目最后还会带来另一篇独家报道，这次是关于白宫计划在先进AI模型公开发布前进行审查的事。

We are bringing on Ross Gerber very shortly, and we will wrap the show with some more exclusive reporting, this time around the White House's plans to review advanced AI models before they get released to the public.

今天内容很满，咱们直接开始。

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

《The Information》独家披露，Anthropic正寻求与微软达成芯片合作协议。

The Information published exclusive reporting that Anthropic is looking to make a chip deal with Microsoft.

这篇报道是我同事亚伦·霍姆斯和刘昌纳写的。

My colleagues Aaron Holmes and Channer Liu wrote that story.

现在我们请亚伦来分享他了解到的情况。

Aaron joins us now to share with us what he knows.

艾伦，欢迎再次回到节目。

Aaron, welcome back to the show.

关于这笔潜在的芯片交易，我们目前知道些什么？

What do we know about this potential chip deal?

Anthropic 正在和微软进行初步洽谈，计划使用微软自主研发的 AI 芯片，代号叫 Maya 200。

So Anthropic is in early talks with Microsoft to use Microsoft's own custom AI chips, which are called Maya two hundreds.

目前还只是初期讨论阶段。

And the talks are early.

还不确定最终会不会真的达成合作。

It's not clear if this is going to actually lead to a commitment.

但如果他们真的决定租用微软的芯片，这对微软的自研芯片项目来说将是重大突破——他们正努力让这个项目成为能和英伟达，还有谷歌、亚马逊以及其他 AI 领域玩家的自研芯片相抗衡的重要力量。

But if they, you know, do decide to to rent the chips from Microsoft, this would be a huge win for Microsoft's, in house chip effort, which it's trying to to, you know, prop up as a meaningful competitor to Nvidia as well as to the in house chip efforts of Google and Amazon and and other AI players in the space.

我们之前在节目里聊过很多种芯片。

So we've talked a lot about a lot about we've talked a about a lot of chips on the show.

但还没专门聊过 Maya 200 这款芯片。

We have not talked about the Maya 200 before.

那这个Maya 200到底是什么？

What what is the Maya 200?

这是微软自研AI芯片的最新版本。

So this is the latest version of Microsoft's in house AI chip.

微软已经研发Maya芯片好几年了，这个版本还多次推迟了发布。

Microsoft has been working on Maya for several years, and this version specifically was delayed a few times.

像英伟达的芯片通常既用于训练新模型，也用于运行现有AI应用，但Maya 200主要专注于运行AI，也就是所谓的推理任务。

And, you know, a a lot of AI chips like Nvidia is are used for both training new AI models as well as running existing AI applications, but the Maya 200 is more focused on just running AI, which is also known as inference.

微软今年早些时候提到，他们已经开始内部使用Maya芯片来运行Copilot工具，发现有些任务用Maya的成本比用英伟达芯片更低。

And Microsoft has said earlier this year that it started using Maya internally to to run its Copilot tools and that it found that it was cheaper to run Copilot using Maya for some tasks than Nvidia chips.

如果微软能让Anthropic这样的客户相信，用Maya跑模型比用英伟达芯片更省钱，那就能帮Anthropic获得更多算力——而他们现在正迫切需要更多算力，因为用户需求暴涨。

So, you know, if it can convince customers like Anthropic that they could also run their models more cheaply on Maya than on Nvidia chips that could help Anthropic unlock more compute, which it desperately needs right now because of all of the surging customer demand that Anthropic is seeing.

我们有没有办法判断，这款芯片跟谷歌的TPU或者亚马逊的Trainium相比，是更强、更弱，还是差不多？

Do we have any sense around whether or not this chip is better or worse or comparable to Trainium or the TPUs that Google has put out there?

它跟谷歌的TPU和亚马逊的Trainium相比，还处于非常早期的阶段。

It's definitely a lot more nascent than Google's TPUs and Amazon's Trainium.

你知道吗，目前Maya除了微软自己之外，还没有其他大客户。

You know, Maya, as of now, doesn't have any big customers other than Microsoft itself.

而谷歌已经成功说服了OpenAI、Anthropic和Meta这些公司使用TPU，亚马逊也说过，一些主要的AI实验室已经在使用Trainium。

Whereas, you know, Google has already been able to convince companies like OpenAI, Anthropic, and Meta to use TPUs, and Amazon has similarly, you know, said that some major AI labs are using Tranium.

总的来说，Nvidia的GPU依然被视为行业黄金标准，但具体效果还得看你要做什么任务。

You know, generally across the board, Nvidia is still seen as the gold standard for GPUs, but it it kind of depends exactly which type of task you're trying to use them for.

如果能拿下Anthropic这样的大客户，对微软来说意义重大，能有力证明这些芯片确实是靠谱的，真正适合大型AI公司使用。

And think that winning a big customer like Anthropic would go a long way in terms of Microsoft's efforts to prove that these chips are the real deal and are useful to major AI customers.

你听到这个消息的时候感到意外吗？

Were you surprised about this when you heard this?

我觉得咱们先暂时把Maya 200放一边。

Because I feel like we just Put the Maya200 aside for a second.

芯片这个领域，我一直觉得微软很少被提到。

Chips were an area that I think we just haven't really talked about Microsoft around that much.

我一直隐隐觉得，大家都在往芯片领域进军。

And I think there was always a hunch that everybody's getting into the chip game.

大家都想往垂直整合的方向走。

Everyone wants to be a little more vertically integrated.

微软 probably 早就该进这个领域了。

Microsoft probably would have gone into it.

但我觉得，如果他们真能拿下Anthropic这个单子，长远来看，可能会成为他们芯片事业的一个重要转折点。

But it strikes me that if they get this Anthropic deal, this could mark a pretty significant turning point for its own chip ambitions in the long run.

是的。

Yeah.

说真的，看到这么大的AI实验室可能用Maya芯片，确实有点意外，毕竟像你说的，这种事还没发生过。

I mean, on the one hand, it is maybe a little surprising to see a big lab potentially adopt the Maya chips, just because, to your point, we haven't seen that happen yet.

但另一方面，Anthropic一直秉持着中立策略，明确表示只要能拿到的芯片，他们都愿意尝试。

But at the same time, Anthropic has actually always had this kind of agnostic strategy where they have, you know, signaled that they're very willing to use any chips that they possibly can get their hands on.

他们不仅是最早采用谷歌TPU的公司之一，也早早用了亚马逊的Tranium，再加上一直用的英伟达GPU。

They were one of the first adopters of Google's TPUs as well as Amazon's Tranium in addition to the Nvidia GPUs that they've always used.

他们现在甚至还在和一些新兴初创公司谈，考虑用它们的新芯片。

They're even in talks with, you know, newer startups about potentially using chips from those companies.

所以，Anthropic 一直秉持着多元化策略，不想只依赖单一架构。

So Anthropic very much has this philosophy of trying to diversify and not just rely on one architecture.

从这个角度看，他们去探索微软的芯片也就不奇怪了。

And so, in that sense, it makes sense that they might explore what that would look like from Microsoft.

没错。

Right.

这其实也正是我想表达的意思。

And that's kind of what I was trying to get at.

这到底是反映微软芯片本身有多强，还是反映 Anthropic 实在太缺芯片，或者太想分散供应链了？

It was, to what extent is this a reflection of how good Microsoft's chip is and to what extent is this a reflection of just how desperate Anthropic is or maybe how eager they are to diversify their chip supply.

这种权力关系里，一直都有个平衡。

I mean, there is always a balance in that power dynamic.

我想再深入聊聊 Anthropic 和微软之间更广泛的合作关系。

I want to dig into though the broader relationship between Anthropic and Microsoft.

这和他们之前签的云服务协议有什么关联？长远来看，会不会影响到这些合作？

So, how does this fit into the cloud deals that they have signed together and whether this could have an impact on that in the long run?

是的。

Yeah.

过去一年里，Anthropic 和微软的关系越来越紧密。

So Anthropic and Microsoft have been getting closer and closer over the past year.

一开始，OpenAI 和微软是绑定在一起的，他们签了独家协议，微软成为 OpenAI 唯一的云服务商，还获得了 OpenAI 模型的使用权。

You know, originally, OpenAI and Microsoft were joined at the hip, through an exclusive deal that made Microsoft OpenAI's exclusive cloud provider and gave Microsoft the rights to OpenAI's models.

但今年早些时候，他们把这个独家协议解除了。

But they opened up that deal earlier this year.

现在，OpenAI 开始和其他云服务商合作，同时微软也在和 Anthropic 越走越近。

So now, you know, OpenAI is striking deals with other clouds, and we're also seeing Microsoft get closer to Anthropic.

而且就在过去六个月左右，Anthropic 已经成了 Azure 的大客户。

And at the same time, Anthropic has become a big customer of Azure in the last, you know, six months or so.

他们在 Azure 上的投入一直在快速增长。

It has been ramping up its spending on Azure.

我听说，Azure 内部团队一直在专门为 Anthropic 扩容，甚至明确说：这部分资源要留给 Anthropic，哪怕其他大企业客户再着急，也不能挪给他们。

I'm told that, you know, inside of Azure, folks have been actively setting up new capacity for Anthropic and essentially saying, we we need this capacity for Anthropic, we can't give it to other customers, even some very big enterprise customers.

这很合理。

And it makes sense.

我的意思是，微软想证明，Azure才是AI实验室运行模型的首选平台。

I mean, Microsoft wants to show that Azure is the the place for AI labs to run their models.

目前来看，Anthropic在AI领域已经遥遥领先。

And at this point, Anthropic is the front runner in the AI race.

所以，微软肯定有动力去维系并进一步发展和这家客户的合作关系。

So it's definitely in Microsoft's interest to to keep, growing and nurturing that customer relationship.

很好。

Great.

Aaron，非常感谢你今天来做客。

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

这位是Aaron Holmes，我们《The Information》的微软领域记者。

That is Aaron Holmes, our Microsoft reporter here at The Information.

英伟达上个季度业绩大爆发，营收增长了85%。

Nvidia had a blowout quarter with revenue climbing 85%.

这家芯片巨头还预测，本季度的营收将增长95%。

The chip giant also projected 95 percent growth in the current quarter.

这将是连续第四个季度实现营收增速加快。

That would be the fourth consecutive quarter of top line acceleration.

我们来详细分析一下，我请来了汇丰银行全球科技硬件与半导体研究主管李凡。

Break it all down, I want to bring on Frank Lee, Global Head of Technology Hardware and Semiconductor Research at HSBC.

弗兰克，欢迎再次做客我们的节目。

Frank, welcome back to the show.

很高兴你能来。

It's great to have you here.

很高兴回来。

All Good to be back.

好的。

Okay.

昨晚公布的业绩中，有什么特别让你印象深刻的地方吗？

So what stood out to you from the results last night?

是的。

Yeah.

我觉得这次财报发布前的预期，跟上个季度很相似。

So I think going into the print, I think it was very similar setup as last quarter.

我们原本预计，他们的实际业绩会超出市场普遍预期，而且还会上调指引。

We expected we'd probably see a beat and raise relative to where most consensus numbers was.

但与此同时，几天前我们也提到，英伟达可能需要更多新的故事来支撑市场情绪。

But at the same time, I think we also highlighted, I think, a few days ago that Nvidia would probably need to see more narratives emerging.

光靠盈利势头和下一代AI芯片的路线图，可能已经不够了。

I think earnings momentum and, you know, the the the next generation AI GP road map wasn't really gonna be enough.

他们真的得开始推动几个新的叙事方向。

I think they really need to start focusing on a couple of new narratives.

我觉得他们已经开始这么做了，比如重新分类数据中心收入，这可能是个重要转折，因为他们现在把超大规模客户和非超大规模客户的收入拆开了。

And and I think they've started to do that with the reclassification of their data center revenue, which we think is is a is a potentially big event because now they're breaking down the hyperscaler and non hyperscaler revenues.

对。

Yeah.

那我们来聊聊这个重新分类的事。

So I want to go to that reclassification here.

听起来你觉得这是个明智的举措，而且你支持。

So it sounds like you think that that was a smart move and you support it.

是的。

Yeah.

如果你回想一下去年的叙事，贾斯汀·黄早就提过企业级和主权AI，但这些一直没在财务数据里体现出来，因为他们没把它们分开统计。

I think if you think about the narrative going back to last year, you know, Jensen Huang's talked about the enterprise and the sovereign AI for quite some time now, but it hasn't really quite materialized or very obvious in their numbers because they haven't really broken that down.

现在他们把云服务提供商和非云服务提供商的收入拆开了，市场就能更清楚地看到他们在非云服务这块的表现了。

And so I think now that they've broken down the difference between, you know, CSP and non CSP revenues, it'll be easier for the market to really track how they're doing on a non CSP.

因为如果你看的话

Because if you look

云服务提供商就是那些大型数据中心运营商。

at CSP being for those

超大规模客户。

the hyperscodes.

对。

Yeah.

超大规模服务商。

Hyperscodes.

就是云服务提供商。

The the cloud service providers.

没错。

Yeah.

因为今年我们看到的一个明显问题是，超大规模资本支出现在被分摊到了多个厂商身上。

So because if you look at one of the issues we're seeing this year is that it's more obvious is that the hyperscale CapEx is now being divided by multiple people.

除了英伟达，内存厂商现在也开始分走这部分资本支出。

So besides Nvidia, you know, memory guys are now eating into this CapEx.

你还要为网络设备和光通信供应链买单。

You're also paying for the networking, the optical supply chain.

还有服务器CPU，现在也成了一个非常重要的方向，尤其是随着AgenTek AI的发展。

You're also looking at, you know, the the server CPU now becoming a very big theme, especially with AgenTek AI.

所以现在有更多人在争这块蛋糕。

So more people are now competing for that pie.

但如果你看看非超大规模厂商的资本支出，这仍然是英伟达主导的市场。

But if you think about the non hyperscaler CapEx, this is still very much Nvidia dominated world.

我认为，投资者希望继续看到这一点。

That's, I think, what investors want to continue to see.

那你能详细说说这两个细分市场的规模和结构吗？

So tell us a little bit about the shape and structure then of these two segments.

电话会议上有人问，哪个细分市场增长更快，还有新云服务提供商应该归入超大规模客户还是非超大规模客户的收入类别。

There was a question on the call about which one they expect to be growing faster, and also where the Neo Clouds would fit if they would fall into the hyperscaler bucket or the non hyperscaler revenue bucket.

这个问题的答案是什么？

What was the answer to that question?

之后我们再聊聊你对这两个市场未来增长的预期。

And then we'll talk a little bit about how you expect these segments to grow afterwards.

是的。

Yeah.

我认为，除了那四五个主要的超大规模云服务商之外，其他所有公司都属于非超大规模类别。

So I think for the most part, everyone else outside of the four or five major hyperscalers are gonna be in a non hyperscaler bucket.

所以像Neo Cloud这样的公司，基本上也都归在这一类里。

So that includes, I think, Neo Cloud guys would largely be in it.

事实上，过去几个月，Nvidia的股价上涨，更多是受Neo Cloud需求乐观情绪的推动，而不是上季度本身业绩超预期或上调指引。

In fact, if you look at the last couple months, Nvidia share price has had a bigger move on the back of the positive bullishness around the Neo Cloud's demand rather than its own earnings beat and raised last quarter.

这说明市场正在期待，非超大规模客户能带来更多的增长机会。

So I think that's a sign that the market's looking for more growth opportunities to come from these non hyperscalers.

所以你们认为，Nvidia和其他公司都预期，第二个类别——也就是非超大规模市场——会成为他们的主要增长来源。

And so you're expecting and the Nvidia and some, they're expecting that second bucket, the non hyperscaler bucket.

没错，他们指望这个市场能带来显著增长。

That is going to be the big growth driver for them.

也许称不上‘主要增长引擎’，毕竟超大规模客户体量太大，但这个细分领域的增速会更快，这才是重点。

Maybe not growth driver, because hyperscalers are so big, but that will be a faster growing segment is the point.

对。

Yes.

我觉得这正是他们想暗示的，而且这个说法可能比单纯靠财报势头更能推动股价重估。

I think that's what they're being alluded to, and I think that is probably going to be a more meaningful narrative to help rerate the stock than purely just its own earnings momentum by itself.

那我想回到你开头说的，你说公司正在寻找新的故事线，而这种重新分类就是其中一部分。

So then I want to go back to what you were saying at the start, is you said the company was looking for its new narrative or its next narrative, and this reclassification being part of that.

那这个新故事到底是什么？

What would be that narrative?

非超大规模云厂商之外的公司，我明白，你可以说广大企业都在转向我们，新兴云厂商也在用我们的产品。

Companies outside of the hyperscalers, I get that, you could say that enterprises broadly are turning to us, the neo clouds are turning to us.

光是这样就能推动股价上涨吗？或者你觉得到底要什么才能让股价再上一个台阶？

Is that enough to move the needle here, or what do you think it actually takes for the stock to move even higher?

因为昨晚的财报出来后，股价其实没怎么动。

Because it didn't really move on the results last night.

是的。

Yeah.

抱歉，让我补充一句。

Because excuse me.

你现在看数据中心收入的构成，超大规模客户和非超大规模客户的占比几乎各占一半。

I think if you look at the split right now, the data center revenue split is now almost fifty fifty between the hyperscalers and the non hyperscalers.

如果非超大规模客户的部分持续增长并超过一半，那将是一个相当积极的信号。

So I think if if we see the non hyperscalers portion continue to rise and exceed 50%, that would be a pretty positive indicator.

另一个可能的叙事方向是服务器CPU，尤其是今年大家对GenAI的关注。

The other thing I think as a second narrative is this focus on the server CPU, especially with the attention being paid this year to Genentech AI.

这一直是英伟达现在更积极发力的领域。

This has been basically an area that Nvidia is now pushing more aggressively.

他们提到，仅服务器CPU这一块，收入可能达到200亿美元。

They talked about potentially 20,000,000,000 in revenues from the server CPU itself.

不过，有人可能会说，对英伟达来说，200亿美元其实不算大数目。

Now I think the I guess the pushback, though, would be 20,000,000,000 for Nvidia is not a huge number.

我的意思是，这听起来很惊人，但考虑到他们现在的总收入规模。

I mean, it's it's it's crazy to think, but given how big the revenue already is.

所以，服务器CPU对整体业务到底有多重要，目前还有争议，但管理层很清楚，市场现在非常关注这一点。

So I think there's still some debate about how meaningful the server CPU could be to its overall business, but it is an area that I think the the management understands that the market is quite focused on right now.

我只是想再澄清一下，让大家明白这些数据是怎么报告的。

I just want to come back to making sure people understand how the results are reported.

CPU收入，也就是服务器CPU收入，在现在的分类体系下会归到哪一类？

CPU revenue, server CPU revenue, where would that fall in that reclassification now under the current structure?

这是个好问题。

Well, yeah, it's a good question.

我觉得它还是属于计算业务这一块。

I mean, I think it still falls under compute.

至于具体怎么细分，我觉得你会看到，无论是超大规模客户还是非超大规模客户，只要他们单独采购，CPU收入都会体现出来。

Now how it breaks down, you know, I think you'll see basically CPU revenues both in the hyperscaler as well as in the non hyperscaler, especially if they sell standalone.

他们提到了200亿美元这个数字，但还没明确说明这部分收入在两类客户之间怎么分配，我认为这两种客户都会贡献CPU收入。

So that I mean, they pulled out a number of, excuse me, of 20,000,000,000, but they haven't really you know, in terms of where it classifies between that, I think you're gonna see CPU revenues in both types of customers.

这次重新分类会不会导致数据的精细度下降？

Do we lose any level of granularity with this new reclassification?

这一直是个权衡的问题，对吧？

It's always kind of a tug of war, right?

从某个角度看，确实更清楚了，但另一方面，细节就少了。

You get a little bit more clarity in one sense, and then you lose some detail in the other way.

这对你来说是个问题吗？

Is that a concern at all for you?

我觉得，从英伟达自身来看，他们的报告也越来越复杂了。

I think from a I mean, Nvidia itself is also getting much more complicated in terms of reporting.

对吧？

Right?

因为他们已经不再只是卖GPU的公司了。

Because it is moving beyond just a GPU revenue company.

对吧？

Right?

正朝着系统解决方案的方向发展。

It's becoming more system oriented.

所以他们以前的报告方式，本来就不完全适用了。

So so I think the the way they have been reporting wasn't gonna be fully reflected anyway.

我觉得最终更重要的是终端客户群体的变化。

I think this I think the end customer base is what ultimately is becoming a more important driver.

所以我相信他们走的这条路是对的。

So I do believe this is the right direction for them.

昨晚电话会议中提到的另一件事是首席财务官科莉特·克雷斯。

The other thing that came up on the call last night is CFO Collette Kress.

她谈到了老款英伟达芯片的租赁价格，说这些价格实际上上涨了。

She talked about rental prices for older Nvidia chips, and she mentioned that they had actually come up.

你觉得是什么原因导致租赁价格上涨？

What do you think is driving the increase in those rental prices?

是因为整个GPU供应紧张吗？

Is it just the GPU crunch broadly?

还是说客户在考虑，也许我们不需要最新最强的型号？

Is it customers saying that, Hey, maybe we don't need the latest and greatest.

用Hopper系列或者A100也能凑合？

We can make do with the Hopper family or the A-100s?

你觉得背后的原因是什么？

What do you think was behind that?

不好意思。

Excuse me.

我觉得你本质上是说，需求依然远远超过现有的供应量。

I think ultimately what you're saying is that you're still seeing demand is just still outstripping whatever supply is out there.

计算需求还是超出了大多数人的预期。

So the compute demand is still exceeding most people's expectations.

所以我认为，这正是租赁费用持续上涨的原因。

So I do think that is why, you know, the rental costs are still going up.

我们早就看到云服务商那边的情况了。

You know, we've seen that already for the neo clouds.

过去几个月我们一直在讨论这个问题。

We've been talking about that in the last couple of months as well.

总的来说，这表明计算资源依然供不应求。

So it still paints an overall picture that, you know, there's just not enough compute demand out there.

但我觉得另一件有趣的事是，供应端的瓶颈正在不断向更多地方蔓延。

But I do think one other thing that's interesting is, you know, the supply side bottlenecks continue to diffuse out to more and more places.

Nvidia 提到过一个说法，就是最近在光通信领域的一系列合作。

So one thing I think that another narrative that Nvidia has talked about is all these recent deals on the optical side.

你看他们跟 Coherent、Lumentum 和 Corning 的合作，都是为了确保光通信环节的供应，我们认为这会成为一个非常关键的推动力。

If you look at these deals with Coherent and Lumentum and Corning, these are all meant to secure the supply side on on the optical side, which we think will be a very big driver.

我觉得 Nvidia 也认同这一点，所以他们才要确保自己能拿到足够的供应。

I think Nvidia agrees with that, which is why they want to make sure that they have enough supply there.

你能再给不太了解光通信的人解释得清楚一点吗？

Can you explain that a little bit more for people who don't follow optical as much?

用大白话再展开说说这部分。

Just unpack that a little bit more in layman's terms.

当然可以。

Sure.

随着这些数据中心不断扩建，其中一个瓶颈就是带宽必须变得越来越快。

So as you see these data centers build out, one of the bottlenecks is really, you know, the bandwidth, the speed has to get faster and faster.

推动这一点的一个关键因素就是光通信的需求。

And one of the things that's driving that is the optical demand.

我们正从铜缆转向光通信，这会带来巨大的需求，从而提升整体性能。

This is gonna be we're gonna move from copper to optical, and this is gonna create a significant amount of demand to help improve performance.

这并不仅仅取决于芯片本身的计算能力。

So it's not purely just based on the actual computing power of the chip.

这是一个更完整的系统集成，包括网络部分。

It's a much more full a system integrated, and that includes the networking piece.

对吧？

Right?

Nvidia 已经通过其 Mellanox 在这方面提前布局了。

So Nvidia has been moving there already with its Mellanox.

但除此之外，他们还得确保整个供应链能跟上步伐。

But in addition to that, they gotta make sure the rest of the supply chain is there with them.

比如内存，这一直是个大瓶颈，他们已经做了大量工作，确保在当前供应紧张的情况下能拿到足够的内存产能。

For example, if you look at memory, this has been a big bottleneck, and they have done a lot of things to make sure they have secured enough memory capacity given the tightness we're seeing.

他们正在把这种做法扩展到供应链的其他环节。

They're just extending that to another part of the supply chain.

明白了。

Got it.

所以你提到的‘光通信’，是指英伟达深入布局的那些网络产品吧。

So when when you say optical, you're you're talking about the the the networking products that that Nvidia has moved deeper into.

是的。

Yeah.

比如那些用在骨干网络里的光收发模块，无论是横向扩展还是纵向提升，都需要它们来提供网络带宽。

Like, they're basically the transceivers that goes into the backbone in terms of in terms of providing the the networking bandwidth, both for scale out and scale up.

对。

Right.

实现这一点需要大量的网络基础设施。

You're gonna need a lot of networking for that.

对。

Right.

我还有几个问题想问问你。

A couple other questions for you.

关于Vera Rubin这个新芯片系列，当你跟渠道伙伴或者客户圈子里的人交流时，他们对这款新芯片最期待的是什么？

So Vera Rubin, the the new family of chips, when you talk to channel partners or people in the orbit of customers, etcetera, what are they hoping for out of this new family of chips?

你对英伟达能否实现这些期望有多大的把握？

And, you know, how confident are you that Nvidia can deliver that?

每次推出新产品，供应链总会有些潜在问题，这很正常。

Well, there's always potentially a concern that whenever you have a new ramp, there could be some, you know, potential issues in supply chain.

之前确实有人担心内存跟不上，或者跟不上得不够快，但我们觉得这应该是短期的。

There has been some potential concerns about memory not catching up or at least not quick enough, But we think there's gonna be relatively short term.

归根结底，英伟达还是在全力追求性能。

At the end of the day, I think Nvidia is still pushing, you know, performance based.

Rubin确实提升了性能。

Rubin does increase performance.

但这也引出另一个问题——我们正看到来自ASIC领域的竞争，尤其是谷歌的TPU。

But that also brings up another question is that we're seeing some competition from the ASIC side, especially Google's TPU.

那边的势头非常强劲，所以市场希望他们能拓展客户群，毕竟你看那些新兴云服务商，他们是不会转向ASIC领域的。

We've seen a lot of very strong momentum there, which is why I think the market is looking for them to extend their customer base because, you know, I think if you look at neo clouds, right, they're not gonna be moving into the ASIC side.

他们还是会牢牢扎根在GPU这一边。

So they're gonna be very much within the GPU camp.

所以我认为，鲁宾芯片的大量需求也会流向这些非超大规模云服务商。

So I think a lot of that Reuben demand will also go into these non hyperscaler customers as well.

很好。

Great.

弗兰克，非常感谢你今天来做客。

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

这位是来自汇丰银行的弗兰克·李，正在TITV为您直播。

That is Frank Lee from HSBC here on TITV.

SpaceX的IPO文件已经正式公布了。

SpaceX's IPO filing is officially out.

这份文件是硅谷乃至全球都翘首以盼的。

It was a document that Silicon Valley markets, the whole world was waiting for.

在某些情况下，内容确实很丰富。

In some cases, there was a lot there.

所以为了详细梳理一下，我请来了罗斯·格伯，格伯川崎财富与投资管理公司的总裁兼首席执行官。

And so to break it all down, I want to bring on Ross Gerber, president and CEO of Gerber Kawasaki Wealth and Investment Management.

罗斯，欢迎再次做客我们的节目。

Ross, welcome back to the show.

很高兴你今天能来。

It's great to have you here.

能回来真好。

Great to be back.

关于SpaceX的IPO文件。

So SpaceX's IPO fund.

里面有不少不错的照片。

There's nice photos in there.

简直可以出一本精装画册了。

Mean, you could make a little coffee table book out of it maybe.

是的，你知道吗，对于S-one来说，这是我第一次看到带图片的版本。

Yeah, you know, for an S-one, it was the first S-one I read with pictures, you know?

对，没错。

Yeah, yeah.

我们待会儿细聊，但先说个大概，是比你预期的好还是差？

So, I mean, we'll get into it, but high level, better or worse than you thought?

说实话，我大概早就知道这些内容了，因为路透社之前不知怎么就拿到了这些信息。

You know, I kind of knew most of the information to be honest because Reuters had gotten ahold of this somehow earlier.

所以我其实早有耳闻。

So I kind of knew.

我还得提一下，我们手里也有一份相关数据。

And I should say, we also had a number on the file.

并不是某些媒体机构。

So not some media outlets.

所以因为我的

So because of my

报道了其中很多内容。

reported a lot of it.

是的。

Yeah.

对。

Yeah.

因为我的人脉关系，我早就知道了一些信息。

So because of my relationships, I kinda had some of the information already.

我觉得特别有意思的是他们如何规划业务，这其实是一个三管齐下的模式。

What I thought was really interesting was the way they laid out, like, their business plan because it's really a three pronged business.

很多人只盯着财务数字和那些曝光出来的细节，比如X AI在服务器上花了多少钱，却忽略了像星链这样的业务其实已经非常赚钱了。

And I think a lot of people are really focused on, like, financial numbers and all the stuff that came out, like how much money x AI is spending on servers versus the fact that Starlink, for example, is a wildly profitable successful business.

航天发射本身也并不是一门差生意。

Space launching is also not per se a bad business either.

当你看到他们完整规划的这三个方向时，第一个就是通过星链实现全球互联。

And so when you see the vision laid out, the three prongs of the vision, which is really about, you know, first, you know, connectivity through Starlink.

其次，就是这些太空数据中心，理论上能大幅降低计算成本，还能为地球之外的生命提供连接。

And then secondly, this, you know, space data center things with which would bring ideally down the cost of compute substantially, but also be a connection to further life outside of Earth.

你懂的。

You know?

这简直是个疯狂的构想。

It's kind of this crazy vision.

但总得有人来做吧。

But, you know, somebody's gotta do it.

我想，这个人就是马斯克。

I guess it's Elon.

所以，这些内容里没什么让你感到意外的吗？

So there was nothing in there that that surprised you really?

确实没让我意外。

Not surprised me.

我唯一不太清楚的是，最近我跟不少SpaceX的员工聊过，因为我们公司跟他们合作很多。

I I think the one thing that I didn't really, I guess, understand is clearly and I I recently met with a lot of SpaceX employees because we're working with a lot of SpaceX employees at my firm.

你能看到这个愿景的不同部分，但看到星舰的设计时，就会感受到一种宏伟得不可思议的未来图景——人类甚至可能灭绝。

And and, you know, you see different parts of the vision, but the way the S-one's lined out, you see this, like, grand grandiose vision of the future and really this idea that humans could be, like, extinct.

所以他想在别处建立一个殖民地，确保人类不会彻底消失。

And so he wants to create this, like, colony somewhere else so that we'll never be extinct.

我个人更倾向于，我们为什么不先好好拯救地球，而不是想着搬去别的地方住呢？

You know, I'm more in the camp of, like, why don't we just, like, save the Earth we're on now versus trying to live somewhere else?

但毫无疑问，埃隆和SpaceX的愿景规模巨大。

But but the certainly, Elon and SpaceX's vision is massive.

这才是让我感到惊讶的地方。

So that's the part that surprised me.

对啊。

Right.

简直让人惊叹。

It's just like, wow.

上亿人生活在火星上？

You know, like, a billion people living on Mars?

得了吧。

Come on.

你懂的。

You know?

我要是能让一百万人离开洛杉矶搬到别的地方，那才怪了。

I can't get a million people to move out of LA to somewhere, you know, else.

你懂的。

You know?

那可太好了。

It'd be great.

我问你个问题。

So let me ask you this.

你还记得特斯拉Model S刚推出的时候吗？

Do you do you, recall when the Tesla s one came out?

我记得是2010年。

I think it was 2010.

你还记得当时读到关于它的内容吗？

Was it if you remember reading through your

我记得那份文件，说实话，我记得IPO那会儿，我们当时一起看了，你知道的。

doc I mean, I I kinda remember the the I remember the IPO because we watched and we you know?

但那时候我对特斯拉其实没那么感兴趣。

But at that time, I wasn't, like, that into Tesla.

你知道的。

You know?

有人跟我提起过，我知道有这么个品牌。

Like, somebody brought it up to me, and I knew of it.

但我真没太放在心上。

And so I didn't actually care that much.

不过等它上市之后，那款车就是Model S，唉，说起来真难过，它昨天刚去世了。

But once it went public, it was actually the Model S, which, rest in peace, you know, died yesterday.

特斯拉昨天还在线上办了追悼会，我本来想看的，但太难过了，根本看不下去——那可是我第一次见到它时的车型啊。

They had the funeral at at, Tesla yesterday online and which I couldn't watch because I was so sad because that was the vehicle when I saw it.

我知道特斯拉有点特别的东西。

I knew Tesla had something very special.

你知道的。

You know?

所以呢，那感觉完全不一样，因为那时候特斯拉才值几十亿美元。

And so, you know, it's a very different experience because Tesla was like a 3 or $4,000,000,000 company.

它还是个初创公司，我们是在赌一个全新的东西。

It was a startup, and we were betting on something new.

SpaceX现在是个成熟的大公司，市值上万亿，已经能给投资者提供参与的机会了。

SpaceX is a very well established multi trillion dollar company that now in its maturity is giving investors an opportunity to get involved with.

我挺幸运的，多年前通过投资推特，间接拿到了SpaceX的股份。

Now I was lucky enough to get into the SpaceX investment through my investment in Twitter many years ago.

所以坦白说，我确实持有SpaceX的股份。

So I'm happy to full disclosure, I do own SpaceX.

所以我对这个阶段特别期待。

So I'm excited about this stage.

但我最兴奋的其实是，美国有人怀揣着如此宏大的愿景。

But what I'm most excited about actually is somebody actually has a grandiose vision in America.

你知道吧？

You know?

没人尝试，就永远到不了那里。

Like, you can't get somewhere if somebody doesn't try.

我也不一定相信人类真能住在火星上。

And I don't necessarily believe we're gonna be living on Mars or not.

这个话题我们未来几十年都可以一直争论下去。

That we can debate that for the next several decades.

我问我孩子，他们就说：谁会想住在火星上啊？

Know, You I ask my kids, my kids are like, why would anybody wanna live on Mars?

你知道的，人类其实根本适应不了失重环境，诸如此类的问题。

Like, you know, humans don't do good with no gravity, it turns out, you know, things like that.

所以我觉得，太空作为一个真正的领域，就像一个全新的产业。

So, but I think space as an actual place, it's like a new sector.

这个领域，据我了解，有着巨大的机会。

And this sector, in what I've learned, has enormous opportunities.

你可以去看那些宏大的愿景，也可以关注实实在在的东西，比如每年能往太空运送多少重量，用来发射卫星、搭建太空基础设施。

And so you can look at this grandiose stuff or you can focus on the stuff that's real, which is, you know, how much weight they can bring into space on an annual basis to put up satellites and put up actual infrastructure in space.

这个数字正在呈指数级增长。

And that number is growing exponentially.

今天下午我们有一场星舰发射。

We have a launch this afternoon, Starship launch.

你知道的。

You know?

希望一切顺利。

Hopefully it goes.

希望它能成功。

Hopefully it

一定会成功的。

It'll go.

问题是，到底能走多远？

The question is, like, how far?

你知道的吧。

You know?

没错。

Yeah.

对。

Yeah.

不过我想聊聊你刚才说的实际情况，财报里有几个数字挺有意思。

So I do want to talk though about You mentioned what's real, and so there were a couple numbers in the filing that were interesting.

我特别想到的是星链的单用户平均收入，也就是ARPU，我们发现，尽管用户数量在增加，这些数字却开始下降了。

Specifically, the numbers I'm thinking about are the ARPU numbers, the average revenue per user numbers for the Starlink And we saw that those numbers have started to come down even despite the number of subscribers having gone up.

我觉得——

And look, I think-

但这也是因为星链迷你版的原因。

But that's also because of the Starlink Mini.

比如，你有两个星链设备。

Like, have two Starlinks.

一个每月120块。

One, I pay 120.

另一个只要50块，因为是迷你版。

The other, I only pay 50 because it's the Mini.

对。

Right.

所以大家其实都明白为什么会这样。

And so I think there's a good understanding out there as to why that's the case.

我的问题是，你预计这些数字会回升吗？

My question is, I mean, do you expect those to come up?

你认为它们还会继续下降吗？

Do you expect those to to continue to come down?

你是怎么看这个数字的？

How are you thinking about that figure?

你知道，这毕竟算是个新兴业务，他们的定价比起普通宽带来说实在很有竞争力。

You know, because it's kind of like a new business, and I find their price point to be very competitive compared to, like, normal Internet.

Starlink 真的是救命稻草。

And Starlinks are a lifesaver.

说实话，它在我经历帕利塞德斯山火的时候救了我一命。

I I you know, it saved my life in the Palisades fire in a lot of ways.

我们整整三天都跟外界断了联系。

Like, we were stuck for three days without any communication.

如果没有 Starlink，乌克兰这场战争根本打不下去。

And without Starlink, it's it's winning the war in Ukraine.

所以咱们千万别小看 Starlink 的价值，还有将来用户最终要付的费用。

So we can't underestimate the value of Starlink and what will eventually be charged to customers.

他们压价明显是为了抢占市场，毕竟 Starlink 的利润率高得吓人。

It seems like they're bringing down prices to to really just be competitive because their their margins are huge at Starlink.

一旦卫星部署到位，运营成本其实很低，赚钱太容易了。

So, you know, once they get it up there in the in the in the sky, you know, it's a very profitable system to run.

所以星链显然是SpaceX的盈利王牌。

So so Starlink is is clearly the, you know, the profit monster of SpaceX.

所以我的真正疑问是，他们要是把用户数量翻几倍，哪怕资费稍微降一点，依然能维持一个快速增长且利润丰厚的业务。

And and so what I'm think I the real question is, they, like, double, triple, you know, the amount of subscribers and and maybe rates go down a little bit, but you still got a rapidly, rapidly growing business that's highly profitable.

从盈利能力来看，星链无疑是SpaceX旗下最出色的业务。

So that's that's definitely the queen of the crop of businesses inside SpaceX as far as from profitability perspective.

另一件有意思的事是，有些员工的股票锁定期比传统的180天要短。

Now, the other thing that was interesting was the lockup period for the shares It's shorter in some cases for some employees than the one hundred and eighty days that everyone has been traditionally used to.

我觉得这对科技公司来说正变得越来越普遍。

My sense is this is becoming a little bit more common for tech companies.

以前可不是这样的。

It wasn't always the case.

那你对关于锁定期的表述怎么看？

But what did you make of the language around the lockup period?

之后我们再聊聊，你认为这种趋势长期来看会不会改变。

Then we'll talk broadly about whether you think this is going to change in the long