本集简介
双语字幕
仅展示文本字幕，不包含中文音频；想边听边看，请使用 Bayt 播客 App。
欢迎大家收看《信息》的TI TV节目。
Welcome everyone to the Information's TI TV.
我是阿卡什·帕什拉查。
My name is Akash Pashracha.
今天是12月15日星期一。
It is Monday, December 15.
今天我们为大家准备了一期精彩的节目。
We have got a great show lined up for you today.
节目开始前先插播一条消息：Roomba吸尘器制造商iRobot公司已申请破产。
Before we get going, iRobot, the maker of the Roomba vacuum cleaner, has filed for bankruptcy.
该公司表示设备仍可正常使用，同时将进行重组——这算是个好消息。
The company says its devices will continue to work as it looks to restructure, which is good news.
破产虽是个坏消息，但吸尘器还能用，这很棒。
Bad news about the bankruptcy, but the vacuums are working, so that's great.
今天的节目将讨论我们最新发布的苹果公司组织架构图，此前该公司经历了一系列高管离职事件。
Today on the show, we are talking about our new Apple org chart that we published, which comes after a series of high profile executive exits.
接下来我们将深入分析在太空建设数据中心的可能性，以及SpaceX备受期待的首次公开募股。
We'll then get into some analysis on building data centers in space and the anticipated IPO from SpaceX.
我们还将独家报道埃隆·马斯克如何推动全球企业购买xAI的Grok模型，节目最后会深度解读我们全新AI基础设施通讯的第一期内容。
We also have some exclusive reporting around Elon Musk's push to get companies around the world to buy x AI's Grok model, and we will end the show with a deep dive into the first edition of our brand new AI infrastructure newsletter.
本期节目内容丰富，让我们立即进入正题。
It is a big show, so let's get right on into things.
苹果公司最近几周出现了一系列高管离职现象。
Apple has seen a series of departures from the company in recent weeks.
其首席运营官和人工智能主管近期相继退休，还有一位设计负责人转投Meta。
Its Chief Operating Officer and AI Chief both retired recently, and also one of its design leaders left for Meta.
这些变动促使我们的新闻团队对苹果新一代领导层进行了追踪报道。
It's all prompted our newsroom to do a bit of reporting on who the new guard is at Apple.
因此今天我们发布了更新版的苹果组织架构图。
And so today, we are publishing our updated appleorg chart.
现在有请长期跟踪报道苹果的资深记者Wayne Ma，为我们详细解读他的调查发现。
I want to bring on Wayne Ma, who has closely covered the company for years, to tell us more about what he's learned.
韦恩，欢迎来到节目。
Wayne, welcome to the show.
早上好。
Good morning.
谢谢阿卡什的邀请。
Thanks, Akash, for having me.
那我们聊聊苹果的组织架构图吧。
So let's talk about the Apple org chart.
你写的报道和我们发布的架构图中内容很丰富，所以我想或许可以从你注意到公司架构图中与上次重大更新相比的三个主要变化开始讨论。
There was a lot to it in the story that you wrote and in the org chart that we published, and so I thought maybe a good way to do this was let's talk about three big changes that you've noticed at the company's org chart from the last time that you did your big update.
然后我们再探讨这些变化分别预示着公司怎样的发展方向。
And then we'll talk about what each of those changes sort of signal about the direction the company's going in.
那就从你指出的第一个重大变化开始吧。
So let's start with one of the first big changes you made.
好的。
Sure.
人工智能领域发生了最重大的变化之一。
One of the biggest changes is in artificial intelligence.
2018年，AI被提升至高管层C级职位，由John G.负责。
So in 2018, it was elevated to its own executive level c suite position run by John G.
还有来自谷歌的Andrea。
And Andrea, who came from Google.
随着他的退休，AI部门被重新划归软件工程组管理。
With his retirement, they're moving it back under the software engineering group.
因此AI更多被视为一个产品，而非独立存在的事物。
So it's more considered a product and not its own type of thing.
既然被视为产品，这暗示着苹果在AI创新方面采取怎样的策略？
And so considered a product, what does that sort of signal about the way Apple is looking to innovate in AI?
因为苹果在AI领域的发展一直有些磕磕绊绊，对吧？
Because it's kind of had a bit of a choppy run here, right?
是的。
Yeah.
我认为这表明他们对AI的看法与Meta或谷歌等公司不同，后者追求的是超级智能，试图让人工智能比人类更聪明，这是一个非常崇高的目标。
I think it signals that they think about AI differently than other companies like Meta or Google who are chasing kind of super intelligence, which is trying to make AI smarter than a human and is a very like lofty goal.
而苹果真正想要的是让AI服务于我们普通人。
Whereas Apple is really trying to make AI, I guess, the rest of us.
对吧？
Right?
让它成为你需要的工具，比如写作、总结或通知。
Making it good for the tools that you need for, like, writing or summarization or for notifications.
好的。
Okay.
在你们的组织架构调整中，你还注意到哪些重大变化？
What's another big change that you noticed in your org chart overhaul?
今年早些时候，苹果宣布其长期担任首席运营官的杰夫·威廉姆斯即将退休。
Well, earlier this year, Apple announced that Jeff Williams, its longtime COO, was retiring.
他不仅负责运营，还负责设计团队、机器人团队，甚至健康业务。
And so he was actually in charge of not just operations, but, the design team, the robotics team, and so and health even.
因此随着他的卸任，他们不得不将他的职责分配给其他苹果高管。
And so they've moved with him stepping down, they've kind of had to divide his responsibilities up with other Apple leaders as well.
他们没考虑提拔某人接任COO职位，还是完全接管了这个角色？
Did they not think to elevate someone to the COO position or just took over the role entirely?
他们确实提拔了萨比·汗担任COO，但并非完全对等接替。
Well, they have elevated Sabi Khan to COO, but it's not one for one match.
我认为最引人注目的是，威廉姆斯本应在紧急情况下接替蒂姆·库克。
And I think the most striking thing about this is that, Williams was supposed to take over from Tim Cook in an emergency.
因此他需要监管许多超出他专业领域的团队，比如设计团队，甚至一度包括自动驾驶汽车项目。
And as a result, he had lots of, overview of teams that were outside of his area of, expertise like design and like, the self driving car at one point.
所以如果我们想看到蒂姆·库克的继任者更早出现，应该会看到他们将部分职责交给某个人，但现在他们又把这些职责分散了。
And so we should have seen if we're going to see a successor to Tim Cook sooner, we would have seen them give some of these responsibilities to to another person instead they'd kind of divide them all up again.
我想深入探讨这种非一对一接替的情况。
And so I I I want to dig into this idea of it not being a one to one replacement.
说说萨比·汗的情况吧。
Mean, tell us about Sabi Khan.
他的背景是什么？
What's his background?
是否因为缺乏经验才犹豫要不要把所有责任都交给他？
Was it sort of a lack of experience that was the hesitation behind giving him all the responsibility?
还是
Or
不是。
No.
我认为只是他在运营方面更专业。
I think it's just he's more specialized in operations.
他主要在运营领域有很长的历史，但在设计或其他这类领域没有经验。
He mainly he's he has a long history in operations, but not in design or these other kind of areas.
因为他刚开始担任COO，我认为他们不想一开始就给他这么庞大的职责范围。
And so because he's just started out as COO, I think they didn't want to give him such a huge portfolio right off the bat.
而且最终可能不是他来接替蒂姆·库克。
And it may not be him that replaces Tim Cook, ultimately.
对吧？
Right?
紧急情况下，我认为他会接任。
In emergency, I think he will.
但从长远来看，我们认为最终人选不是他——实际上我们觉得应该是苹果的硬件工程主管约翰·特尼斯。
But long term, don't think we don't think it's not become we actually think it's John Turnis, this hardware engineering chief at Apple.
对。
Right.
你注意到第三个能反映公司动向的重大变化了吗？
Was there was there a third big change that you noticed that that tells us something about the company?
是的。
Yeah.
在AI部门里有个核心团队叫基础模型组，他们负责开发支撑'苹果智能'的底层软件。
So within the AI group, there's this key group called the Foundation and Models team, and they're the group that, is responsible for the underlying software that powers Apple Intelligence.
该团队原负责人管理着60人，今年早些时候离职了。
And so, the head of that team, who ran a group of 60 people, he left earlier this year.
于是我们找到了他的接替者。
And so we found his replacement.
是陈继芬（Jifen Chen），他从谷歌跳槽而来。
It's Jifen Chen, who joined from Google.
然后我们注意到，在他直接管理的30人团队（总共60人）中，超过半数来自谷歌。
And then we noticed that, you know, of his 30 direct reports, 60 overall, more than half come from Google.
这就引出一个问题：在吸引人才从谷歌过来的主要卖点——约翰·凯和安德莉亚（现已离职）——离开后，这些人还会留多久？
And so then the question is, how long will those people stay given that one of the big selling point for bringing people from Google over was John Key and Andrea who is now leaving.
你在报道中指出，尽管发生了所有这些高管层面的变动，但还有很多东西没有改变。
Now, one of the things you called out in your story was that despite all of these executive level changes, there's a lot that hasn't changed.
是的。
Yeah.
有个流传的笑话说，十年前苹果就有大约100位副总裁和高级副总裁。
There's like a running joke that, you know, ten years ago, Apple had like a 100 VPs and SVPs.
而现在公司规模增长了约23%。
And now the company's grown by like 23%.
但如今有多少高级副总裁和副总裁呢？
But how many, SVPs and VPs are there today?
依然是100人。
Still a 100.
这表明公司在晋升方面非常保守，同时也维持着非常扁平的组织结构，不会不断新增层级。
And that just shows, the company, I guess, is very conservative when it comes to promotions, but also, they run a very flat organization where there's not constantly adding new layers.
对。
Right.
对。
Right.
最后问你一个问题。
And last question for you.
你提到了约翰·特尼斯，以及围绕蒂姆·库克的继任问题——这在苹果相关圈子里是个热门话题，至少对关注公司的人来说是这样。
You you mentioned John Turnis and the idea that, look, succession around Tim Cook is a is a very popular topic in the apple apple orbit, at least amongst people who follow the company.
那么从业务架构中，有没有看出他们对待继任问题的任何端倪？
So did anything about the structure of the business tell you anything about how they're approaching succession?
真的只有这么几个人被考虑作为继任者吗？
Is it really just these few people that are, you know, being thought of as successors?
还是
Or
是的。
Yeah.
实际上，没有任何重大变动，也没有给特尼斯增加更多职责的表现，这可能意味着他们并未立即考虑继任问题，或者对启动这一程序有所犹豫。
Actually, the absence of any major changes and the absence of giving turn this into sort of additional portfolio shows that maybe they're not thinking about succession right away or, they're hesitant to to start it.
所以我认为最大的变化或至少我注意到的重要一点是，目前确实没有太多迹象表明库克很快会有继任者。
And so I think that was the big change or the big thing that noticed that I noticed at least, was that there wasn't really much movements that show that there is, a successor, for Cook anytime soon.
很好。
Great.
韦恩，这次讨论非常精彩。
Well, Wayne, it's a great discussion.
感谢你参与节目。
I wanna thank you for coming on.
我们非常感谢。
We appreciate it.
这位是The Information的记者Wayne Ma。
That is Wayne Ma, a reporter here at The Information.
好的，接下来，太空数据中心最近备受关注，或者说至少未来在太空存储数据的可能性让亿万富翁们津津乐道。
Okay, next up, data centers in space have gotten a lot of attention recently, or at least the prospect of having data in space one day is one that billionaires cannot stop talking about.
这种兴奋的根源在某种程度上与我们最近的报道有关——SpaceX正考虑明年上市，但一如既往，这个故事还有更多内情。
The roots of that excitement in some ways tie back to our recent reporting that SpaceX is looking to go public next year, but as per usual, there is a lot more to that story.
我想请我们的专题编辑尼克·温菲尔德加入讨论，他上周末与人合著了一篇关于该主题的精彩文章。
I want to bring on Nick Wingfield, our features editor who co authored a great piece on that topic over the weekend.
尼克，欢迎回到节目。
Nick, welcome back to the show.
很高兴你能来。
It's great to have you here.
谢谢邀请。
Thank you for having me.
你上周刚结束节目，这周又由你开场。
You ended off last week and you're starting us off this week.
这简直就是尼克·温菲尔德的专场秀。
It is very much the Nick Wingfield show here.
那么把数据中心建在太空这个想法的起源是什么？我是说，是埃隆·马斯克先提出的吗？
So where did the origin of this idea of putting data centers in space I mean, did it start with Elon Musk?
还是说早有初创公司长期致力于此？
Were there startups who have been working on this for a long time?
请简单谈谈这方面的情况。
Just talk a little bit about that.
其实并不是这样。
It actually didn't.
你会以为埃隆·马斯克应该是首批讨论这个话题的人之一。
I mean, you would think Elon Musk would have been one of the first to talk about this.
他创立并领导着全球领先的火箭公司。
He founded and leads the leading rocket company on the planet.
他拥有近地轨道上规模最大的通信卫星群。
He has the largest constellation of communication satellites out there in low Earth orbit.
所以对他来说，开始讨论这个话题应该是自然而然的事。
So it would have been a natural thing for him to start talking about it.
但实际上，他直到上周才开始真正高调发声。
But in fact, he really just started to get very vocal about it this past week.
在此之前，我注意到更多是初创公司在讨论。
Prior to that, I had noticed more talk from startups.
杰夫·贝索斯对这个话题的讨论相对更多些。
Jeff Bezos has talked a bit more about this topic.
我们今年早些时候其实就写过相关报道，讲述太空数据中心、轨道数据中心的构想如何开始兴起。
And we wrote a story actually earlier this year really about how the notion of space data centers, orbital data centers, was starting to pick up.
在此之前，这还属于科幻范畴。
And then for that, it's been a science fiction.
我是说，可能不是数据中心这部分，但关于让卫星在太空中执行工业任务的构想已经存在相当长时间了。
I mean, maybe not the data center part of it, but the notion of having satellites doing industrious things in space has been around for quite a while.
现在，这是个重大挑战，我想谈谈其中的一些难点。
Now, it's a big lift, I want to talk about some of the challenges here.
但如果我们只关注这个十年愿景，它能解决哪些问题？
But if we just stick with the ten year vision here, what problems could this solve?
电力是这里的主要问题吗？
Is it power that's the big problem here?
电力确实是个大问题。
Power is a big one.
我上周末写的报道主要探讨了：这个趋势现在真正兴起的事实，对地球上的数据中心意味着什么？
The story that I wrote over the weekend was largely about what does this say, the fact that this is really starting to pick up right now, what does this say about data centers on earth?
有些人认为，24/7不间断的太阳能供电是将数据中心建在太空的一个非常吸引人的理由。
There are people out there who believe that, well, that free 20 fourseven solar power is a very compelling reason to put data centers in space.
此外还有冷却因素，这与电力以及水资源都有关联。
There's also a cooling element to it too, which is related to power and also water.
不过在太空，冷却难题确实尚未得到解决。
In space though, the cooling challenges have really not been figured out.
如果你想寻找对轨道数据中心概念持怀疑态度的人，他们往往会重点关注这一点。
If you want to look for skeptics of the notion of orbital data centers, they really tend to focus in on this.
马斯克认为或者说这本质上是免费冷却，但我不确定这个技术问题是否已经解决。
Musk thinks that it's or says that it's free cooling essentially, but I'm not sure that that technical problem has been solved.
我认为这还说明了一个问题，那就是地球上的数据中心在很多情况下都遇到了接入电网的障碍。
Now, the other thing that I think this says is that data centers on earth are in many cases running into obstacles connecting to the power grid.
此外，人们不希望这些数据中心建在他们的某些社区里。
Also, people don't want these data centers in some of their communities.
你开始看到有关抗议活动和进程放缓的报道。
You're starting to see reports of protests and slowing down the process.
理论上，把它们放到太空可以让你绕过所有这些繁文缛节。
Theoretically, putting them up in space could allow you to cut through all of that red tape.
没错。
Right.
就挑战而言，他们还没解决冷却问题。
In terms of the challenges, they haven't figured out the cooling.
我记得大约一年前我们写的那篇报道里提到过，太空垃圾也是另一个问题，对吧？
I remember in that story that we wrote about a year ago or something like that, also space debris was another issue, right?
这些东西可能会撞上其他卫星之类的。
The idea that these things could, don't know, knock into other satellites or something like that.
嗯，我看到的一些模型，特别是一家名为星云（Star Cloud）的初创公司展示的，显示这些巨大的太阳能电池阵列。
Well, some of the mockups that I've seen, in particular from a startup called Star Cloud, show these vast solar arrays.
我记得他们的模型大约是两公里乘两公里的太阳能电池阵列。
I think that their mockup is something like a two kilometer by two kilometer solar array.
占据着巨大的空间。
Just taking up a massive amount of space.
如果你想象太空中有很多这样的东西，那么是的，我认为太空碰撞确实是个值得担忧的问题。
If you imagine a lot of these things in space, then yeah, I think space collisions become a real concern.
或许可以通过确保废弃的数据中心脱离轨道，有效减少太空垃圾来缓解这个问题。
Now, there may be ways to mitigate that by making sure that obsolete data centers are deorbited and you reduce the amount of junk effectively in space.
但我仍然认为这是这些公司未来必须解决的一个大问题。
But I still think that's a big issue that these companies are going to have to figure out in the future.
你指出的其中一点，而且报道中有个很棒的图表，就是你谈到了将物品发射到太空的成本一直在下降，而SpaceX在其中起到了很大作用。
One of the things that you pointed out, and there's a great chart in the story, is you talked a little bit about how the cost of launching things into space has been coming down, and SpaceX plays a large part of that.
我是说，这真的是推动我们能在太空开展任何项目的关键数据点吗？
I mean, is that really the key data point here that is propelling the idea that we could get anything in space conceivably?
其中一个肯定是的。
One of them for sure.
而现在，SpaceX作为全球发射火箭最多的公司，通过可重复使用的火箭大幅降低了成本。
And right now, SpaceX, which is the most prolific launcher of rockets anywhere, has brought the cost down dramatically through the reusable rockets.
我认为我们图表中的估计是针对猎鹰重型火箭的，这是他们最大的火箭，现在每公斤有效载荷的成本已降至约1500美元。
I think the estimates that we have in our chart is for the Falcon Heavy, which is their biggest rocket, that they're now down to about $1,500 per kilogram a payload.
他们还有这个巨大的新型火箭——星舰，他们正努力让它真正投入使用。
They have this giant new rocket, Starship, that they are trying to really get rolling.
有人认为这种火箭在几年内成本可能会降到几百美元。
And that rocket, some people think could come down into the hundreds of dollars within a few years.
这就开始为发射更多有效载荷打开了可能性。
So that starts to open up the possibility of bringing up a lot more payloads.
此外还有SpaceX的竞争对手。
And then you also have competitors to SpaceX.
确实，SpaceX已经独领风骚相当长一段时间了。
Really, it's been a kind of SpaceX show for quite some time now.
但蓝色起源——杰夫·贝佐斯的火箭公司——已成功用他们的大型火箭（相当于SpaceX的星际飞船，名为新格伦号）发射了有效载荷。
But Blue Origin, Jeff Bezos' rocket company, has successfully launched payloads from their big rocket, their Starship equivalent, which is called New Glenn.
但他们确实需要加快这些火箭的发射节奏。
But they really have to get those rockets rolling.
此外还有众多小型初创企业，它们也可能带来更多竞争并助力降低价格。
And then you have a number of other smaller startups that could also introduce more competition and help lower prices.
没错。
Right.
那么尼克，最后问你一个问题。
So last question for you, Nick.
你作为记者，过去二十年间至少从宏观层面研究过航天领域。
Now, you've studied the space sector, at least from afar, for the past couple decades as you've been a reporter.
鉴于你目前看到的行业发展趋势，你认为这现实吗？
Given what you've seen now with the momentum in the sector, do you think it's realistic?
我是说，十年后我们会在太空建立数据中心吗？
I mean, ten years from now, do we have data centers in space?
还是说这仅仅是个构想，也许是个埃隆·马斯克式的宏大构想？
Or is it really just an idea, an Elon Musk sized idea maybe?
在科技领域，很少有话题能让我坦言自己真的不清楚。
There are not too many topics in tech where I can say that I don't really know.
看到科技界最聪明、最具影响力、最富有的人——黄仁勋、埃隆·马斯克、杰夫·贝索斯——都开始讨论将太空数据中心作为长期AI算力问题的唯一解决方案，这确实非比寻常。
It's extraordinary to see some of the smartest, most powerful, richest people in tech, Jensen Huang, Elon Musk, Jeff Bezos, all starting to talk about data centers in space as really the only solution to the long term AI compute problem.
但每当报道这个话题时，总会有工程师、航天专家和数据中心从业者跳出来说你是个白痴，并列举各种XYZ理由说明这永远行不通。
Then whenever you write about this, you will get engineers, space people, data center people telling you why you're an idiot and why this will never work for X, Y, and Z reason.
宇宙辐射会损坏芯片。
Radiation zapping the chips.
散热问题尚未解决。
The cooling issue is not solved.
我是说，对于把计算设备送上太空这件事，有太多不同的反对意见了。
I mean, are just so many different objections to Will this whole they get compute into space?
这已经在发生了。
That's already happening.
已经开始有英伟达芯片上太空的演示了。
There are starting to be demonstrations of NVIDIA chips going up.
就在上周，星云公司宣布他们成功在卫星上训练了一个小型LLM模型。
Just this past week, Star Cloud announced that they had trained an LLM, a small one, on a satellite effectively.
人们会这么做的。
People will do this.
但这是否能规模化到真正解决AI算力问题的程度？
Will it ever scale up to the level of actually solving the AI compute problem?
我认为这真的很难说。
I think that's really uncertain.
我们得先看看人们尝试的结果。
We'll have to start to see people try it.
让我再问你一个问题，尼克。
Let me ask you one more question, Nick.
我是说，我一直在想这些创新讨论有多少是牛市行为。
I mean, I always wonder how much of these innovation discussions are bull market behavior.
我是说，市场下跌了，对吧？
I mean, the market was down, right?
如果我们看到股市下滑，可能泡沫会以人们期待的方式破裂。
If we were sort of watching the stock market slip and maybe the bubble bursts in a way that people are waiting for it to.
我是说，如果人们更担心基本面，你觉得这个话题还会受到这么多关注吗？
I mean, do you think this gets as much attention if people are worried about the fundamentals a little bit more?
嗯，AI算力才是更大的问题。
Well, AI compute is the bigger question.
如果我们进入熊市，情况会怎样？
What happens to that if we tip into a bear market?
突然间，AI可能开始增长放缓，我们就能在地球上满足AI算力需求了。
All of a sudden, maybe AI starts to grow more slowly and we can meet the AI compute needs here on the planet.
但如果你假设AI增长真的会直冲月球，那么太空数据中心确实开始显得更有意义。
But if you assume that AI growth is going to the moon literally, then yeah, space data centers start to make more sense.
我认为我们还没深入讨论的另一点是，现在之所以听到这个消息，是因为埃隆·马斯克想让SpaceX上市。
I think the other thing here that we haven't really talked too much about is that the reason we're hearing about this right now is because Elon Musk wants to take SpaceX public.
这给了他一个绝佳机会，能有效利用AI叙事为SpaceX造势。
This gives him an opportunity to tap into the AI story effectively for SpaceX.
这本身已经是相当吸引人的业务了。
It's already a pretty compelling business.
他们主导着发射市场。
They dominate the launch market.
他们在卫星互联网领域占据统治地位。
They the dominant position in satellite internet.
如果他们能分得AI计算的一杯羹——
If they can get a piece of this AI compute-
没错。
Right.
这个故事变得更有吸引力了。
Story becomes much more compelling.
完全同意。
Totally.
完全同意。
Totally.
没错。
Right.
尼克，感谢你今天的参与。
Well, Nick, I want to thank you for coming on.
这是一次很棒的讨论，期待很快能再次与你交谈。
It was a great piece, I look forward to talking to you again soon.
谢谢邀请。
Thanks for having me.
好的。
Okay.
继续我们的太空话题，投资者在展望SpaceX首次公开募股时面临的一个关键问题是：在评估该公司价值时，他们可以参照哪些其他公司进行合适对比。
Sticking with our space coverage, one of the big questions that investors looking ahead to the SpaceX IPO will have to answer is which other companies they may turn to for a suitable comparison as they try to sort through what the company is worth.
据彭博社报道，SpaceX目前正寻求达到1.5万亿美元的估值目标。
As of now, SpaceX is reportedly targeting a $1,500,000,000,000 valuation that is according to Bloomberg.
现在有请卫星咨询公司TMF Associates总裁蒂姆·费拉尔为我们分享他的见解。
I want to bring on Tim Ferrar, president of satellite consulting company TMF Associates, to offer his thoughts.
蒂姆，欢迎回到节目。
Tim, welcome back to the show.
很高兴你能来参加。
It's great to have you here.
非常感谢。
Thanks very much.
稍后我们想深入探讨SpaceX的基本面，不过现在先说说太空数据中心，你是看涨派吗？
I do want to get into the fundamentals of SpaceX in a second here, but look, data centers in space, are you bullish?
还是看跌派？
Are you bearish?
你认为目前情况如何？
Where do you think this stands?
我觉得最有趣的是，我们这么快就从SpaceX的传统故事——过去十年间一直围绕发射和卫星业务——转向了新话题。
I think the most interesting thing is how quickly we've pivoted from a story for SpaceX, which has been about launch and then it's been about satellites for the last decade.
所有人之前都在说，好吧，星链产生的收入将足够送埃隆·马斯克上火星，而现在我们突然开始讨论数据中心了。
Everyone was saying, okay, well, Starlink's gonna generate enough revenue to take Elon Musk to Mars, and and now we're suddenly talking about data centers instead instead.
而且，我认为这正说明SpaceX已经连续三年未能实现其预测目标。
And, you know, I think that's indicative of the fact that SpaceX has missed its forecast for the last three years.
我是说，三年前的2023年，我们听说营收目标是80或90亿美元。
I mean, three years ago, 2023, we're hearing 8 or 9,000,000,000 of revenue.
现在被告知只有74亿美元。
Now we're told it's 7,400,000,000.0.
今年马斯克曾说会达到155亿美元。
This year, Musk said it's gonna be 15,500,000,000.0.
现在看来他们最多只能争取到150亿。
Now it looks like they're trying to get to 15 at best.
所以，他们确实需要一个新的故事，我认为太空数据中心就是为此准备的。
So, yeah, they need a new story, and I think that's what data centers in space do for them.
那么，让我们来谈谈支撑SpaceX至今发展的故事。
Well, let's talk about the story that has carried SpaceX so far through its growth.
对于没有深入研究过这家公司的人，请简单介绍一下公司目前主要的盈利来源。
I mean, for people who haven't studied the company in great detail, just walk us through how the company actually makes most of its money right now.
没错。
Right.
公司最初显然是从事发射业务的。
So the company started off obviously in launch business.
在这方面已经占据了非常主导的地位。
It's got very dominant in that.
大约十年前，马斯克注意到人们正考虑发射卫星到太空提供宽带服务。
And then about ten years ago, Musk saw that people were thinking about launching satellites into space to do broadband.
于是他顺势加入了这一浪潮。
And he jumped on that bandwagon.
他在这一领域取得了惊人的成功，正是这一点推动SpaceX的收入从五年前的20或30亿美元增长到今年接近150亿美元。
He's been incredibly successful at that, and and that's what's driven SpaceX's revenues from 2 or $3,000,000,000 five years ago to close to $15,000,000,000 this year.
这项业务确实人气飙升。
And that's really surged in popularity.
目前已有800万人使用Starlink的宽带服务。
We've now got 8,000,000 people using Starlink's broadband service.
他们刚在今年夏天与T-Mobile合作推出了新的直连手机服务，仅在美国及国际市场就已拥有超过800万用户。
They've just launched a new direct to cell phone service with T Mobile in the summer, and that's already had over 8,000,000 users in The US and internationally.
因此人们将其视为巨大成功，但话说回来，实际表现还是略低于他们的预期。
So people are looking at that as great success story, but again, it's not been quite as good as they'd hoped for.
目前Starlink构成了他们收入的大部分，是这样吗？
And Starlink right now makes up most of their revenue, is that right?
没错。
That's right.
是的，今年Starlink很可能将占总收入的三分之二左右。
Yeah, I mean Starlink this year is probably going to be around two thirds of the total revenue.
显然，这既包括他们为消费者提供的服务，也包括他们为美国政府所做的许多工作。
Obviously, that includes both the services that they provide to consumers and a lot of stuff that they're doing for the US government.
他们正在为国家侦察局建造一个价值18亿美元的系统。
So they're building a $1,800,000,000 system for the National Reconnaissance Office.
有传言称他们将参与'金穹顶'项目。
There's talk about them getting into Golden Dome.
还有关于为国防部建立'MillNet'通信系统的讨论。
Talk about a MillNet communication system for the DOD.
因此，除了面向消费者和企业的服务外，这也是他们业务的重要驱动力。
So that is also on top of the sort of consumer and enterprise services, a big driver of their business.
但发射业务目前只有30到40亿美元规模。
But launch, yeah, it's only 3 or $4,000,000,000 right now.
我想再谈谈发射业务，因为在AI数据中心的未来，这可能变得更具相关性。
And I just want to go back to the launch business because this could become more relevant, I guess, in this AI data center future.
对于发射业务来说，传统上是谁在支付SpaceX将他们的卫星或其他载荷送入太空？
But for the launch business, who traditionally has been paying SpaceX to get their satellites in space, whatever they're launching in space?
没错。
Right.
所以，我的意思是，这是政府任务的混合体。
So, I mean, it's a mixture of government missions.
他们正在为国防部发射卫星。
They are launching satellites to the Department of Defense.
他们为NASA将宇航员送入太空。
They're launching astronauts into space for NASA.
他们获得了NASA为阿尔忒弥斯计划提供的资金支持。
They're getting funded by NASA for the Artemis program.
当然，他们还主导了商业发射业务。
And of course, they dominated commercial launch business.
所以他们一直在为许多其他客户发射，包括像亚马逊和Warmweb这样的强大竞争对手。
So they've been launching for a lot of other people, including great rivals like Amazon and Warmweb.
那么现在当我们展望IPO时，我的意思是，在这个市场还有哪些上市公司可以作为可比对象？
So now as we look ahead to the IPO, I mean, what other public companies are there that play in this market that you might look to as comparable?
嗯，我认为这确实很难，因为航天业里既有传统巨头，比如波音、洛克希德·马丁这样的企业，它们的估值完全不具备可比性。
Well, I think that's really difficult because what you've got in the space industry is a mixture of legacy players, the Boeings, Lockheed Martin's of this world, that really aren't comparable at all in terms of valuation.
然后还有很多新兴太空公司，规模比SpaceX小得多。
And then you have a lot of other new space companies that are way, way smaller than SpaceX.
基本上它们都在说，既然SpaceX现在估值8000亿，
And in the main, they're mostly just saying, well, SpaceX is worth 800,000,000,000 now.
明年可能就值1.5万亿了，
Maybe it's worth 1,500,000,000,000.0 next year.
那我们至少应该占它估值的一个零头。
So we ought to be a percentage of that.
所以估值逻辑就是这样传导的。
So it goes that way around.
所有公司都以SpaceX为估值基准，而不是反过来。
Everyone takes their valuation from SpaceX, not the other way around.
所以确实，要找到可比对象，恐怕只能看埃隆·马斯克的其他公司了。
And so, yeah, really, you've got to look at Elon Musk's other companies to find any comparison.
我是说，这已经超出了航天工业的范畴。
I mean, that's outside the space industry.
但如果你能证明特斯拉的估值合理，那么或许也能证明SpaceX的估值合理。
But if you can justify that you know, Tesla's valuation, then maybe you can justify SpaceX's valuation as well.
归根结底，这取决于对埃隆·马斯克的信任。
It's about belief in Elon Musk after all.
所以1.5万亿。
So 1,500,000,000,000.0.
我是说，
I mean,
你觉得这个数字，我是说，依据是什么？
do you this number, I mean, comes from what?
只是埃隆·马斯克在定价吗？
It's just Elon Musk setting a price?
我是说，这背后有任何基本面支撑吗？
I mean, is there any fundamental backing behind it?
嗯，我是说，上周我们看到的8000亿这个数字的讽刺之处在于，它似乎是为了每股420.69美元这个方便取整的价格。
Well, I mean, you know, the irony of the 800,000,000,000 number that we saw last week is that it seems to be a conveniently rounded number for $420.69 a share.
要知道，这类估值背后未必有太多基本面支撑。
You know, these things tend to have not necessarily a lot of fundamentals behind them.
我认为这只是他们认为能获得的价位。
I think it's what they think they can get.
是的，今年营收的100倍看起来疯狂，直到你看到马斯克其他公司的成就。
Yeah, a 100 times this year's revenues seems crazy until you look at what Musk has done with his other companies.
现在，如果我们把IPO视为融资活动，你认为SpaceX需要为实现其描绘的野心上市吗？
Now, if we think about the IPO as a fundraising event, do you think SpaceX needs to go public for the ambitions that it has painted?
我是说，我们看到它能够保持私有化这么久，而且我们在节目中也讨论过，至少对大型企业而言，私募市场是可触及的。
I mean, we've seen that it's been able to stay private for so long, and we've talked on this show about the private market being at least tappable for the largest companies.
你认为IPO对公司来说是个好主意吗？
Do you think an IPO is a good idea for the company?
嗯，我认为过去很多公司上市的原因是为了给员工和投资者提供流动性，但这其实并不必要，因为他们能每六个月进行一次大规模股权回购，既给员工流动性，也给了他们留下的理由。
Well, think the reasons why a lot of companies went public in the past has been to give liquidity to employees and investors, and that's not really been necessary because they're able to do these really large tender offers every six months to give their employees liquidity and, you know, give them a reason to stay.
所以，这可能已经达到了极限，他们现在不仅需要资金将数据中心发射到太空，还需要开发星际飞船，以及实现火星计划。
So, you know, that's gone as far as it can perhaps now, and they need some money not just for launching these data centers into space, but also for developing Starship, but going to Mars.
而且他们几年后还要欠EchoStar大约170亿美元。
And and they owe EchoStar about $17,000,000,000 in a couple of years time anyway.
我的意思是，其中一半是股权，但他们也需要现金来处理这类事情。
I mean, half of that's in equity, but they're going to need some cash for those sorts of things as well.
所以现在是个好时机。
So it's a good time.
但归根结底，如果明天有人对埃隆·马斯克说，我无条件给你100亿美元，你可以保持私有化，他可能会认为这比公开市场中的所有问题都要好，就像我们在特斯拉身上经常看到的那样。
But at the end of the day, if someone came to Elon Musk tomorrow and said, I'll give you $10,000,000,000 no strings attached, you can stay private, He might think that's a better idea than all the problems in the public markets, which we've seen so often with Tesla.
没错。
Right.
因为在我看来，航天领域未必是最可预测的市场，就预测收入和季度结果而言，这些数据实际上无法让人对接下来六到九个月的情况产生任何信心。
Because it doesn't really seem to me like the space sector is necessarily the most predictable market in terms of forecasting revenues and quarterly results that actually give people any kind of confidence about what's coming in the six to nine months to come.
对。
Right.
而且由于公司未能达到预期，你知道，在私有化状态下，它可以讲述这些关于明年220或240亿美元收入的故事。
And because the company has fallen short of expectations, you know, while it's private, it's able to, you know, tell these stories about 22 or $24,000,000,000 of revenue next year.
埃文会认真对待这些说法，因为他们确实没有太多数据可供分析，无法得出准确的预测。
Evon will take them seriously because they don't really have a lot of data to analyze to, you know, come up with predictions that have, been true.
我是说，这家公司确实表现不佳。
I mean, the company has fallen short.
私有化很棒，因为你可以选择最有利的指标呈现给投资者。
It's great being private because you can just select the best possible metrics to give out to investors.
而当他们上市后，你会看到更多数据，也有更多能力来分析公司的实际运营状况。
And when they're public, you'll see a lot more data and a lot more ability to analyze how the company's actually doing.
没错。
Right.
让我再问你一个问题。
Let me ask you one more question here.
我在考虑SpaceX首次公开募股时，一直在想那些可能被SpaceX带入其轨道的公司。
As I've been thinking about a SpaceX IPO, I've been thinking about the companies that SpaceX may have helped lift in its orbit.
我在考虑供应商，那些为火箭制造零部件等相关产品的企业。
I'm thinking about suppliers, people that make the parts for these rockets and stuff like that.
再次展望IPO，我的理论是，一旦SpaceX上市，它很可能会成为一只网红股，因为人们都想参与马斯克最新公司的投资。
Again, looking ahead to an IPO, look, my theory is that if and when it does go public, I think there's a lot of potential for this to turn into a bit of a meme stock in terms of people just wanting to get in on Musk's latest company.
然后我开始思考，如果SpaceX的IPO进展顺利，或者股价最终飙升，对吧？
And then I started thinking about, okay, well, if SpaceX IPO goes well, or if the stock ends up surging, right?
我的意思是，是否会有小型航天公司看到这个机会后选择上市。
I mean, are there smaller space companies that then would see an opportunity to then go public.
你是否认为存在一批可能跟随SpaceX脚步进入公开市场的企业？
Do you see sort of an orbit of companies here that could follow SpaceX in taking to the public markets?
嗯，我认为我们四年前已经经历过一轮SPAC泡沫，当时不少航天公司都借此机会上市。
Well, I think we've already seen a sort of SPAC bubble about four years ago, and quite a lot of space companies took advantage of that then.
但你知道，并非所有这些公司都发展得很好。
And, you know, not all of those have gone super well.
特别是在IPO之后，它们的股价都出现了大幅回落。
Certainly after their IPOs, they they fell back a long way.
过去一年到一年半的时间里，随着SpaceX估值不断攀升，许多相关公司的股价也被推高了。
Now in the last year or eighteen months, a lot of them have been inflated on the back of this growing SpaceX valuation.
我认为，就在最近几周，自从SpaceX可能IPO的消息开始传出，我们已经看到其中一些公司的股价进一步上涨。
And I think that even in the last couple of weeks since we saw the news of a potential IPO for SpaceX start to emerge, we've seen some of those share prices increase further.
所以我认为人们真正关注的是SpaceX的竞争对手，因为大家的假设是这不可能永远垄断下去。
So I think people are looking really at competitors to SpaceX because people's assumption is this can't be a monopoly forever.
那么替代方案是什么？
So what are the alternatives?
人们会说，好吧，也许这个能成，也许不行，但或许它值得SpaceX估值的5%、10%或1%。
And people say, well, okay, maybe this will work, maybe it won't, but maybe it's worth 5% or 10% or 1% of SpaceX's valuation.
因此，这给那些试图与SpaceX竞争的其他类网红股票带来了大量关注——坦白说，到目前为止它们都相当不成功。
And so, you know, that gives a lot of attention to those other sort of meme stocks in some cases who are trying to compete with SpaceX, so far pretty unsuccessfully, to be frank.
很棒。
Great.
好的蒂姆，再次感谢你参加我们的节目。
Well, Tim, I want to thank you for coming on the show once again.
每次和你交谈都很愉快。
It's always a great conversation.
我们很快会再联系你。
We'll talk to you again very soon.
非常感谢。
Thanks very much.
XAI正致力于成为一家企业软件公司，或者说至少正在组建一个向企业销售软件的团队。
XAI is looking to be an enterprise software company, or at least it is trying to build a group that will sell its software to businesses.
根据The Information的最新报道，我想请这篇报道的记者西奥·韦特来为我们详细讲述埃隆·马斯克在该领域的雄心壮志。
That is according to a new report from The Information, and I want to bring on Theo Waite, the reporter behind that story, to tell us more about Elon Musk's ambitions in that arena.
西奥，欢迎再次来到节目。
Theo, welcome back to the show.
很高兴你能来。
It's great to have you here.
很荣幸来到这里。
Good to be here.
今天是埃隆·马斯克的主题日。
It's the Elon Musk show today.
我们要讨论太空探索，还要聊聊xAI。
We're talking about space, we're talking about xAI.
虽然还没谈到特斯拉，但很快就会涉及。
We haven't yet done Tesla, but we'll have to get there soon.
我们来谈谈你发布的这篇报道吧。
Let's talk about this report that you published.
简单介绍一下xAI试图向企业销售的产品。
Talk a little bit about the product that XAI is trying to sell to businesses.
他们想向企业推销Grok。
They wanna sell Grok for business.
这是我们在X平台上常见AI模型的企业版本。
It's the it's the enterprise version of the AI models that, you know, we we've all seen on on X.
他们过去半年组建了约十二人的团队在推广这个产品。
And they they have about a dozen people or so that they've they've built up over the past six months that are out there pitching this.
好的。
Okay.
所以团队规模还不算太大。
So it's not it's not too big a team yet.
我是说，这个...
I mean, it's it's
十几个人。
a dozen people.
规模很小。
It's tiny.
是啊。
Yeah.
明白了。
Okay.
那目前进展如何？
And how has the progress been?
展开剩余字幕（还有 94 条）
你知道，据我听到的消息来源说，Palantir、摩根士丹利等一些大公司进行过一些规模相当有限的测试，但与微软、谷歌或OpenAI相比，规模简直小得可怜。
You know, there have been some some somewhat limited tests I I've heard from sources that Palantir, Morgan Stanley are some of the bigger names, but the the scale is just incredibly tiny compared to Microsoft or Google or OpenAI.
很大程度上，这些公司本来就与埃隆有业务往来，比如摩根士丹利通常就是他们的首选投资银行。
And for the large part, you know, these are these are companies that have existing relationships with Elon already, like Morgan Stanley, for example, is, you know, generally has preferred investment bank.
根据我的消息来源所说，这些公司的测试规模通常在几十万或几百万美元左右。
And these are tests that at these companies that are seemingly in the you know, based on what my sources have said, generally, hundreds of thousands or or low millions of dollars.
对于像xAI这样在芯片上投入巨资的公司来说，这些测试金额真的微不足道。
They're they're really not very significant for a company that's spending as much money on chips as x AI is.
很多时候当企业转向企业软件业务模式，，是因为他们的消费者业务更具周期性，可能没有达到预期效果。
And so a lot of times when people or when companies turn towards an enterprise software type of business model, it's because their consumer business is a little more cyclical and maybe hadn't seen the results that they would have wanted.
消费者业务目前表现如何？
How is the consumer business doing right now?
目前消费者业务绝对比企业业务规模更大。
It's definitely bigger than the enterprise business right now.
比如，你看看x平台这类面向消费者的平台，Grok已经无处不在。
Like, you know, if you think about, you know if you if you look at x, if you look at kind of, you know, these consumer facing platforms, like, Grok is everywhere there.
它正在不断地被推送给X平台的用户。
It's being it's being, you know, pushed on x users all the time.
而且他们还有一种意愿，敢于涉足其他AI公司避之不及的领域。
And they also have these kind of they have a kind of willingness to go to places that other AI companies have been squeamish about.
所以现在Grok应用上有AI女友功能，并且对图片视频等内容审核更为宽松。
So they have, you know, AI girlfriends now on the Grok app, and they have, looser content moderation in general for, like, images and videos.
嗯哼。
So Mhmm.
我不确定人们会怎么使用这些功能。
I'm not sure what people do with that.
目前来看，这确实在推动他们的消费者业务比企业业务增长得更显著。
And for now, that's kind of helped, you know, boost their consumer business, more significantly than their enterprise business.
这么说消费者业务至少通过这些公司推出的实验性产品获得了一些关注度。
So the consumer business is is getting at least some traction by way of these experimental products that the company has rolled out.
这个说法很贴切。
That's a great term for it.
是的，没错。
Yes, that's right.
那么现在请告诉我，公司在创收方面有多紧迫，毕竟它已经筹集了这么多资金。
So now tell me a little bit about how much of a rush the company is in to generate revenue, because it's raised all this money.
我的意思是，它还在建设这些数据中心。
I mean, it's still building these data centers.
我是说，现在存在资金紧张的情况吗？
I mean, is there a funding crunch here?
你知道，我不一定会说存在完全的资金紧张，但即使是埃隆·马斯克，当xAI花费如此巨额资金时，最终也需要外部融资。
You know, I I I I wouldn't say there's necessarily complete funding crunch, but, like, you know, even if you're Elon Musk, you know, at some point, you're you're gonna need to raise outside money when you're spending the amount of money that that x AI is.
而且，这些公司在某种意义上一直在进行融资活动，它们需要在这个过程中展示收入增长。
And, you know, that's kind of a like, these companies are kind of constantly fundraising in a sense, and they're gonna need to show revenue growth when they're doing that.
我的意思是，它们短期内肯定不会盈利，但它们正试图通过建立可重复的企业收入业务来达到这个目标，这对它们会有很大帮助。
I mean, they're definitely not gonna generate profit anytime soon, but they're they're trying to, you know, get get there with you know, if they can go out and say we have this, you know, recurring revenue enterprise business going, like, that'll help them a lot.
嗯，我个人觉得，一家同时销售AI女友服务又试图将其模型卖给财富500强企业的公司，这有点...
Well, and I tend to think that, I don't know, a company that is selling AI girlfriends, at the same time trying to sell its models to Fortune 500 corporations, I don't know.
你觉得，人们真的认为这个企业软件业务会成功吗？
Do like, do people actually think that this enterprise software business is gonna take off?
还是说这很尴尬。
Or is this awkward.
我的意思是，是的。
I mean, yeah.
从埃隆的角度来看，他们想要一个无可争议的好产品，因为其卓越性而让所有人不得不使用。
I, you know, I think from Elon's perspective, they they wanna have, like, an undeniably good product that, you know, everyone will be forced to use because it's just far and away better.
但我认为他们还没达到那个水平。
But I don't think that they're they're there yet.
如果你从一个财富500强公司高管的视角来看，他们决定要达成哪些企业交易时。
I mean, if you think about it from the perspective of, like, an executive at a Fortune 500 company that's, you know, deciding which enterprise, you know, deals to make.
当他们想到Grok时，他们会联想到围绕它的许多头条新闻。
Like, when they think of Grok, they're gonna think of a lot of the headlines around it
是啊。
Yeah.
你知道的，就是那个麦加希特勒闹剧和AI女友的新闻。
Which are the, you know, Mecca Hitler fiasco and, AI girlfriends.
相比之下，如果你面对的是一个合作了十年的微软销售代表，他们提供的是某种定制化的微软模型或通过微软路由的OpenAI模型，你大概率会选择更保守稳妥的方案。
And, you know, if you have that versus, like, a Microsoft salesperson that you've worked with for ten years that's offering you some kind of, like, specialized, you know, Microsoft model or OpenAI model routed through Microsoft, like, you're you're just gonna go with the more conservative and safe option generally.
是啊。
Yeah.
嗯，我觉得你说'这相当尴尬'确实很好地概括了现状。
Well, I I think your your comment that it's pretty awkward is is a good way to to sum it up.
那么，西奥，感谢参加我们的节目。
So, Theo, thanks for coming on the show.
我们非常感激，很快会再邀请你回来。
We really appreciate it, and we will have you back again soon.
谢谢。
Thanks.
好的，今天《信息报》将推出一份由Nebius赞助的新简报，专注于AI基础设施领域。
Okay, today, The Information is launching a new newsletter sponsored by Nebius, focused on AI infrastructure.
这已成为近期最受关注的话题之一，数据中心和电力问题成为讨论的核心。
It has become one of the most closely watched stories lately, with data centers and power being at the center of the conversation.
今天的专栏文章对所有这些计算设施所需的电力需求进行了现实检验。
Today's column offers a bit of a reality check on the power that's going to be needed for all these compute facilities.
我想请安妮·戴维斯·沃恩来为我们讲述她写的这篇文章。
I want to bring on Anne Davis Vaughan to tell us about the piece that she wrote.
安妮，欢迎来到节目。
Anne, welcome to the show.
很高兴你能来。
It's great to have you here.
非常感谢你，阿波什。
Thank you so much, Aposh.
很荣幸能来到这里。
It's great to be here.
我说欢迎回来，因为我预感随着你为这份通讯撰稿，我们会越来越多地邀请你上节目，我们非常期待你继续供稿。
I say welcome back because I anticipate that we're going to have you on more and more as you write for this newsletter, and we're so excited for you to do it.
恭喜项目启动。
Congratulations on the launch.
听着，我想先就AI基础设施这个话题展开一些高层面的讨论。
Look, I want to start with a bit of a high level discussion here on the AI infrastructure story.
我是说，这里有巨额资金投入，也有大量工作要做。
I mean, there's so much money, there's so much work to do.
这个过程耗时漫长。
It takes so long.
根据你的初步报道，你研究这个领域已久，我是说，进度比预期提前了吗？
From your initial reporting here, you've been studying this space for so long, I mean, things ahead of schedule?
他们目前进展是否符合预期？
Are they sort of right in the pocket here?
还是落后于人们当初的预期？
Are they behind where people thought they were going to be?
当前整体形势如何？
What's the lay of the land right now?
当然。
Sure.
我认为至少对一些项目而言，更大的风险在于它们会落后于市场预期。
I would say that there is a bigger risk, at least for some projects, that they'll be behind where the market wants them to be.
而高水平的执行力在2026年将变得极其珍贵。
And that execution at a high level is going to be at a huge premium in 2026.
这远不止是模型和芯片那么简单。
There's just so much more to it than the models and the chips.
硅谷正不得不深入实体经济领域来实现这一目标。
And Silicon Valley is having to go deep into the fiscal economy to get this done.
企业已经意识到，没有其他机构或行业能像它们这样愿意按照自己的时间表推进。
The the companies have realized that there's nobody, there's no other, you know, institution or industry that is as willing as they are to move on their timeline.
所以他们正挽起袖子，真正投入到实体工作中。
So they're rolling up their sleeves and getting really physical.
这实际上是一个软件行业不得不学习如何融入工业经济的过程。
And this is a, you know, in effect a software industry that is having to learn to get involved in the industrial economy.
因此我认为，世界其他地区将这场竞争视为计算竞赛，但参与者若想在模型竞赛中立足，就必须真正投身工业战场，因为他们所需的一切资源都极为稀缺。
And so the way I think about it is that the rest of the world is thinking about this as a computing race, but the, the the players are really gonna have to get on the industrial battlefield if they wanna even compete in the model race because of scarcity of everything that they need.
所以他们正在自建发电厂，建立变压器和太阳能组件的专属供应链。
And so they're building their own power plants, their own supply chains for transformers and solar modules.
他们正在与公用事业公司共同设计新型费率结构。
They're coming up with novel rate structures with utilities.
他们同时推进着诸多前所未有的举措，并同步加速能源创新。
They're doing a lot of things that have never been done before and jumpstarting energy innovation all at the same time.
那么当你考虑要在通讯简报中深入探讨的主题时，未来几周我们可以期待你重点聚焦哪些领域？
And so as you think about the topics that you want to drill into with your newsletter, what are sort of the areas that we can expect you to double click on in the weeks to come?
当然。
Sure.
我们将持续关注这个行业试图实现的惊人规模与推进速度。
Well, we're always going to be talking about the colossal scale and speed that the industry is trying to move at.
其他一些主题还包括：试图理解这个高度泡沫化的经济中哪些部分具有持久性。
You know, the other some of the other themes are trying to understand what about this highly frothy economy is durable.
当你思考这场基础设施竞赛时，我们讨论的是多个需要参与进来的数万亿美元级产业。
And when you think about this infrastructure race, we are talking about several multi trillion dollar industries that have to get involved for this to happen.
这不仅是硅谷的事，还涉及华尔街、电力公用事业、石油天然气行业、清洁能源行业——每个领域都需要提升自身的执行力和速度来应对这一时刻。
It's not just Silicon Valley, it's it's Wall Street, but then it's the utility industry power, the oil and gas industry, the clean energy industry, everybody, you know, needing to improve their own execution and speed to meet this moment.
因此要审视哪些是可持续的，因为部分基础设施显然需要数十年时间才能建成。
And so looking at what's durable because those some of the infrastructure will obviously be, you know, built with like decades long timeframes.
我们正试图协调这些长期规划与AI竞争者的需求之间的关系。
And we're having to figure out how to match those timeframes against what the AI contenders want.
我想说，我们在专栏开篇就指出存在两个被误解的方面。
And I would say, you know, we start the column by saying there are two things that are misunderstood.
我认为人们没有意识到，从听到千兆瓦级项目的宣布到电力真正输送到这些服务器和巨型数据中心需要多长时间。
You know, I believe people don't realize what it's you know, how long it's gonna take to get the gigawatts you hear about an announcement actually pulsing through these servers and big colossal data centers.
但另一个被忽视的是，这种高强度约束条件正在催生多少与AI模型完全无关的创新。
But the other thing people don't realize is how much innovation this heavy amount of constraint is is really producing wholly apart from AI models.
比如，我们正在见证数据中心能效的阶梯式提升，以及针对大型能源登月计划的巨额研发投入和创新。
Like, you know, we're we're talking about step changes and data center energy efficiency and massive R and D and innovation into large scale energy moonshots.
这将带来巨大的红利。
There will be huge dividends on it.
嗯，这确实很有趣。
Well, and it's certainly interesting.
我是说，事情的发展速度显然比以往任何时候都快。
I mean, things are obviously moving faster than they ever have.
问题是，它们的发展速度能否快到满足那些以更快速度涌入该领域投资的投资者们？
Question is, can they move fast enough to satisfy all the investors that have been moving even faster to pour money into the space?
我非常期待你能撰写这份通讯，我迫不及待想阅读它。
I'm so excited that you are authoring this newsletter, and I look forward to reading it.
这位是我们的AI基础设施专栏作家Ann Davis Vaughan，来自The Information。
That is Ann Davis Vaughan, our AI infrastructure columnist here at The Information.
好了，今天的节目就到这里。
Well, does it for today's show.
提醒一下，我们的直播时间是太平洋时间周一至周五上午10点，东部时间下午1点。
A reminder, we are on this stream Monday through Friday at 10AM Pacific, 1PM Eastern.
我要感谢亚马逊网络服务，他们是我们本次节目的主要赞助商。
I want to thank Amazon Web Services, who is our presenting sponsor for this production.
感谢您的收看。
I want to thank you for tuning in.
我们非常感激您的观看。
We really do appreciate your viewership.
我已经对明天的下一期节目充满期待了。
I'm already excited for our next show tomorrow.
祝您周一剩余时光愉快。
Have a great rest of your Monday.
现在先再见啦。
Bye bye for now.
关于 Bayt 播客
Bayt 提供中文+原文双语音频和字幕，帮助你打破语言障碍，轻松听懂全球优质播客。