在OpenAI的广告推广与潜在1000亿美元融资背景下，预览SpaceX上市 | 2025年12月24日

欢迎各位收看Information的TI TV。

Welcome everyone to the Information's TI TV.

我叫阿卡什·马斯里查。

My name is Akash Masricha.

今天是圣诞前夜，12月24日。

It is Christmas Eve, December 24.

假期前我们还有一期节目。

We have got one more show before the holiday.

我们首先讲述一个关于OpenAI计划在ChatGPT上启动广告业务的故事。

We are kicking off with a story about how OpenAI is looking to kickstart an ads business on ChatGPT.

接下来是本周的编辑精选：Information最新独家披露，OpenAI正在寻求新一轮融资，金额可能高达1000亿美元。鉴于亚马逊已开始洽谈投资100亿美元，我们将深入探讨OpenAI如何实现这一千亿美元的目标，并在节目最后分享一个有趣的圣诞故事。

Then it is time for this week's Editor's Cut, the information recently scooped that OpenAI is seeking new funding, potentially as much as $100,000,000,000 With Amazon already in talks to invest $10,000,000,000 we will dig into how OpenAI might reach that $100,000,000,000 target, and we will end the show with an interesting Christmas story.

我们的埃隆·马斯克记者西奥·韦特加入我们，讲述一个关于他家乡的故事，标题已说明一切：X AI的圣诞灯饰揭示了SpaceX上市的哪些信息。

Our Elon Musk reporter, Theo Wait, joins us for a story about his hometown, and the title says it all, What x AI's Christmas Lights Say About the SpaceX IPO.

这将是一期有趣的节目，让我们马上开始。

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

OpenAI正在为ChatGPT中的广告整合奠定基础，今天《信息》杂志独家报道了这些广告可能的实际形态和即将实现的功能。

OpenAI is laying the groundwork for ad integrations into ChatGPT, and today the information has some exclusive reporting out on what that could actually look like and what features could actually become a reality.

我想邀请共同撰写这篇文章的Shree Mappidi，来告诉我们她了解到的更多信息。

I want to bring on Shree Mappidi, who co authored that piece, to tell us more about what she is hearing.

Shree，欢迎再次做客我们的节目。

Shree, welcome back to the show.

很高兴你来到这里。

It's great to have you here.

很高兴能来。

Great to be here.

那么，关于OpenAI在实际操作中的广告策略，你听到了哪些消息？

So what are you hearing about OpenAI's ad strategy on the ground?

没错。

Totally.

所以，首席执行官萨姆·阿尔特曼以及其他高管曾暗示过一种联盟收入模式，即OpenAI将从用户购买的任何产品推荐中抽取一部分收入。

So CEO Sam Altman, as well as other execs have hinted at an affiliate revenue model in which OpenAI takes a cut of revenue for any product recommendations that a user purchases.

但最近，尤其是通过我们的报道，我们了解到OpenAI的员工一直在探讨广告在ChatGPT中可能呈现的形式。

But more lately, especially through our reporting, we've been hearing that OpenAI employees have been actually working through ideas of what ads could actually look like within ChatGPT.

例如，员工们讨论过调整AI模型，以便在结果中优先展示先前或赞助的信息。

For instance, employees have discussed tweaking the AI models to be able to show prior or to show sponsorship sponsored information as a priority to within the results.

例如，如果你在寻找睫毛膏推荐，Sephora的广告可能会弹出，同时他们还制作了广告在ChatGPT中显示的模拟原型。

For example, if you're searching for mascara recommendations, a Sephora ad might pop up, as well as actually making mock ups for what ads would look like within ChatGPT.

我必须说，当人们想到这些平台上做广告时，Sephora似乎成了每个人都会举的例子，这真有意思。

I have to say, it's so funny to me that when people think about people advertising on these platforms, Sephora seems to be the example that everyone uses.

我昨晚刚看了一场电视节目，Sephora简直就是这类故事中的首席广告商。

I I was just watching a TV show last night, and Sephora is like the advertiser in chief for these types of stories.

当然，里面还有很多有趣的细节。

Look, there were a lot of interesting details.

我在你的报道中发现的一个有趣点是，你提到广告可能并不会基于你当前的搜索结果或提问内容。

One of the things that I found interesting in your story is you talked about how the ads could potentially actually pull not from the acute search result that you've put in or the inquiry, guess.

它实际上可能从你两三天前的其他提问，甚至在某些情况下从你的聊天历史中提取信息。

It could actually pull from other inquiries that you've made two, three days ago, and really your chat history in some cases.

更重要的是，你通过聊天所表达的聊天历史和偏好，将成为个性化广告的一种方式，例如向你推送特定的广告。

It's more that the chat history and the preferences that you indicate through your chats will be a way to be able to personalize, for example, a specific ad to you.

所以我认为，广告方面的努力还处于非常早期的阶段。

So I believe that the ad effort is still super early.

目前还不清楚，比如你几天前搜索过的内容是否会再次出现，但更关键的是，聊天机器人了解关于你的各种信息。

It's not clear in terms of whether or not, for example, if you search something a few days ago, it will pop up again, but more that the chat bot knows the different types of information about you.

例如，如果你在搜索去西班牙的旅行，它推荐你去圣家堂，而用户对此感兴趣，那么ChatGPT可能会在下一步展示一个关于圣家堂的赞助信息链接。

So if for example, you're searching for a trip to go to Spain and it recommends you to go to Sagrada Familia and a user is interested, then perhaps, for example, ChatGPT might present a sponsored information link to Sagrada Familia as that secondary step.

OpenAI员工讨论的其他一些想法包括：添加披露说明，告知用户你搜索到的部分结果包含赞助信息；以及在侧边栏显示广告，而不是在主聊天界面中造成杂乱，这样你可以侧面查看这些信息。

A few other ideas that OpenAI employees have been discussing is a disclosure note to actually indicate that some of the results that you're searching has sponsored information, as well as, for example, a sidebar where instead of cluttering, for example, the main chatbot interface, it may pop up to the side so you could sort of look at it on the side.

而这主要涉及用户界面在用户端的呈现方式。

And then this is primarily on sort of what the user interface would look like for users.

但在模型层面，你需要确保模型能够真正捕捉到这些赞助信息。

But then on the model side, you need to make sure that the model can actually capture that sponsored information.

因此，它基本上会优先展示赞助信息，比如在西班牙这个例子中，然后在搜索结果中弹出任何相关的赞助内容。

So it will basically prioritize sponsored information, for example, in that Spain example, and then be able to pop up, for example, anything that is sponsored to show that within the results themselves.

我觉得他们做广告的方式挺有意思的。

It is kind of interesting to me, again, thinking about the way they're doing ads.

在某种程度上，这有点像谷歌的广告模式，就像我们说的，取决于你搜索的内容。

They're sort of in some ways, it's the Google ad model where, like we said, it's depending on what you search.

就像你提到的，在其他情况下，它又有点像社交媒体，比如Instagram。

Like you said, in other cases, it's a little bit more like social media, like Instagram.

这是对你所有搜索内容的综合，从而实现精准广告投放。

It's an amalgamation of all the things that you've been searching for and them being able to target ads.

所以看到他们是如何思考这些问题的，确实挺有意思的。

So it is kind of interesting to see how they're thinking about things.

我们有没有来自公司本身、分析师或公司相关人士的估算，关于这些广告究竟能为公司带来多少收入？

Do we have any estimates from either the company or maybe analysts or people in the company's orbit about how much revenue these ads could actually generate for the company at all?

完全正确。

Totally.

根据今年夏天公布的收入预测，该公司曾表示，预计到2030年，通过免费用户的变现将实现约1100亿美元的收入。

So according to revenue projections from over the summer, the company had actually shared that the they expect to make about $110,000,000,000 through 2030 from free user monetization.

因此，免费用户变现基本上指的是广告。

So free user monetization basically hints at ads.

这是来自大量非付费用户收入。

This is revenue for many non paying users.

对于这一免费用户变现收入类别，从明年开始，每个用户的平均收入约为2美元，到本十年末将上升至15美元。

And for this free user monetization bucket of revenue, the average revenue per user starting next year would roughly be $2 but then goes up to $15 by the end of the decade.

这类产品的毛利率也与Facebook的毛利率相当，大约在80%到85%之间。

The gross margins of these types of products similarly match Facebook's margins as well in terms of roughly 80 to 85%.

我们的理解是，广告例如就属于这一免费用户变现类别。

Our understanding is that ads, for example, would fall within this for user monetization bucket.

那购物呢？

What about shopping?

OpenAI那边的购物业务进展如何？

How's that going over at OpenAI?

购物是OpenAI在向商家和外部人士分享方面较为积极的领域，相比之下，他们在广告方面则更为低调。

Shopping is an area that OpenAI has been pretty forward with in terms of sharing with merchants and sharing externally with folks versus they've been more quiet about, for example, ads.

因此，OpenAI迄今为止推出的某些购物举措包括与Zillow和DoorDash等公司合作的应用程序，用户可以通过聊天直接实现家居配送，或查找旧金山的房屋。

So some of the shopping efforts that OpenAI has launched so far are, for example, apps partnerships with companies like Zillow and DoorDash, where as a user, you can actually chat to be able to have home delivery directly to you or look for a house in San Francisco, for example.

此外，他们还与Shopify和Stripe合作，以支持聊天购物的支付处理。

And then they also work with Shopify and Stripe to process in chat shopping.

还是以Sephora为例，如果我在寻找睫毛膏，而且正好在圣诞节假期急需，我会推荐一款产品，然后快速完成购买并直接配送到我亲家的家中。

So again, back to the Sephora example, if I'm looking for mascara and I'm really in a need for the Christmas holidays, I'll recommend something and then be able to purchase it quickly and deliver it to my in law's house.

这些就是OpenAI在购物方面公开推进的一些举措。

And so those are some of the shopping efforts that OpenAI has been really public with.

此外，就在OpenAI推出一项让商家能够向其发送产品目录的规范时，为准备这一电商浪潮，当时有数千家商家在等待名单上。

Also just in terms of merchants preparing for this e commerce push when OpenAI had launched, for example, a specification to be able to have merchants send their product catalogs to OpenAI, OpenAI had thousands of merchants actually on the waitlist.

这充分表明了OpenAI在电商领域的推动力，以及商家们参与其中的极高热情。

And so, that really indicates their push in e commerce and how merchants too are really excited to participate.

对。

Right.

说到这个等待名单，让我印象特别深的细节是，不仅商家如此，广告商也同样如此。

And speaking of that wait list, again, the detail that was really interesting to me was it's not just the merchants, it's also going back to the advertisers.

你提到，广告商们至少已经开始尝试与OpenAI建立某种关系，感觉他们都在等着看这件事如何发展，仍然在说：请再给我们一点时间，弄清楚这一切会如何运作。

Advertisers, you reported, have tried to at least start to build some kind of a relationship with OpenAI, and it feels like they kind of have the world waiting on how this thing is going to play out, and they're still like, Just give us a second to figure out how it's all going to work.

与此同时，所有这些商家都在尽力争取建立联系，我想，因为这肯定很重要？

Meanwhile, all these merchants are they're trying to get their relationship points in where they can, I guess, because it's certainly going to matter?

在我们结束之前，我问你个简单的问题。

Let me ask you a quick question before we go.

你认为广告对ChatGPT用户体验有多大风险？

How much risk do you think ads might have to the ChatGPT user experience?

你使用这个平台，我也在用。

You use a platform, I use it.

我曾试着设想一下，如果我看到广告，会不会因此质疑结果的可靠性？

I've tried to do the exercise of, if I were to see the ads, would it make me question the result at all?

你觉得用户会怎么看待这个问题？

How do you think customers think about that?

完全正确。

Totally.

我认为这才是真正关键的问题：我们如何不失去用户的信任。

I think this is the real big question about how do we not lose users' trust.

我认为这就是为什么OpenAI在早期对广告策略如此谨慎和实验性，因为我们的数据显示，广告在ChatGPT中呈现的形式必须与Facebook或Instagram上的方式完全不同。

I think that's why OpenAI has been so slow and early and experimentive with their approach, because our reporting essentially shows that the format that these ads could show up in in ChatGPT would have to be totally different for from, for example, how it shows up in Facebook or Instagram.

因此，这种全新的体验需要大量的努力和实验，但我认为这存在巨大风险。

And so this type of new experience would require a lot of effort and experimentation, but I think it is a huge risk.

这是一次重大押注。

It's a big bet.

如果他们能做对，就意味着到2025年能获得1100亿美元；如果做不好，我认为……

If they can get it right, that means $110,000,000,000 through If 2,000 and they don't, I think that

他们可能会更少。

Little they're going less.

好的。

Okay.

好了，Shree，感谢你前来做客。

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

祝你假期愉快，等你回来我们再聊。

Have a great holiday, and we will talk to you when you get back.

这是我们在《The Information》的OpenAI和Anthropic记者Shri Mupiti。

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

好的，本周的编辑精选，我们将深入探讨OpenAI的融资计划，新一轮融资总额可能高达1000亿美元，估值为7500亿美元（投前）。

Okay, for this week's Editor's Cut, we are unpacking OpenAI's fundraising ambitions, which in a new round could total up to an additional $100,000,000,000 at a $750,000,000,000 pre money valuation.

即使对于萨姆·阿尔特曼和OpenAI的融资团队来说，这也是一个巨大的成就。但为了厘清这笔资金的可能来源，我想邀请我们的联合执行主编阿米拉·弗拉迪和执行主编劳拉·曼达罗。

It is a big feat even for Sam Altman and the OpenAI fundraising machine, But to break down where this money could come from, I want to bring on our co executive editor, Amira Fradhi, and our managing editor, Laura Mandaro.

欢迎你们两位。

Welcome to you both.

很高兴见到你们。

It's great to see you.

嗨，各位。

Hey, guys.

好的。

Okay.

所以，劳拉，我想先从你开始，因为你穿着圣诞色系的衣服。

So Laura, I want to start with you because you are wearing Christmas colors.

我想了解，在萨姆·阿尔特曼所有可能的选择中，你认为这1000亿美元可能来自哪里。

So I want to understand where you think the $100,000,000,000 could conceivably come from, of all the options that Sam Altman has open to them.

正如我们两位财经记者凯蒂和迈尔斯昨天在《DealMaker》上所写的，这可能是一场丰盛的自助餐。

Well, as Katie and Miles, two of our finance reporters, wrote yesterday in DealMaker, there's it could be sort of a smorgasbord.

显然，我认为金融投资者都想参与进来，这些投资者可能包括已经投资过的风险投资公司。

Obviously, I think financial investors will wanna get in, And these could range from venture capital firms that have already invested.

红杉资本、安德森·霍洛维茨，这些我们都知道，但这些都是他们过去的投资者，还有像龙耳这样的交叉基金，以及那些通常投资于后期公司的大型共同基金。

Sequoia Andreesen, these We don't know these names, but these are some of their past investors, the crossover funds like Dragon Ear, as well as the big mutual funds that tend to invest in late stage companies.

然而，那只是100亿美元。

However, that's 10,000,000,000.

那么，你如何才能达到1000亿美元呢？

So how do you get to 100,000,000,000?

OpenAI已经获得过一些重大的战略投资，比如亚马逊，或者预计将会获得。

There have already been big strategic investments in OpenAI, such as Amazon or it's expected.

因此，我们可能会看到一些云服务提供商也像对其他模型制造商那样投入股权。

And so we could see some of these cloud providers also putting in equity as they have for some of the other model makers.

然后是第三个类别，即这些规模庞大的主权财富基金。

And then there's a third bucket, which are these very big sovereign wealth funds.

是MGX吗？也就是阿联酋的基金？

Is it MGX, which is UAE fund?

我们曾报道过，他们已与沙特阿拉伯的公共投资基金（PIF）进行过接触，这是一家规模巨大的基金。

We've reported that they've talked to Saudi Arabia's PIF, public investment fund, which is a huge fund.

我们不清楚，也不认为他们已经投资了。

We don't know, we don't think that they've invested.

确实有多种选择，但这个数字依然极其庞大。

There are options, but it's still a colossal number.

好的。

Okay.

所以，阿米尔，正如劳拉所说，有非常多的选择。

So Amir, like Laura said, there are a lot of options.

我想知道，你听到哪些选项最有可能率先启动这轮融资？

I'm curious what you're hearing in terms of which of these options is actually the, you know, the the most likely to maybe start the round?

我们看到了亚马逊的新闻。

We saw the news with Amazon.

你知道的？

You know?

天啊。

Gosh.

超大规模云服务商？

Hyperscalers?

你怎么看？

What do you think?

是的。

Yeah.

我的意思是，在硅谷，一些后期的私营公司确实会开始接触公开市场的投资者。

I mean, you do have situations in Silicon Valley where late stage private firms start to tap public market investors.

对吧？

Right?

世界上那些专注管理公开股票的基金，有时也会在早期阶段进入投资。

The the fidelities of the world, the kind of funds that manage largely public stocks, but they do get in early earlier sometimes.

有时候这很棒。

Sometimes it's great.

有时候则不然，他们进入时的估值高得离谱。

Sometimes it's not, and they come in at a crazy high valuation.

我觉得很多这类投资者在Instacart上市前，就涌入了这家估值过高的公司。

I think a bunch of those kind of folks piled into, like, Instacart, you know, when it was overvalued before it went public.

所以，这有利有弊，但确实存在这样的资本。

So, so there are pluses and minuses to that, but there is there is that kind of capital around.

这本质上就像不上市而进行IPO，你知道的，筹集这么多资金。

It's essentially like almost like going public, without going public, you know, raising that amount of money.

我觉得有趣的是，2023年当这种趋势真正兴起时，萨姆·阿尔特曼曾对他的团队说，这将成为有史以来资本投入最密集的业务。

I think it's it's interesting that in 2023 when this kind of train really got rolling, Sam Altman told his staff, you know, this is gonna be the most capital intensive business of all time.

在这个过程中，我们甚至可能从外部投资者那里筹集1000亿美元。

And we may even raise a $100,000,000,000, in in in this process from outside investors.

当时，我记得我写下了这句话。

And, at the time, that was like I remember writing that down.

我当时想，好吧。

I was like, okay.

哇。

Wow.

对此，我不知道该说什么。

That's I I don't know what to say about that.

假设软银在年底前注入这225亿美元，作为较早一轮投资的一部分，他们已经完成了60%的进度。

They're already 60% of the way there, assuming this, 22,500,000.0 from SoftBank gets in before the end of this year as part of an older round.

所以，他们实际上在讨论这下一个1000亿美元之前，已经远远超过了一半的进度。

So they're already, you know, they're already more than halfway there to to a 100,000,000,000 before, talking about this next 100,000,000,000.

但我想说，我们得到的反应大多是困惑。

But I would say that the reaction we've gotten is one of a lot of confusion.

人们基本上在问，好吧。

People are basically asking, okay.

他们有一年时间筹集了400亿美元。

They had a a year in which they they raised $40,000,000,000.

为什么他们现在或未来几个月内还需要筹集1000亿美元？

Why do they need to raise a 100,000,000,000 right now or the next few months?

这有什么急的？

Like, what's the rush?

为什么不能分阶段进行呢？

Like, why can't they do this in increments?

所以这几乎就像一个试探气球，你知道的，这类融资轮次。

So this is almost, a bit of a a a trial balloon, if you will, you know, these kind of rounds.

有很多情绪、动力以及各种非常有趣的影响因素。

There's a lot of, like, emotion and momentum and all these really interesting factors.

但从基本面来看，这家公司的估值并没有疯狂高估。

But on the fundamentals, it's it's not as if this company is crazily valued.

我的意思是，他们收入相当充足。

I mean, they they have plenty of revenue.

他们的增长速度非常快。

They're growing extremely quickly.

如果看看他们目前的收入和增长率，他们在这一轮融资中获得的收入倍数并不离谱。

The the multiple, on revenue that they would be getting in this route is not insane if you look at the revenue that they currently have in their growth rate.

而且，你知道，如果他们停止所有研发支出，实际上还是可以实现盈利的。

And the fact that, you know, if they stopped all research spending, they could actually turn a profit somehow.

这会很困难，但他们能做到。

It it would be it would be tough, but they could do it.

所以，是的，这并不离谱。

So so, yeah, it's not insane.

所以让我问你一下，劳拉，这笔钱将用于哪些方面？显然，研发支出是其中之一。

So let me ask you this, Laura, in terms of what this money is going to, there's obviously the research spending.

而其中一部分就是招聘人员。

And I mean, part of that is the people hiring the people.

但你看，这是在建设数据中心。

But look, it's building the data centers.

这是在获取芯片。

It's getting your hands on the chips.

我的意思是，那句老话，债务比股权更便宜。

I mean, the age old adage, debt is cheaper than equity.

也许这个问题很傻，但债务在这里不是一个选择吗？

Maybe it's a silly question, but is debt not an option here?

为什么他们必须选择股权呢？

Why do they have to go equity here?

这是个好问题。

That's a good question.

我的意思是，他们确实有一些债务。

I mean, they do have some debt.

我的意思是，他们离盈利还很远。

I mean, they are far from profitable.

所以他们有收入，但没有现金流，正在烧钱。

So they have revenue, they don't have cash flow, they're burning cash.

因此，这并不是理想的借款方式，也不是有利的借款条件。

So that's not exactly the best means of borrowing or the best conditions for borrowing.

我的意思是，他们拿什么做抵押借款吗？

I mean, do they borrow against something?

他们并不拥有这些芯片，或者也许拥有一些，但他们主要是租用芯片。

They don't own these chips or maybe they own some, but that's not they're renting the chips.

所以我认为，对于任何正在大量烧钱的初创公司来说，这是一个非常好的问题：你能获得债务融资吗？

So I mean, I think it's a very good question for any private startup that's burning tons of cash, can you get debt?

我认为这里存在一些问题，正如我刚刚列出的那些。

And I think there's some problems there, right, that I've just listed.

而股权融资却非常容易获得。

And equity has been very forthcoming.

我的意思是，我们已经看到了许多规模巨大的融资轮次，但我认为一个不变的事实是，这些公司似乎都能成功融资。

I mean, we've seen a lot of clearly very huge funding rounds, but I think one constant has been that all of these companies have seemingly been able to raise.

就在一年前，这家公司的估值高达1600亿美元，当时我们还在惊叹：这太疯狂了，对吧？

Just a year ago, the valuation was at $160,000,000,000 and we were writing, Wow, this is crazy, right?

而现在，我们谈论的是7.5美元，或者如果算上资金，超过800亿美元。

And now we're talking about $7.50 or over 800 if you include the money.

所以，如果存在回调迹象，对这些最大公司而言并不明显。

So if there's a sign of pullback, it's not very apparent for the largest companies.

它们的增长速度非常快。

Their growth is rising very quickly.

我们上周日的一篇文章中有一张很好的图表，展示了Anthropic和OpenAI收入增长有多快。

We had a good chart in a story on Sunday about how fast Anthropic and OpenAI's revenue has grown.

我的意思是，是一年增长三倍，还是增长了八倍？

I mean, is it tripling in a year or up eight times?

我认为，作为投资者，这正是你希望看到的。

I mean, this is what you'd wanna see, I think, as an investor.

而且我认为，正如阿米尔所说，市场上存在大量资本。

And I think, you know, like Amir said, there are very big pools of capital out there.

他们已经从共同基金、T. Rose和富达筹集了资金。

They have already raised money from mutual funds, the the T.

但还有更多这样的机构。

Rose and and Fidelity.

但还有更多这样的机构。

But there are there are many more out there.

我不知道。

And, I don't know.

关于你问到他们需要资金做什么，我认为这是这场对话中最有趣的部分，因为萨姆·阿尔特曼对这家公司的愿景和涉足领域、以及它建立的合作伙伴关系，实际上没有任何上限。

I mean, in your answer to your question about what what they need the money for, I think that is the most interesting part of this conversation because, there's really been no upper limit to how Sam Altman has envisioned this company and what it gets into and the partnerships it makes.

所以，这是否会超越？他们已经与软银合作建立自己的数据中心，这些数据中心不通过云公司来获取芯片。

So does it go beyond I mean, they're they're already kind of working with SoftBank about setting up their own data centers, that are not going through the cloud companies that are gonna be accessing chips.

他们是否会进一步推进这种垂直整合？

Do they, you know, push that verticalization even further?

我们曾报道过，他们正在研究可能自行设计芯片，或者是否在利用某种能源？

We've reported that they're, you know, working on maybe designing their chips, or do they tap some energy source?

我认为这将是我们的团队在新的一年里报道的有趣内容。

I think that's what's gonna be fun for our team to report on in the new year.

阿米尔，我问你一个问题。

Amir, let me ask you this.

有没有哪家公司曾经单独筹集过这么多资金的历史先例？

Is there any historical parallel here for a company raising a single company raising this much money?

除了显而易见的缺乏多元化风险之外，这么多资金涌入一家公司，是否会对风险投资生态系统带来更广泛的系统性风险？

Are there any broader systemic risks to venture capital ecosystem to so many people pouring into one company other than the obvious risk, which is that you're not exactly diversified?

但你能想到有什么类似的先例吗？

But can you think of there being any parallel here?

没有。

Nope.

我们无疑正处于未知领域。

We're we're in uncharted territory for sure.

不过我想说的是

The thing I would say though is that

这真是个很好的过渡。

Well, that's a that's a good segue.

也许我们就到这里吧。

Maybe we'll just call it there.

嗯，很多

Well, a lot of

许多公司已经在这笔投资上获得了丰厚的账面收益，而且这家公司也做得不错——具体效果如何，取决于你站在哪一边——它允许股东出售股份。

firms have already made, plenty of paper gains on this, entity, and it has done a good job of, depending on on which side of it you you're on, of allowing shareholders to sell shares.

因此，已经出现了相当大的流动性。

So there has been quite a lot of liquidity.

我不太记得了，是110亿还是120亿美元，仅在过去几年里，就有这么多资金通过出售股份退出，这简直令人难以置信。

I I can't remember if they're like at 11,000,000,000 or 12,000,000,000, of money that has gone out or sold, shares shares that have been sold in the last few years alone, which is pretty incredible.

所以，确实已经获得了巨大的收益。

So, yeah, there have been enormous gains already.

我认为唯一的风险在于，那些以这种估值进入的公司，以及这件事还能走多远。

I think the the only risk is for the firms that come in at this kind of evaluation and how much more can this thing, go.

我认为Chateapiti明显被低估了，所以我们还没谈到的一个故事是今天早上报道的关于广告推广的内容。

I think clearly Chateapiti is under monetized, and so one of the stories we haven't talked about is one that ran this morning, on the ads push.

也许你说得对。

Maybe you Right.

对。

Right.

其他同事很快就会谈到这一点。

Other colleagues about that, soon.

这是一次巨大的推广，而且它将被货币化、货币化、货币化，即使订阅业务在明年有所放缓，这个方向仍有很大的增长空间。

That's a massive push, and it's it's gonna be monetized, monetized, monetized, and there's plenty of runway for for that even if the subscription business, kind of, slows down in the next year.

所以我不确定。

So I don't know.

我只是从纯粹的基本面角度来看，这并不疯狂。

I just from a purely fundamentals perspective, it's it's not crazy.

他们显然拥有我们历史上前所未有的巨额支出计划。

They obviously have the most unprecedented amount of spending plans that we've ever seen, in history.

我认为公众股市并不相信这是真实的。

I think the the public equities don't believe that that is real.

他们不相信OpenAI能够真正履行其义务。

They don't believe that OpenAI can actually fulfill its obligations.

我认为筹集大量额外资本将有助于改善这一情况。

Raising a tremendous amount of additional capital, I think, will will help in that regard.

所以，是的。

So, yeah.

我想不起来最初的问题是什么了，但是

I can't remember what what the original question was, but

好吧，最初的问题只是，有没有什么相似之处？

Well, the original question was just, is there is there any parallel?

答案是没有。

And the answer was no.

所以但是

And so But

是的。

Yeah.

优步是我唯一能想到的例子，当时他们筹集了前所未有的风险资本，同时烧掉了大量资金。

Uber Uber is the the only one that I can remember where, you know, they were raising at the time unprecedented amounts of of venture capital, and they were burning a lot of money.

当时很多人怀疑他们是否能真正盈利。

And there was a ton of skepticism about whether they could ever make money.

如果你仔细审视他们的财务数据，你可能会发现类似的情况：如果他们停止砸钱去建立市场、支付司机报酬——也就是他们当时补贴乘车的高额费用——他们其实是可以盈利的。

And if you did squint and look at the financials, you could say something similar, which is that if they stopped throwing cash around to, create marketplaces and to to pay, you know, to pay drivers the the the amounts they were paying to subsidize rides the way they were, then they could actually make a profit.

很多人实际上对此表示怀疑，但事实确实如此。

A lot of people actually doubted that, but it was true.

因此，这是目前我唯一能想到的类似历史案例。

And and so that that's the only that's the only kind of historical example that I that I think about in this moment.

而且这种情况确实发生了。

And it did happen.

只是花了一些时间。

It just took a while.

我的意思是，那正是

I mean, that's

嗯，我当然知道他们确实如此，我的意思是，他们上市了，现在我们到了一个阶段，我记得优步已经盈利了。

Well, I certainly, know they had I mean, they went public, and now now we're at a point I remember Uber turned a profit.

整个股市都在看，瞧。

The whole the whole stock market went, look.

看看发生了什么。

Look what happened.

所以，我

So I

我的意思是，对于OpenAI来说，情况不同，我们报道的预测显示，几年内就能看到现金流的路径。

mean, with OpenAI, it's not you know, it's the projections we've reported on are the you you can see the path to cash flow in a couple years.

我认为问题在于，他们在真正实现之前就去测试公开市场吗？

I think the question is, do they test the public markets before that's actually really happened?

而且可能会。

And it could.

我的意思是，我们之前见过大型科技公司上市。

I mean, we've seen big tech companies go public before.

它们是盈利的。

They're profitable.

大家都拿亚马逊举例。

Everybody points to Amazon.

当然，也有一些公司在上市后仍然不盈利。

And certainly some are still not profitable in public.

所以，觉得这种情况并不疯狂。

So, it's it just doesn't seem that crazy to think of.

只是这些公司的规模、支出和现金消耗都非常高。

It's just that the the scale of these, the spending and, the cash burn is quite high.

对。

Right.

当然，看看事情如何发展，这确实是个有趣的故事。

Well, it is certainly an interesting story to see how it plays out.

我期待你们两位团队就这个话题提供更多报道。

I'm looking forward to more reporting from both of your teams on the topic.

感谢你们的到来。

I want to thank you for coming on.

这位是我们的执行主编劳拉·门德罗，以及我们的联合执行主编阿米尔·阿夫罗蒂，他们都来自《信息》杂志。

That is Laura Mendero, our managing editor, and Amir Afrotti, our co executive editor here at The Information.

好的。

Okay.

在结束之前，我们还有一个故事要分享，这是一个圣诞故事。

We have one more story for you before we go, and it is a Christmas story.

我们的埃隆·马斯克记者西奥·韦特今天发表了一篇关于他家乡田纳西州孟菲斯的文章。

Our Elon Musk reporter, Theo Wait, has a piece out today about his hometown, Memphis, Tennessee.

这个标题非常棒。

It is a great headline.

X AI的圣诞灯饰揭示了SpaceX上市的哪些信息。

What x AI's Christmas lights say about the SpaceX IPO.

我想请Theo来聊聊这件事。

I wanna bring on Theo to talk all about it.

嘿，Theo。

Hey, Theo.

最近怎么样？

How you doing?

嘿，Akash。

Hey, Akash.

圣诞快乐。

Merry Christmas.

谢谢你用你的故事带来了一些圣诞气氛。

Thank you for bringing some of the Christmas spirit with your story.

很棒的文章，我们期待在简报中阅读它。

A good piece, we're looking forward to reading it in the briefing.

但我想趁这个机会营造一下节日氛围，所以我准备了一个小把戏，稍后马上给你展示。

But I thought we'd get in the spirit here, and so I've got a little bit of a trick I want to show you at the snap of a fingers.

怎么样？

How's that?

哇。

Wow.

很酷吧？

Pretty cool, right?

太棒了。

That's amazing.

是的。

Yeah.

感谢我们的制作团队让这一切成为可能。

Well, thank you to our production team for making that happen.

其实这不是我的主意。

Wasn't my idea, actually.

是他们的主意。

Was theirs.

好吧，我们来聊聊孟菲斯和孟菲斯人，我之前都不知道当地人这么称呼自己。

Okay, let's talk about Memphis and the Memphians, which I did not know is what they call them in Memphis.

你有一个关于圣诞灯饰的故事，以及它与数据中心的关系。

You had a story about Christmas lights and how that relates to data centers.

所以，我把话筒交给你。

So I'm gonna I'm gonna give the floor to you.

确实如此。

It's true.

是的。

Yeah.

X AI 给了我一个完美的隐喻，让我能在圣诞节期间谈论人们对数据中心以及企业试图赢得邻居支持的普遍反感。

X AI kind of handed me the perfect metaphor to talk about on on Christmas for the kind of general backlash against data centers and and and companies, you know, efforts to win over their neighbors.

孟菲斯有一个历史悠久的圣诞灯展，名叫‘星光之夜’，位于城市郊区的一个大型公园里，人们开车绕行，观赏这些昂贵而壮观的灯饰。

So there there's this long running Christmas light show in Memphis called Starry Nights, and it's in this huge park, kind of on the outskirts of the city, and people go and drive their cars around and look at these expensive large light shows.

而今年，这个灯展被命名为‘由 X AI 赞助的星光灯饰’。

And this year, it's called Starry Lights presented by x AI.

如果你去看看Starry Lights的Instagram账号和Reddit帖子，会发现很多评论者对此非常不满，有人呼吁抵制，甚至说：‘我无法相信我热爱的这个本地公园会接受埃隆·马斯克的钱。’

And if you go and and look at the Starry Lights Instagram accounts and and Reddit threads and things, there are lots of people in the comments that are, you know, pretty upset about this and, you know, calling for boycotts or saying, you know, I can't believe this local park that I love would take money from Elon Musk.

这仅仅是XAI在孟菲斯存在的最新例证——它在该市拥有两个大型数据中心，已成为民众不满的焦点。

And it's it's just kind of the latest example of XAI's presence in Memphis where it has two very large data centers kind of being a flashpoint for for people.

好吧。

And Okay.

所以人们对于这些圣诞灯饰感到不满。

So so the people are upset about the Christmas lights.

政界人士也生气了吗？

Are the politicians upset?

目前孟菲斯掌权的政界人士普遍对XAI持温和甚至大力支持的态度，因为他们正是当初把XAI引进孟菲斯的人。

Generally, the politicians in Memphis that are in power currently are are pretty lightly to extremely pro x AI because they're the same ones that kind of brought x AI to Memphis in the first place.

这种情况才发生了十八个月而已。

Like, this has only been happening for eighteen months.

所以，实际上还没有举行过选举。

So, you know, there haven't really been elections yet.

话虽如此，但明年中期选举即将到来，一些初选挑战者和其他因素让人们对这件事变得更加怀疑。

That being said, though, there are a number of, you know, primary challengers and, you know, other things coming up for the midterms next year that people are being a lot more skeptical.

好的。

Okay.

现在，跟我谈谈你认为这与埃隆在太空中的数据中心野心有什么关系。

Now, talk to me a little bit about what you think this has to do with Elon's data center ambitions in outer space.

好的。

Okay.

与此同时，随着xAI的发展，埃隆一直在为SpaceX明年的潜在IPO造势。

So, you know, at the same time this has been happening with xAI, Elon has been, you know, hyping up SpaceX ahead of a possible IPO next year.

而他最近谈论的最新一种未来科技，就是将数据中心放在太空中。

And kind of the latest, you know, flavor of futuristic technology that he's talking about is is putting data centers in space.

他说，要实现我们满足他对人类梦想所需的AI规模，唯一的办法就是把数据中心放到太空中，因为太空中有巨大的能源，而且更容易为数据中心降温。

And he's saying the only way to scale AI to, you know, the degree that we need to fulfill his his dreams for humanity is to put data centers in space because there's so much power up there, and it's easier to cool to to cool data centers.

而且，基本上来说，地球上存在的所有资源限制，在太空中都不以同样的方式存在。

And, you know, it it basically, like, all the resource constraints that exist on Earth don't exist in space in the same way.

许多持怀疑态度的人正确地指出，把数据中心送到太空非常非常昂贵。

And a lot of skeptics, you know, rightfully say it's very, very expensive to put data centers in space.

如果芯片坏了怎么办？

What if the chips fail?

这些工程和财务上的障碍都是真实存在的。

Like, there there are all these kind of engineering hurdles and financial hurdles that are totally real.

但在我看来，只有当你把太空中的数据中心视为对地球上的反对力量的一种对冲时，它们才有意义。

But, you know, in my opinion, I I I think the only way the data centers really make sense in space is if you think of them as a hedge for opposition on Earth.

我认为，许多科技界的人可能正在意识到，未来几年内，地球上的数据中心遭到强烈抵制的可能性虽然很小，但确实存在，比如5%左右，以至于你根本无法再建造它们。

Like, I think, you know, a lot of people in in tech are probably coming to realize that there's maybe a, you know, a small percentage chance, you know, maybe 5% or whatever, that in the next few years, the backlash to data centers on Earth will be so big that you won't really be able to build them.

在这种情况下，建造太空数据中心的成本高低就不再重要了，因为那将是你唯一能建造它们的方式。

And in that scenario, all of a sudden, it doesn't really matter if building data centers in space is expensive because it'll be the only way you can build them.

好吧。

Okay.

所以我只是想理清一下这里的论点。

So I'm just trying to track the argument here.

所以你的意思是，如果地球上对这些数据中心的反对如此强烈，从孟菲斯人们对圣诞灯饰的反应开始，如果反对情绪如此之大，那么太空中的数据中心反而可能变得合理。

So you're saying that if there is so much backlash on earth to these data centers, starting with the people reacting to the Christmas lights in Memphis, If there's so much backlash, then maybe the data centers in outer space actually make sense.

这就是你的意思。

That that's what you're saying.

是的。

Yeah.

我的意思是，当然，我把这和圣诞灯饰联系起来有点荒谬。

I mean, you know, I I'm being a bit silly by connecting it to the Christmas lights, of course.

不。

No.

我同意你。

I'm with you.

我同意你。

I'm with you.

但没错。

But but yes.

但是，是的。

But, yes.

我认为，圣诞灯饰只是一个小小的例子，说明了人们普遍有一种这样的感受

I think the Christmas lights are a small example of how people in general, you know, have this feeling

关于这种大规模的反对情绪。

about the the backlash the backlash at large.

是的。

Yeah.

对。

Right.

是的。

Yeah.

对。

Right.

我可以问你一下吗？你是否广泛报道过美国各地对这些数据中心的反对声音？

Can I ask you, you know, you have covered the backlash to these data centers broadly around The US?

你认为这会走向何方？

Where do you see this going?

你认为会不会真的有数据中心不得不关闭或搬迁？

Is this something where are there likely to actually be data centers you think that may have to shut down, may have to move?

我的意思是，这些小镇上反对这些项目的努力有多强？你知道的，至少为他们改善一下状况？

I mean, like, how strong are these efforts in these local towns to, you know, at least make things better for them?

在孟菲斯，这是一个非常具体的本地问题，因为明年将选举县市长，理论上，县市长可以控制卫生部门，从而阻止XAI使用发电机为其数据中心供电。

So in in Memphis specifically, you know, it's a it's a pretty particular local issue where there's election for county mayor next year, and theoretically, the county mayor could control the health department, which could then make it so XAI couldn't use generators to power its data center.

这在当地政治中有点复杂，但这也是全国范围内正在发生的事情的一个极端例子。

And it's, like, kind of a convoluted local politics thing there, but, you know, it is kind of an extreme example of what's going on around the country.

就在本月早些时候，参议员伯尼·桑德斯呼吁在全国范围内暂停新建数据中心。

Like, earlier this month, senator Bernie Sanders was, you know, calling for a nationwide moratorium on new data centers.

但参议院不太可能通过这项法案，或者让特朗普很快签署它。

And, like, the senate's not gonna go and, you know, pass that or, you know, have Trump sign it anytime soon.

但总的来说，趋势似乎正在发生一些变化。

But, you know, in general, the the tides do seem to be possibly shifting some.

比如，罗恩·德桑蒂斯，几年前还在总统初选中得到埃隆·马斯克的支持，现在却频频发表反数据中心、反人工智能的言论。

Like, you know, Ron DeSantis, who, you know, a couple of years ago was, you know, getting endorsed by Elon Musk in the in the primary for for presidency, you know, now is going out and making pretty anti data center, anti AI statements.

确实，两党中都有些政客在这方面发声很响亮，而且觉得这可能是个能赢得选票的议题。

Like, there there are, you know, some politicians in both parties that are being pretty loud about this and, you know, see it as a potentially winning issue.

但这些措施要真正付诸实施，恐怕还得好几年。

But it is, you know, almost certainly years away from, you know, coming in in action.

对。

Right.

这些数据中心的一个关键是，最大的那些目前还处于规划阶段，所以我预计，一旦这些数据中心开始运营并消耗大量电力，这类讨论会越来越多。

Thing about these data centers, obviously, is the biggest ones are still in the planning stages, and so I anticipate we will see more of these discussions popping up once these data centers actually start operating and using a lot more of the power.

所以，进入新的一年，这确实是个值得持续关注的有趣故事。

So it's certainly an interesting story to keep track of heading into the new year.

西奥，非常感谢你今天来做客。

Theo, I want to thank you very much for coming on.

这位是西奥·韦特，我们的埃隆·马斯克记者，为TI电视带来全部的圣诞气氛。

That is Theo Waite, our Elon Musk reporter, bringing all the Christmas spirit here on TI TV.

谢谢，阿卡什。

Thanks, Akash.

好的。

Okay.

好了，今天的节目就到这里。

Well, that does it for today's show.

我们明天和周五放假过节。

We are off tomorrow and Friday for the holidays.

我们下周一上午10点（太平洋时间），下午1点（东部时间）再见。

We will be back on Monday at 10AM Pacific, 1PM Eastern.

我要感谢亚马逊网络服务公司，作为本节目的冠名赞助商，也感谢各位的收看。

I want to thank Amazon Web Services who is our presenting sponsor for this production, and I want to thank you for tuning in.

我们非常感谢大家的支持。

We really do appreciate your viewership.

祝大家圣诞快乐。

Merry Christmas to you all.