标普 500 指数对未来 AI 上市的影响

欢迎收听本档播客。

Welcome to the podcast.

首先，标普500指数拒绝为SpaceX加入开绿灯，坚持原有规则。

Kicking this off, the S and P five hundred is refusing to change the rules for SpaceX to join the S and P five hundred.

这也会让OpenAI和Anthropic在即将来临的IPO后，无法走快速通道加入。

This is also going to block OpenAI and Anthropic when they do their inevitable IPOs here coming up quickly from joining on a fast track.

另外有个消息，谷歌每月要付给SpaceX九亿二千万美元，租用他们的AI算力。

We also have news that Google is going to be paying SpaceX $920,000,000 a month to rent x AI compute capacity.

这还不包括Anthropic目前已经做的那些。

This is on top of what Anthropic is already doing.

企业们在AI令牌预算上的花费，已经超出了2026美元的限额三倍之多，很多人正头疼该怎么管住这笔开销。

Enterprises are blowing past their $20.26 AI token budgets by three x, and a lot of them are trying to figure out how to put controls on this.

Airbnb的首席执行官布莱恩·切斯基，正在创办自己的AI实验室。

The CEO of Airbnb, Brian Chesky, is starting his own AI lab.

他以前曾给萨姆·阿尔特曼当顾问，我觉得他还会继续提供支持，但现在他要自己搞一个实验室，很可能会直接跟对方竞争。

He's kinda was formerly advising Sam Altman, and I think he's gonna keep doing that, but now he's gonna be spinning up a lab which could be, you know, directly competing.

说到竞争对手，目前大约有90位投资者同时持有OpenAI和Anthropic的股份。

And speaking of competitors, about 90 investors currently hold stakes in both OpenAI and Anthropic.

我想具体分析一下，为什么会出现这种情况，以及为什么在硅谷这种事其实相当少见。

I wanna break down some of the numbers about why this is and why it is pretty uncommon for this type of thing to happen in Silicon Valley.

如果想每天收到这些科技动态的邮件摘要，欢迎订阅《AI Chat》简报。

I wanted to mention, if you wanna get all of these news stories in an email to your inbox every single day, go subscribe to the AI Chat newsletter.

网站是aichatdaily.com，是我自己运营的平台。

It's on aichatdaily.com, my own site.

点击页面右上角的订阅按钮，就可以轻松加入。

You can click the subscribe tab in the top corner, and you can subscribe to this newsletter.

完全免费。

It's free.

每天准时发送，每一条新闻我都亲自筛选核实。

It goes out daily, and I vet every story, what's going on.

我会重新整理内容，补充一些播客里没来得及展开的细节。

I reorganize it and try to give you basically all the extra details that sometimes aren't covered in the podcast.

你可以看到深入的分析，或者如果你没时间听播客，我也会在描述里附上 aichatdaily.com 的链接。

You get a deep dive or sometimes if you don't have time to listen to the podcast, I'll leave a link in the description to aichatdaily.com.

我简直惊呆了。

I'm blown away.

打开率高达百分之四十。

It has a 40% open rate.

听这个通讯的人特别喜欢它。

People that listen to it love it.

我们的读者遍布各大顶尖人工智能公司，赶紧去看看吧。

We have listeners listeners from all the top AI companies, so go check it out.

链接在描述里。

Link in the description.

好的。

Okay.

咱们聊聊标普500最近的情况。

Let's talk about what's happening with the S and P 500.

他们拒绝为SpaceX放宽规则，而这将来也会阻碍OpenAI和Anthropic。

They have refused to bend the rules for SpaceX, and, of course, this is gonna block OpenAI and Anthropic in the future too.

SpaceX曾申请跳过一些常规流程，快速加入标普500，但标普500拒绝了。

But basically and SpaceX went and they requested that they could skip kind of some of the standard rules for getting fast tracked into the S and P 500, and the S and P 500 said no.

SpaceX为什么想进这个指数，原因很明显。

It's pretty obvious why SpaceX wanted it.

只要能入选，就能自动带来一千四百亿美元的被动基金买入。

It would have basically given them $14,000,000,000 in automatic passive fund buying just by getting included into this.

但标普500确实有明确的入选标准。

But the S and P actually has rules to get included.

其中一条是必须有十二个月的上市观察期。

One of them is a twelve month seasoning period.

也就是说，公司上市后得等满一年才能被纳入。

So basically, once you IPO, it's gotta you gotta wait twelve months before you can get added.

这主要是为了防止公司短期暴涨暴跌，避免对标普500指数造成负面影响。

I guess this is just to make sure that companies don't just boom and bust, and it doesn't have any sort of negative impact on people in the S and P 500.

还有个10%的流通股要求，再加上盈利门槛，现在很多大型IPO公司，尤其是这些AI公司，目前根本达不到。

There's also a 10% float requirement, and there's a profitability screen that some of these huge IPOs, especially these AI models, don't really have right now, these AI companies.

所以综合来看，SpaceX其实并不符合资格。

So because of all of that, SpaceX didn't actually qualify.

这么做的目的，是为了保护标普500这个规模高达7.5万亿美元的基金，避免它被那些没有盈利、市值虚高但长期前景不明的公司拖累。

And the idea here is that it's gonna protect the $7,500,000,000,000 fund, the S and P 500, right, one of these mega funds, from basically absorbing any sort of unprofitable or these companies that have these huge caps, but maybe they're they're not gonna be successful long term.

那这是否意味着SpaceX长期来看不会成功呢？

Now is this some sort of indication that SpaceX isn't gonna be successful long term?

我不这么认为，他们只是想说：看，

I don't think so, but I think they're just saying, look.

我们有明确的规则。

We have these rules.

你得遵守这些规则。

You gotta follow them.

虽然SpaceX因此错过了大约140亿美元的被动资金流入，但OpenAI和Anthropic也同样无法入选——根据彭博情报的估算，如果它们早些时候能加入，OpenAI会错过80亿美元，Anthropic则会错过46亿美元。

And while SpaceX is missing out on about $14,000,000,000 in automatic passive buying, OpenAI and Anthropic are both gonna you know, because they're not also gonna get included, they're both gonna miss out on $8,000,000,000 for OpenAI and about $4,600,000,000 for Anthropic according to Bloomberg Intelligence if they had kind of got in there early as well.

但我觉得，既然他们拒绝了SpaceX，那另外两家也基本没戏了。

But I think because they're saying no to SpaceX, those other two are basically nos as well.

纳斯达克100指数将在15个交易日内接纳SpaceX，富时罗素的罗素500指数则会在5天内加入，这比他们平时的节奏快得多，但这些指数的规模远不能和标普500相比。

The Nasdaq 100 is going to let SpaceX in fifteen trading days, and FTSE Russell's Russell top 500 is gonna let it in in five days, which is far faster than they normally do, but these are in no way as big in size as the S and P 500.

所以虽然规模小一些，但SpaceX还是会获得一些被动资金流入。

So there's gonna be smaller, but still there it's gonna get some passive inflows.

我觉得这很重要，因为标普500现在这个做法，其实是在强调指数规则的严肃性。

The reason why I think this matters is because S and P 500 right now, this is sort of, interesting move for index discipline.

对吧？

Right?

他们就是想说：我们得按规矩来。

They're they're they're just saying, look.

规则比资金流动更重要。

We're gonna follow our rules over the capital flows.

他们确实错过了不少钱。

They're they're missing out on a lot of money.

很多人可能会直接投资SpaceX，其实他们只要买标普500，就能间接拥有SpaceX的股份。

A lot of people may go and invest their money directly into SpaceX when they could have just bought the S and P five hundred knowing that they would get a piece of SpaceX.

所以现在他们要么从标普500里撤资，要么把原本打算投进标普500的钱，直接转去投资SpaceX，只为拿到这支股票的敞口。

So now they have to go and take either their money out of the S and P five hundred or put money that they might have put into the S and P five hundred alternatively and put it directly into SpaceX to get exposure to this stock.

尤其是像OpenAI、Anthropic和SpaceX这些公司，一旦估值达到万亿级别，根据规则，至少十二个月内都不会被纳入标普500。

Specifically, as these companies are, you know, hitting these trillion dollar valuations, OpenAI, Anthropic, and SpaceX, all of those are not gonna be in the S and P 500 for at least twelve months because of the rules.

所以如果你想分一杯羹，只能直接下手。

And so if you wanna get a piece of those, you gotta go direct.

标普500这个决定挺有意思，但我还是认可他们这种坚持规则的态度。

So an interesting move by the S and P 500, but I do respect some of the discipline there.

谷歌每个月要付给SpaceX九亿二千万美元，租用他们的AI计算能力。

Google is going to be paying SpaceX $920,000,000 a month to rent x AI compute capacity.

这笔交易是为了获得十一万台NVIDIA显卡的算力支持。

This is to get access to a 110,000 NVIDIA GPUs.

这个合作将持续到2029年6月。

They're gonna get this through June 2029.

这是一份长达32个月的合同，而就在SpaceX即将筹备1750亿美元的IPO之际，这个消息刚刚被披露出来，IPO就在下周。

It's a thirty two month deal, and this is just being disclosed as SpaceX is prepping for their $1,750,000,000,000 IPO, which is happening next week.

这么大的IPO马上就要来了，他们却突然扔出这个消息。

So we have this massive IPO coming up, and they've they've dropped this news on us.

很明显，SpaceX想在IPO前尽可能提升收入。

Obviously, SpaceX is trying to increase their revenue as much as possible ahead of this IPO.

这笔交易将从今年10月开始生效。

The deal is gonna run starting in October this year.

如果他们真的能大获成功——比如现在正在全力训练Grok——那到时可能早就把事情搞定，顺利收尾了。

So if they have a massive run, you know, they're training something huge with Grok right now, they could probably have it everything kind of done, buttoned up, and moved on by the time this is happening.

很多人在问，这是否意味着Grok的使用量没达到SpaceX的预期？

A lot of people are saying, well, does this mean that, you know, SpaceX isn't getting the usage they hoped for Grok?

但不管怎样，他们都在大把赚钱，看来状态不错。

Either way, they're making a lot of money, so I think they're probably in a happy place.

这笔交易涵盖GPU、CPU和内存组件，全部部署在SpaceX的数据中心里，有效期将持续到2026年。

But this is gonna cover GPUs plus CPUs and memory components, and it's all gonna be inside of SpaceX's data center, and this is going through 2026.

这是SpaceX的第二笔大规模计算资源交易。

This is SpaceX's second huge compute deal.

众所周知，Anthropic上个月签下了为期一个月12.5亿美元的合同，租下了整个Colossus 1孟菲斯设施。

Anthropic like we know, they signed a $1,250,000,000 a month agreement last month for the entire Colossus 1 Memphis facility.

谷歌的Alphabet在2026年的资本支出预计将达到1800亿到1900亿美元。

Google's Alphabet is facing a 180 to $190,000,000,000 in 2026 CapEx.

他们本周宣布了850亿美元的股票发售，用来筹措这笔资金。

They announced an $85,000,000,000 stock sale this week to help fund that.

我觉得这表明他们预计2027年的AI基础设施成本还会更高。

I think they're showing that they're expecting even larger AI infrastructure costs in 2027.

说白了，大家都想要更多算力，而SpaceX似乎提前建得太多了，现在正把多余资源卖给Anthropic和谷歌。

So basically, everybody wants more compute, and SpaceX had kind of an overbuild it would seem as they're kind of selling off to Anthropic and to Google.

当然，有些人可能会说，嘿。

And while, you know, some people would have said like, hey.

你看。

Look.

你们这次过量建设可要亏大了。

You guys are gonna be in the hole for overbuilding.

结果呢，你们完全可以转手把它卖个好价钱，给那些对AI算力需求无穷无尽的竞争对手。

Well, turns out you can just turn around and sell it to all of your competitors who seem to have, an endless demand for their AI.

尤其值得注意的是，SpaceX的AI部门每个季度亏损25亿美元，但已经锁定了每月20亿美元的合同收入。

This is particularly important when you look at the fact that SpaceX's AI unit is losing $2,500,000,000 a quarter, and they have $2,000,000,000 in monthly contracted revenue locked in.

所以我觉得他们会继续想办法把这笔收入再往上推。

So I think they're going to try and make sure they can keep pushing that number up.

我预计可能还会有更多公司来找他们签合同，买下一部分算力资源。

I would expect perhaps that we might even see more people, signing deals to buy some of their compute.

明白了。

Alright.

目前企业们的AI算力预算已经远远超出了202.6亿美元的限额。

Enterprises right now have blown past their $20.26 AI token budgets.

超出了整整三倍。

They've done it by three x.

很多人现在都在忙着控制成本。

A lot of them are scrambling for controls.

我之前在播客里聊过优步。

I've talked about Uber on a previous podcast.

优步在四个月内就把全年编程预算花光了。

Uber burned through their entire coding budget in four months for the whole year.

还有一家公司因为没给开发者设置使用上限，收到了五亿美元的Claude账单。

There's another company that is facing a $500,000,000 Claude bill after they failed to set usage limits for their developers.

我觉得眼下这场爆炸式增长，主要来自去年十一月发布的那些AI模型。

I think this massive explosion right now is coming from a lot of these AI models that were released in November.

去年十一月，Anthropic突然推出了几款特别出色的模型。

In November, it felt like Anthropic just had some really incredible models come out.

从那以后，我个人就转用Anthropic作为主力工具，不再用OpenAI了，还开发了很多SaaS产品、应用程序，做了不少酷炫的项目。

Personally, since then, I've been using, you know, I switched from OpenAI to Anthropic being my primary tool, and I built a ton of SaaS, a ton of apps, a ton of really cool projects.

这确实很有趣，但你很容易就把使用量耗光，尤其是Anthropic刚推出了工作流功能，你只要说一声‘去完成这个复杂又冗长的任务’，它就真能干下去。

It's a lot of fun, but you definitely can burn through an insane amount of usage, especially when Anthropic just released their workflows, which, you know, you can tell it like, hey, go do this really long lengthy complex task.

它能同时启动几十个下属代理，让它们一起运行。

It can spin up dozens of agent agents below it and get them all running at the same time.

这样一来，你的令牌消耗速度就快得惊人。

So you just burn through so many tokens.

就我而言，我用的是Quad Max订阅，每月200美元。

Now in my case, I'm using the Quad Max subscription, is $200 a month.

不过这个价格其实有大幅补贴，按我的估算——当然可能有人不认同——如果你把使用量拉到极限，每月相当于能用掉5000到11000美元的Anthropic API费用。

But from and and it's very heavily subsidized, I I think according to my calculations, but, you know, some people may disagree with me, you can use somewhere between 5 and $11,000 a month in what would be, you know, Anthropic's API costs if you absolutely max it out.

所以，没错，每月200美元确实划算，但很多人根本用不上最高档的套餐。

And so, yes, $200 a month, great deal, but a lot of people don't have the max plans.

很多人都是直接调用API，按需付费。

A lot of people are just grabbing directly from the API.

他们公司现在只允许用GitHub这类不再有补贴的平台。

Their other company is just letting them use, you know, something like GitHub that no longer has a subsidized plan.

你只能按使用量一块一块地买API令牌。

You just have to buy the the API tokens as they come.

在这种情况下，你在开发过程中可能会花掉一笔惊人的钱。

And so in that situation, you can be spending an insane amount of money as you build things.

这之所以重要，是因为Priceline在续签光标合同时，花费比预期高出四到五倍。

Now the reason why this is important, Priceline's cursor contract renewal cost them four to five x more than they were expect expecting.

在他们追踪的多家公司中，每位开发者的令牌消耗量在九个月内飙升了18.6倍。

Their per developer token consumption rose 18.6 x in nine months across a bunch of different companies that they were tracking.

根据Jellyfish的研究，顶级令牌使用者的生产力是轻量级用户的两倍，但为了获得这些提升，他们消耗的令牌却是十倍之多。

The top token users were two times more productive than Lite users, but they burned 10 times more tokens to achieve those gains according to jellyfish research.

我觉得，有些人可能会说，这根本不值，因为你虽然比轻量用户效率高一倍，但用掉的令牌却多得离谱。

Now I think, you know, some people would be like, well, you know, it's not even worth it because you might be two times more productive than the light users, but you're only you know, you're using so many more tokens.

归根结底，最好的办法就是申请最高档的套餐，因为它是有补贴的。

At the end of the day, the solution is getting the max plan, is subsidized.

但关键问题是，这种补贴能持续多久？我觉得不可能永远持续下去，毕竟Anthropic不可能无限补贴下去。

How long will that last is the real question, and I don't think it can last forever because Anthropic can't subsidize this forever.

但在这期间，如果你是企业用户，我强烈建议别直接通过API调用，一定要选有补贴的套餐。

But in the meantime, if you are an organization, I 100 recommend, not going through just, you know, directly using APIs.

你可以大幅降低成本。

You can cut you can cut costs dramatically.

本周，Linux基金会推出了Tokenomics基金会，旨在统一AI成本衡量标准，已有180家供应商参与。

The Linux Foundation launched the Tokenomics Foundation this week to standardize AI cost metrics with a 180 vendors already participating.

看来其他人也在尝试解决这个问题，让成本变得更透明一些。

So it seems like some other people are trying to, tackle this problem and put a little transparency into this.

Airbnb的首席执行官切斯凯即将创办自己的AI实验室。

Airbnb's CEO, Chesky, he's gonna be starting his own AI lab.

这个实验室将专注于用户交互与设计。

It's gonna be focused on user interaction and design.

大家都知道，Airbnb以设计著称。

Everyone knows that Airbnb is famous for design.

布赖恩一直被认为是那里的设计负责人。

Brian is famous for being kind of the the head of design there.

他对这个领域着了迷。

He's obsessed with it.

每当我考虑如何设计一个应用或工具的界面和用户体验时，我总会去看看Airbnb是怎么做的，找找参考。

It's something that I, you know, whenever I'm thinking about how to design the UI, UX on an app on a tool, I always go and look at what Airbnb is doing and try to find, some precedent there.

所以名气很大，而这正是他专注的方向。

So very famous, and this is what he's focusing on.

但他还是会继续担任Airbnb的首席执行官，同时直接跟OpenAI竞争。

But he's actually gonna keep his CEO role of Airbnb, and he is going to be directly competing with OpenAI.

这家公司他其实已经指导好多年了。

This is a company that he's advised for years.

他说现有的AI工具在旅行和电商领域缺乏丰富的界面设计，说实话，我觉得很多AI工具确实不擅长设计。

He said he believes that the existing AI tools lack the rich interfaces needed for travel and ecommerce, which to be fair, I think that a lot of these AI tools are not great at design.

不过也有办法绕开这个问题。

There's ways to get around it.

我最近用得最好的方法就是，截一张我欣赏的网站截图，发给Claude，让它照着这个风格重新做，效果确实不错。

I think the best way that I've been using lately is just screenshotting a website that I like the design of, giving it Claude and telling it to, you know, make it in that same style, and and it seems to do better.

但如果你只是简单地让它‘建个网站，实现X、Y、Z功能’，那出来的设计肯定不会好看。

But if you just ask it stock like, hey, build a site, build a thing that does x y z, the design is not gonna be a beautiful site.

所以我认为，布赖恩对Airbnb的观察正是这一点。

And so I think this is what Brian is seeing specifically for Airbnb.

不过，他本人并不会亲自管理这个新实验室。

He's not gonna be running this new lab himself though.

他会任命另一位高管担任CEO，而自己则担任创始主席。

He's gonna have another executive that's gonna be the CEO, and he is going to act as the founding chair.

这本质上是他以创始人身份深度参与Airbnb运营的方式。

It's basically kind of his hands on founder mode approach at Airbnb.

据布赖恩说，Airbnb之所以拒绝与ChatGPT插件合作，是因为底层工具还不够稳固。

Airbnb has rejected ChatGPT plug in partnerships because underlying tools aren't robust enough, according to Brian.

但这也可能为他打造一个竞争性实验室铺平道路，就像埃隆·马斯克那样，创立了XAI，直接整合到X平台，并用Grok作为核心模型。

But also, this might set the stage for him to create a competing lab and use his own tool, kind of like how Elon Musk, you know, created x AI and plugged it straight into x and had Grok as kind of the big model there.

如果Airbnb直接使用你开发的模型，你很可能很快就能获得大量用户。

You could probably make a model and get a lot of usage if Airbnb was using that directly.

布拉德·阿德科克的Hark去年底才上线，我认为它也在做类似的押注，聚焦于AI模型在界面和体验上的短板。

Brad Adcock's Hark launched late last year, and I think it's doing something in making kind of a similar bet, focusing on the fact that AI models are not great at UI and UX.

大约有90位投资者同时持有OpenAI和Anthropic的股份。

About 90 investors hold stakes in both OpenAI and Anthropic.

这种情况在硅谷可不常见。

Now this is not something you see in Silicon Valley a lot.

人们通常不会同时押注两个竞争对手。

People don't usually bet on two different competitors.

但根据最新报道，42%的OpenAI投资方也持有Anthropic的股份，这看起来像是投资者在试图对冲风险。

But according to a new report, 42% of OpenAI backers also hold Anthropic stakes, which feels a lot like investors are trying to hedge their bets on this.

投资方包括红杉资本、Greylock创始基金和红点创投。

We have Sequoia Capital, Greylock Founders Fund, and Redpoint Redpoint Ventures.

这些机构同时在两家公司的股东名册上。

They all sit on both cap tables.

Anthropic最新一轮融资的31家投资机构中，至少有13家已经持有OpenAI的股份。

At least 13 of the 31 invested names in Anthropic's latest raise already own OpenAI shares.

这两家公司在IPO前都完成了超过1000亿美元的融资，估值基本都在一万亿左右，大概在8500亿到9500亿之间。

Each company has raised over a $100,000,000,000 at, you know, basically a trillion dollar valuations ahead of their IPOs this year, close to a trillion dollars, you know, somewhere between 850, 950.

但他们融的钱实在太多了，以至于传统意义上的利益冲突和治理问题基本就不存在了。

But they're raising so much money that I think a lot of the traditional conflict of interest type of governance concerns are basically gone.

这些融资规模都太大了。

These are these are massive raises.

大约30%的重叠投资者是对冲基金、私募股权或财富管理机构，这些人早就习惯在竞争对手之间分散投资了。

Roughly 30% of the or 30 of the overlapping investors are hedge funds, private equity, or wealth managers that are these people are very used to spreading their bets across competitors.

我觉得传统的风投公司现在也开始采纳这种策略了。

Traditional venture firms, I think, are now starting to adopt that same strategy.

说白了，大家都想在大 IPO 季节到来前，分一杯羹。

It feels like everybody wants a piece of every pie before the big IPO season is about to be honest.

各位，非常感谢收听这个播客。

Guys, thank you so much for tuning in to the podcast.

如果你还没试过 AI Box，现在每月花 8.99 美元，就能在一个平台访问 80 多个 AI 模型。

If you haven't tried AI Box out already, you can access 80 different AI models in one place for $8.99 a month.

这是我自己的创业项目。

This is my own startup.

你可以使用所有图像、音频、文本和视频模型。

You get access to all of the image, audio, text, and video models.

真心推荐你试试看，告诉我你的感受。

I would love for you to try it out and let me know what you think.

不用再在多个不同AI模型的标签页之间来回切换，也不用订阅一堆服务了。

You can stop juggling tons of tabs with different AI models, tons of subscriptions to different places.

所有功能都集中在一个地方，还能在同一个对话线程里跟它们全部交流。

You can get everything in one place and chat with all of them in the same chat thread.

特别方便，链接在简介里，aibox.ai。

It's super useful, and the link is in the description at aibox.ai.

好的。

Alright.

下一期再见。

I'll catch you guys all in the next episode.