The 'people’s airline,' SpaceXAI, and the Enterprise AI Race

大家好，欢迎回到《Equity》，这是TechCrunch关于初创企业商业的播客。

Hello, and welcome back to Equity, TechCrunch's podcast about the business of startups.

今天是5月8日，星期五，我是TechCrunch的交通板块编辑柯尔斯滕·科罗塞克。

Today is Friday, May 8, and I'm Kirsten Korosec, transportation editor here at TechCrunch.

和往常一样，我身边坐着我们的资深记者肖恩·奥凯恩，还有周末版编辑安东尼·哈。

And I'm joined, as always, by our senior reporter Sean O'Kane and Weekend editor, Anthony Ha.

精神航空已经破产了。

Spirit Airlines went bankrupt.

它已经停摆了。

It's folded.

这事算是结束了。

It's done.

但我还是想提一下，因为现在有人在努力让它重新复活。

But I wanted to mention it because there is an effort to bring it back.

安东尼，我想问问你，你有没有做过什么承诺

And, Anthony, I'm wondering if you if you made a pledge

在TikTok上。

on TikTok.

我还没做任何承诺，至少在这档播客里我不会承认我有。

I have not made a pledge yet, or at least I'm I'm not admitting to it on this podcast.

不过，就像你说的，确实有公告说精神航空要破产了。

But, yeah, as as you said, there was this announcement that Spirit Airlines is going bankrupt.

而且，说实话，整个过程看起来特别突然。

And, I mean, it was it seemed like a very, very sudden process.

很多人都被困住了，不得不紧急订新航班，诸如此类的事情。

People were stranded, had to, like, frantically book new flights, all of that.

然后，这位TikTok博主亨特·彼得森说：嘿。

And then this TikToker, Hunter Peterson, said, hey.

如果我们大家一起集资，把精神航空买下来，改成全民航空公司，怎么样？说实话，我觉得这个想法在好几个层面都挺有意思的。

What if we all got together and crowdfunded and and bought Spirit Airlines and turned it into the people's airline, which, I mean, I I find for interesting on multiple levels.

但其中一点是，我并不觉得精神航空是那种特别受人喜爱的航空公司。

But but one of them is I don't necessarily think of Spirit as a particularly beloved airline.

说实话，如果我要救一家航空公司，Spirit 并不会是我最先想到的那一个。

It's not like if I was trying to save an airline, Spirit is not necessarily the first on the top of my list.

你觉得呢？

What would you say?

哪一家航空公司还有救？

Which one is savable?

哪一家才配得上叫‘人民的航空公司’，安东尼？

Which one is the people's airline, Anthony?

其实吧，如果能回到过去，我年轻那会儿，二十多岁到三十出头的时候，最钟意的是维珍美国航空。

I mean, I I think, actually, if if I could go back in time, I mean, I my favorite when I was, like, in my twenties and early thirties was Virgin America.

要是真有台时光机，再给我几十亿美金，我肯定先救它。

So if, like, if I had a time machine and billions of dollars, that's what I would have saved.

但我觉得，这背后反映的可能恰恰是——虽然我没捐款，但我感觉大家的热情，与其说是有多爱 Spirit，不如说是对如今航空业的整体不满，想说一句：

But I think part of what it speaks to is maybe again, I didn't pledge, but my sense is that the excitement about this is less about how much people love Spirit and more just like a general dissatisfaction with where air travel is nowadays and sort of a desire to say, hey.

你看，机票价格一直在涨。

Like, you know, prices are going up.

体验太差了。

Experiences are bad.

感觉处处都被额外收费，简直被抠得精光。

It really feels like you're getting nickel and dimed a lot.

要是能做个更好、但价格还亲民的选择，该多好。

What if we could do something better, but also that was still affordable?

就这么简单。

That's it.

让我反过来问问你们俩。

Let me let me turn this back around on on either of you.

有谁想猜一猜吗？

Does anyone wanna guess?

你觉得这次能成功挽救精神航空的概率有多大？

Like, what what do you think is, like, the percentage likelihood that this will succeed in actually reviving Spirit?

零。

Zero.

但这种事情只能发生在精神航空身上。

But but it could this could only happen with Spirit.

我的意思是，只有在一家航空公司让乘客们因为糟糕的体验而团结起来时，才可能发生这种情况。

I mean, this can only happen with an airline that, you know, people can bond together because of their miserable experience.

如果你说我们要把捷蓝航空翻新一下，大家根本不会在意。

Like, if if you were to say, oh, we're gonna take, like, JetBlue and turn it around and turn it into a big no one cares.

因为大家对捷蓝航空的体验，可能不好，

Because everybody's experience with JetBlue, like, it might be bad.

也可能还行，

It might be kinda good.

无所谓吧。

Whatever.

非得是全国公认最差的航空公司，影响了那么多人，才能激发集体情绪，让大家觉得：咱们能做得比达美航空更好。

It it it takes the the sort of well known worst airline in the country that has affected so many lives for for the collective psychosis to happen where people say, you know, we can do this better than Delta.

对。

Yeah.

正是这种共同的糟糕经历，才让人们对彼此产生共鸣，真心祝愿你一切顺利。

It is the shared experience of misery that I think will bring people together to pledge, So best of luck to you.

咱们继续下面的内容吧。

We should get on with the rest of the show.

今天还有很多话题要聊，比如 xAI 和 Anthropic 的合作，还有 Haun Ventures 最新一期的基金。

There is a lot to talk about, from the xAI and Anthropic deal as well as Haun Ventures latest fund.

接下来我们会深入探讨企业级 AI 意味着什么，跟进国防部的动态，再聊聊 SpaceX、xAI 和 Anthropic 的其他动向。

Then we're gonna get into enterprise AI and what that means, check-in with the DOD, and, you know, explore other things in the realm of SpaceX, x xAI, Anthropic.

今天的内容可太丰富了。

We've got it all today.

我想先提一个我们 AI 编辑提的问题：xAI 现在算不算一个新型云平台？

I wanna start with a question that our AI editor posed, which is, is xAI a neo cloud now?

这个标题真棒。

Which was a great headline.

我特别喜欢这个说法。

I loved it.

但这其实是在分析 xAI 和 Anthropic 之间的这笔交易。

But it actually is sort of an analysis on this deal between xAI and Anthropic.

是啊。

Yeah.

这周的这个意外挺有意思的，对吧？

This was a fun surprise this week, wasn't it?

我特别喜欢这种意外，尤其是当大家的注意力都放在别的大事上，比如一场重大的庭审。

I always love a surprise, especially when everybody's eyes on, you know, another ball like a a major trial that's happening.

没错。

Yeah.

这周突然之间，SpaceX 及其 AI 子公司 xAI——虽然现在看来 xAI 似乎已经不存在了，或者马上就要被取消了，这个我们待会儿再说——和 Anthropic 达成了一项合作。

Out of seemingly out of nowhere this week, SpaceX and and therefore, it's AI subsidiary xAI, which apparently no longer exists now or is imminently not about to exist, which we can get to, Start a deal with Anthropic.

目前这笔交易的实质是，Anthropic 将全面接管位于田纳西州孟菲斯的 Colossus 1 数据中心的所有算力，重点支持他们面向企业客户的 AI 产品。之前已经有很多报道提到，他们一直在寻找更多算力，尤其是我们过去几周讨论过的那些需求。

And basically, what the the real version of the deal at the moment is that Anthropic is essentially taking over all of the compute at the data center known as Colossus 1 in Memphis, Tennessee to especially focus on Anthropic's more sort of enterprise focused AI products, which, you know, there's been a lot of reporting about how they've been looking for more compute, especially for those they've we've seen and we've talked about in past weeks.

他们之前一直在边缘上做些调整，比如看看现有模型能怎么优化、成本多高，以及什么时候会遇到算力瓶颈。

They've been tinkering a little at the edges of, you know, what's available in those models, how much it costs, when you run into a wall, as far as how much you can use it.

所以这笔交易对他们来说，像是个临时的出路，能让他们短期内获得这么多算力。

And so this seems like an escape valve for them to be able to strike this deal and get access to all this compute in the near term.

你知道，他们现在确实像个新云服务商了，毕竟得想办法处理这些正在建设的算力资源——毕竟很明显，他们根本用不上这些算力来跑Grok，而Grok除了在X平台上，根本没在外界掀起什么波澜，离成为下一个爆款聊天机器人还差得远。

You know, yes, they are a Neo Cloud now in the sense that they kind of had to do something with all this compute that they were building because it certainly seems like they were not gonna need it for Grok, which, you know, outside of x is not burning up the the world, you know, as far as, like, becoming the the new hot consumer chatbot.

没错。

Yeah.

我们得解释一下什么是‘新云服务商’，对不了解的人来说，就是买下英伟达的GPU，然后把这些算力租出去，而不是用来训练自己的AI模型。

And we should say that, like, in terms of what a Neo Cloud is, for those who don't know, this is the idea of buying GPUs from Nvidia and the like and renting those out as opposed to using those for their own AI, training their own AI models.

这完全是另一种商业模式。

So this is a different kind of business.

我们AI编辑拉塞尔·布兰登提到，很多公司都在建数据中心，但如果让他们选，是租出去还是自用——像谷歌这种做云业务的公司，他们还是更倾向于把算力留给自己训练AI模型。

And, you know, the point that our AI editor, Russell Brandon, makes is that a lot of companies are building out data centers, but they are if given a choice between do they rent them out, you you know, because they're doing cloud businesses and stuff like this is I'm talking about, like, Google, they are still prioritizing using this compute for their own internal AI model training.

所以我觉得这点很重要，也暗示了xAI可能在AI模型训练上并没有投入太多。

So I think that's an important point and and one that kind of suggests that maybe xAI isn't doing so much on the the AI model training.

尤其是当我们谈到企业级AI时——我知道后面这期节目会深入聊这个——你几乎听不到有人用Grok来做关键的业务任务。

Especially if we start talking about enterprise AI, which I know we're gonna be getting into later this episode, I don't know you don't hear a lot about people using Grock for, like, work critical tasks.

那么问题来了，xAI 究竟怎么赚钱呢？

And so the question becomes how can xAI actually make money?

看来，单纯出售基础设施可能是主要途径之一。

And apparently, you know, just selling the infrastructure could be one of the main ways to do it.

是的。

Yeah.

你也可以从积极的角度来看待这件事。

And you could take a positive view on that.

对吧？

Right?

嘿，他们总算找到了赚钱的方法。

Like, oh, they figured out a way to make money.

当然，这么做也没错。

And and and, sure, you could do that.

但如果你把自己的公司——比如 SpaceX 和 xAI——定位成一家前瞻创新的企业，那如果只是把显卡租出去，而不用来推动真正的技术创新，这种形象就很难让人信服了。

But I think that when you are positioning your company, in this case, SpaceX slash xAI, as a forward looking innovative company, that's tougher to sell if you are simply just renting out your GPUs and not using them for that innovation.

但既然可以 cynical，干嘛还要积极呢？

But why be positive when you can be cynical?

换句话说，这明显是在 IPO 前搞的一次热度测试，准备把 SpaceX 一股脑儿塞进市场。安东尼，你提到 Grok 根本没被用来做企业级任务，还有报道说 xAI 的员工连内部都没用它，反而在用别的模型。

Which is to say that, like, this seems like a major heat check before the IPO that we're about to see, you know, sort of get rammed into the markets with SpaceX, which is, you know, Anthony, you mentioned not only is Grok not being used for, you know, big enterprise tasks, it also there's been reporting that xAI employees were using other models, like, weren't even using it internally.

这导致了 xAI 被 SpaceX 收购后的一场大地震，除了马斯克，几乎所有联合创始人都离开了。

And that caused this sort of big shake up inside of xAI post acquisition from SpaceX that involved a bunch of the essentially, the co founders leaving other than Elon Musk.

他居然说要从零开始重建 xAI，可 SpaceX 为了这个项目在 IPO 前已经砸了两千五百亿美元。

Him basically saying he's starting from scratch on xAI despite the fact that SpaceX paid $250,000,000,000 for it in the run up to this mega IPO.

现在他又说要把 xAI 从 SpaceX 里彻底拆掉，干脆改叫 SpaceX AI——这人就爱毁掉那些本来有品牌价值的东西，看看 Twitter 就知道了。

And now he's saying that they're gonna dissolve xAI as a separate entity inside SpaceX altogether, and he's starting to call the whole thing SpaceX AI because the this man loves nothing but to ruin a brand that has some value to it, see Twitter.

从短期来看，这种商业模式可能更靠谱，某种程度上，IPO 时投资者或许会觉得它更吸引人，毕竟比吹自己是前沿实验室更实在。

And, you know, this may be a more believable business in the near term, and so on some level, I could see this being maybe more attractive to investors come IPO time because it's like a bit more reliable and certainly more real than them being a leading, you know, frontier lab developer.

但这种模式也吸引不了像前沿实验室那样疯狂的外部投资，至少在正常市场环境下是这样。

But it's also not the kind of business that's gonna draw the same kind of, like, I mean, at least in in a normal environment, outside investment that we're seeing go into all the Frontier Labs.

这大概就是 IPO 过程中出现的最明显的矛盾点了，我特别想知道下个月市场会怎么定价这件事。

And so that's just that's maybe one of the biggest tension points we've seen develop during this IPO process, and I'm really interested to see how that gets sort of priced into the the action come next month.

是的。

Yeah.

我觉得这一点特别值得强调，我们以前也说过，埃隆·马斯克在公开市场——估计在私募市场也是——的成功，很大程度上在于他描绘了一幅未来可能是什么样子的愿景。

I think it's really worth underlining, which we've talked about before, the extent to which Elon Musk, his his, you know, success on the public markets, and presumably on the private ones too, is about sort of presenting this vision of what the future could be.

别太纠结于眼下的业务，而是要想想它最终能变成什么样。

And, like, don't pay too much attention to the current business, but, like, to think about what this could become eventually.

但确实，当你只是在卖基础设施，而不是做前沿实验室的时候，这种操作就难多了，就像你刚才说的。

And it and it does seem like that's harder to do when you're when you're sort of just selling the infrastructure rather being a frontier lab, like you said.

不过话说回来，像英伟达这样的公司不也做得挺好的吗？

Although that said, I mean, you know, like, you know, companies like Nvidia are doing just fine.

所以也许这样就已经足够了。

So so maybe that that's totally enough.

我还想最后再提一句，SpaceX AI，这名字听着真别扭，也不知道以后到底会用到什么程度。

I do also just wanna come back one last time to say that, yes, SpaceX AI, like, what a name it's not clear how widely that's gonna be used.

到目前为止，这名字就只在他发的一条X动态里提过。

I mean, so far, this is basically mentioned in one of his posts on x.

而且我觉得，希望我们的听众都能明白，对这种命名的普及程度得打个问号。

And, you know, I think, hopefully, all of our listeners know to take that with some grain of salt of, like, how widely is this gonna be implemented?

把 xAI 从一家独立公司拆出去，这到底意味着什么？

What does it even mean to dissolve xAI as a separate company?

不过他们倒是真想把这个名字当正经公司名来用。

But, yeah, they try to make that a real, like, company name.

这名字真不怎么样。

It's bad.

我年纪大了，安东尼，还记得当初刚报道特斯拉时，他们因为要把‘电机’从公司名里删掉闹出的风波。

Well, I'm so old, Anthony, that I remember when I was first covering Tesla, all of the kerfuffle around dropping motors from Tesla's name.

所以我真没法想象这种转型会有多顺利。

So I can't imagine the transition being very easy.

当然，我提到特斯拉，是因为我们有些人觉得，迟早有一天，所有这些业务都会合并成一家公司，比如什么‘Space xAI Tesla’，一个超级企业，最后可能干脆就叫‘X’——反正都这样了，为什么不呢？

And, of course, you know, I mentioned Tesla because some of us out here believe that eventually it will all become like one company, and it'll be, you know, Space xAI Tesla, some kind of version of that, some super company in which it will be completely renamed, probably x because why not?

马斯克早就想用‘X’做公司名了，毕竟那原本是 PayPal 的初名。

I mean, Musk has long wanted to have a company called x after that was the original name for PayPal.

他早就注册了x.com这个域名。

He had the domain x.com for a long time.

现在他搞了个社交媒体公司，就叫x。

Now he has a social media company named x.

我个人觉得，最后可能会叫K2。

I I personally, I maintain that it's probably gonna be called K two.

那是他们在某个州注册的实体之一，用来整合这些公司。

That was one of the registered entities they've they've created in one of the states to roll up some of these companies.

他对卡尔达舍夫等级着迷得不行。

He's obsessed with the Kardashev scale.

说到这点，安东尼，咱们是不是该歇歇，别再提SpaceX AI了？

And, you know, to your point, Anthony, and then maybe we can stop talking about SpaceX AI.

这周的这笔交易里，有一部分跟未来规划有关——Anthropic不只是从xAI租用GPU。

There is an element of this deal this week that involves that future looking thing, which is that Anthropic is not just renting the GPUs from xAI.

他们还跟xAI和SpaceX提过一个想法：等几年后SpaceX真把轨道数据中心搞起来，能不能用那些在轨计算资源。

They have also sort of approached xAI and SpaceX about the idea of using this presumed orbital compute that will exist in a few years when SpaceX is supposedly launching orbital data centers.

所以呢，他们打算抛出点东西，让大伙儿保持兴趣，至少在马斯克看来是这样。

So, like, there's the thing that they're gonna dangle to keep people, you know, interested in where things will be going, at least in Elon Musk's view.

马斯克以前说过Anthropic一大堆难听的话，可真到了关键时刻，他还是低三下四地去找人家，好像在X上说过，他的警报没响，觉得能跟这些人讲道理，毕竟人家要给他一大笔钱。

Musk has said so many terrible things about Anthropic, And yet, like, when push comes to shove, he comes hat in hand to Anthropic and, you know, basically says I think he said on x, like, he his evil detector didn't go off, and so he was able to sort of reason with these people because they're gonna hand him a bunch of money.

所以。

So

说到马斯克老是贬低公司，我想顺便提一下，现在正好有个案子在审理，从中爆出的料真是太多了。

Well, speaking of Elon saying bad things about companies, I want to just briefly touch on the fact that there's a trial going on right now, and this is there's been so much interesting fodder that has come from it.

我们已经做了大量报道。

We have had a lot of coverage.

我们有个记者叫蒂姆·富恩霍尔茨，就在法庭现场，这特别有帮助。

We have a reporter, Tim Furnholz, who's who's in the courtroom, which has been great.

但有一件事我觉得得提，尤其是在眼下这些交易的背景下——我们越来越清楚，公众场合，尤其是X（以前叫Twitter）上说的话，跟幕后实际发生的事，往往对不上号。

But one of the things that has come out, you know, and I think it's important to mention it against the backdrop of all these deals because what we are learning more and more, and many of us have known for years, which is what is said in the public sector, particularly on x, formerly Twitter, and what is actually going on behind the scenes don't necessarily align.

关于OpenAI从非营利机构转为营利性公司的权力斗争，也爆出不少惊人细节。

And there's been some incredible details about what was happening in this power struggle of moving OpenAI from a nonprofit to a for profit.

我们还有一篇文章，会在节目笔记里附上链接，讲的是OpenAI联合创始人的一本私人日记，现在他肯定后悔了——这本日记居然在庭审中被公开，里面详细描述了马斯克有多生气。

And we have a story which we'll link to in the show notes that talks about one of OpenAI's, cofounder's personal diary, which I'm sure he deeply regrets now being, you know, right here in this trial, really talking about Musk getting quite angry.

有一段特别生动的描述，说有人送了他一件特斯拉主题的艺术品，结果被人拿走了，他当场发火，摔门就走。

There's a whole scene where he you know, this gift he had been given, which was a a a Tesla, like, art you know, framed piece of art being taken, and he storms out the room.

这简直该给所有可能想跟马斯克做生意的公司提个醒：事情真可能闹到这一步。

I think it should be, like, a warning shot to any other company that maybe is gonna do business with, you know, Elon Musk that it could come to this.

对吧？

Right?

是的。

Yeah.

这场庭审有太多值得聊的内容了。

Feel like there's so much to talk about with this trial.

我真后悔上周没在场，没能跟你们一起跟进第一周的证词。

And I really regret not being here last week to show over the first week of testimony with you guys.

不过我觉得我们还会继续深入讨论这件事。

But I think we're gonna talk about it more.

当然，我们接下来还会聊很多关于人工智能的事，但咱们先暂时跳开一下，聊聊加密货币吧——说实话，2026年我还真没想到自己会再谈这个话题。

Obviously, we're gonna talk a lot more about AI, but let's take, like, a a little little tiny temporary break from talking about AI to talk about crypto, which I am very surprised to be talking about in 2026.

接下来要聊的是凯蒂·豪恩，我记得她以前在安德森·霍洛维茨主导加密货币投资，后来自己单干，刚宣布拿到了十亿美元的新基金，专门投加密领域，这事儿真让我有点意外。

But our our next deal is Katie Haun, who, as as I recall, basically was leading crypto investments that are or involved in in crypto investments at Andreessen Horowitz, kind of struck out on her own and just announced raising $1,000,000,000 in for new investments in crypto, which seems pretty wild to me.

但这可不只是加密货币那么简单。

Well, it's not just crypto, though.

她投的还包括区块链和自主型人工智能。

It's crypto, blockchain, and agentic AI.

也就是说，她瞄准的是一些当下特别火、前沿感十足的技术方向。

So, you know, there are some there are some more current, you know, pieces of technology that are quite trendy that she's gonna be investing in.

不过说实话，看到加密货币又上新闻，我也挺惊讶的。

But, yeah, I was actually surprised to see crypto back in the news.

这到底是历史在重演，还是说这次因为有了自主型AI的加持，情况真不一样了？

And, you know, maybe is it just life repeating itself, or is there something new here with maybe an agentic AI edge to it?

也许吧。

I mean, maybe.

这事儿吧，我觉得真没法说得更肯定了。

Like, that seems like I don't know that you can say anything more definitively than that.

就在她这周完成融资的同时，安德森·霍洛维茨又推出了一支新的加密基金，规模高达22亿美元。

As she raised this money this week, Andreessen Horowitz also raised another fund for crypto involving $2,200,000,000.

这周突然有这么多资金涌向这个领域，过去几年可从来没见过了。

So all of a sudden, there's a bunch of money this week being directed at this stuff in a way that we haven't really seen over the last few years.

克里斯·迪克森和安德森·霍洛维茨的理由是，这属于周期性波动——之前经历了一波低谷，但现在联邦层面的监管政策已经到位了。

And the the justification for Chris Dixon and Andreessen Horowitz is that this is cyclical, and so now we're in, you know, we saw a bit of a bust, but we've got the regulation, you know, the policy in place at the federal level.

说白了，现在的政府对加密货币是真有兴趣，不管这兴趣具体意味着什么。

We clearly have boy, let's put it nicely, an administration that is interested in crypto in whatever way that means.

所以他们觉得，未来几年投进去的钱，有很大机会获得丰厚回报。

And so they feel like they're set up to be able to, you know, invest in things over the next couple years in a way that will that will make them big returns.

我很好奇，到底有多大程度上， startups 和风投真的认为 AI 也能推动加密货币的发展。

I am very curious to what extent I mean, I I have seen, you you know, startups and VCs talking about AI also being a big opportunity for crypto.

但我没觉得这些说法有多 convincing，我觉得更可能的情况是，大部分钱还是投给 AI，加密只是顺带沾点光。

I haven't found any of those arguments to be particularly compelling, so I could see a scenario where essentially just a it just ends up being a lot of investments in AI and then a little bit of crypto as well.

我想我们只能拭目以待了。

I I guess we'll see.

再回到肖恩关于环境的观点，没错。

And and I think just to go back to Sean's point about the environment, yes.

很明显，很多坚定的加密货币支持者都会说，你看。

Like, obviously, there's I think a lot of crypto true believers are just like, look.

听好了。

Listen.

有些事情本来就是有周期的。

Some of this stuff is cyclical.

加密货币真正的潜力还没完全展现出来。

You know, the true arrival of crypto's potential is is still to come.

但确实，监管环境也已经发生了很大变化。

But it it's also yes.

很明显，特朗普政府已经采取了很多措施，不再对加密公司施加压力了。

Like, the regulatory environment has just changed a lot, and it's clear that, like, you know, the Trump administration has taken a lot of steps to back off from anything that might make crypto companies uncomfortable.

所以，确实现在经营这类公司，比几年前要好多了。

So, yeah, it could seem like this is a a good time to be running one of those companies as opposed to a few years ago.

对。

Yeah.

我不觉得这是周期性的，更多是抓住了当前政治环境带来的机会。

I mean, I wouldn't say cyclical as much as opportunistic based on what is happening politically.

不过在我们开始录之前，我正好想到一件事：既然现在监管和政策层面都这么有利，crypto的表现本该比现在好得多才对。

But here's one thing that I was kind of thinking about a little bit right before we, you know, started recording today, which was, you know, you would think actually with all of the tailwinds that they have, at least on a, like, a regulatory policy, like, political level, that crypto would be doing much better than it is right now.

也许再过六个月到一年，我们就能看到这种宽松监管环境带来的成果。

And and maybe six months to a year from now, we will see sort of all the fruits from from this very, like, regulatory free environment.

但我现在没看到它像我预期的那样发展起来。

But I don't see it the way I thought it would have translated by now.

我得承认，参与crypto的人确实挺谨慎的，而且有道理——因为现在的政府对crypto的态度反复无常，好像只是为了照顾那些掌权的人。

I'll I'll give some credit to people who are involved in crypto and say that, like, I think there's some actual caution being exercised here, and for good reason, which is that you have so much there's just so much choppiness in this area from the administration that we have right now, sort of getting involved in crypto seemingly for the benefit of the people who are running the the government.

另外还有像孙宇晨这样的人，他以前跟特朗普政府走得挺近，参与了不少crypto项目，但现在却公开抱怨特朗普政府和特朗普家人对他态度恶劣，而对方还反过来针对他的儿子。

And you also have other stuff, like, you have Justin Sun, you know, who was getting involved with the Trump administration and some of the crypto stuff that they're doing, but he's now, you know, sort of speaking out about how badly he thinks that the Trump administration and the Trump family is treating him, and they're coming back at at Justin's son.

还有这么多基于加密货币的营销活动、Kalshi的操作，以及内幕交易这些事。

And then, you know, that you've got kind of all of this crypto based poly marketing, Kalshi activity, and and insider trading stuff.

有太多类似这样的情况，这让我明白，为什么政府虽然对加密货币开了绿灯，比特币价格却没涨到二十万美元之类的样子。

There's just a lot of stuff that's happening like that that I feel like helps me understand why the administration giving its blessing to crypto hasn't set the price of Bitcoin to, like, you know, $200,000 or something like that.

现在加密世界里真是风波不断。

Well, lots of intrigue right now happening in the crypto world.

在深入探讨一些更深层的主题之前，我想提一下一笔刚好赶上的交易，因为我们最近采访了这家公司的创始人之一。

I wanna mention before we get into some of the deeper themes, a deal that is timely base basically, because we've recently, interviewed one of the founders.

我说的是Aurora Innovation这家公司。

And I'm talking about Aurora Innovation.

这家公司专注于自动驾驶技术，目前主要应用在无人驾驶卡车上。

This is the company that's working on autonomous vehicle technology, specifically right now applying it to self driving trucks.

他们最近达成了一笔交易，我这周刚写过，而且这是一份实实在在的商业合同。

And they had a deal which I wrote about this week, and it's specifically a commercial contract.

这一点很重要，因为很多这类公司虽然有不少试点项目，但真正落地的商业合作却不多。

And I think this is important to note because so many of these companies have a lot of pilot programs.

当这些技术真正转化为商业合同时，就特别值得关注了。

It's notable when these translate to commercial contracts.

奥罗拉是一家上市公司，但说实话，我依然觉得它更像一家初创公司。

And Aurora is a publicly traded company, but I still think of them, you know, honestly, as a start up.

他们和一家叫麦克兰的公司签了商业合同。

They they have a commercial contract with a company called McLane.

麦克兰是伯克希尔·哈撒韦的子公司，可以说是全美最大的分销商之一。

This is a Berkshire Hathaway subsidiary that is basically one of the, like, largest distributors in the country.

他们主要做餐饮品牌的配送，给快餐连锁店供应易腐食品之类的商品。

They really focus on, like, restaurant brands delivering sort of perishable foods and things to, like, fast food chains.

如果他们能进一步扩大这项合作，确实值得密切关注。

So this is gonna be interesting to watch if they expand upon that.

在我们的节目笔记里，有一段我们最近对联合创始人兼CEO克里斯·厄姆森的采访链接。

And in our show notes, there's a link to an interview that we recently did with the cofounder and CEO, Chris Urmson.

那次采访是在HumanX大会上做的，但到现在依然非常相关。

And this was back during the HumanX conference, but it's absolutely timely.

所以我强烈建议你去听一下。

So I highly recommend that you listen to it.

如果你们对自动驾驶卡车没什么想说的，那咱们就进入企业级AI的话题吧。

And unless you guys have some comment on self driving trucks, I kinda think we should jump into enterprise AI.

好，开始吧。

Let's do it.

行。

Alright.

结果发现，各地都在频频发生交易。

Turns out there's a lot of deals happening everywhere.

人人都想切入企业级AI领域。

Everyone wants to get into enterprise AI.

要是一年前，我可不敢这么说，但现在这事儿确实在发生。

I couldn't have said that maybe a year ago, but it's absolutely happening now.

你们两位能给我讲讲现在到底发生了什么吗？

Can either one of you give me some of the rundown of what's happening?

最大的新闻可能是 OpenAI 和 Anthropic 宣布联手专注企业级人工智能。

Maybe the the biggest headline is is OpenAI and Anthropic announcing a joint venture to focus on enterprise AI.

与此同时，Sierra 也在行动，虽然它没 OpenAI 或 Anthropic 那么出名或规模大，但它的创始人是布雷特·泰勒，很多人应该都熟悉他，毕竟他以前在 Salesforce 很有名。

And that's also happening kind of at the same time as Sierra, which is, you know, not quite as well known as or large as companies, but is is led by Brett Taylor, you know, who who a lot people may know from Salesforce.

他们融了九亿五千万美元。

But they raised $950,000,000.

跟 Anthropic 或 OpenAI 的融资比起来，这数字不算特别大。

Again, compared to some of the Anthropic or OpenAI rounds, not huge.

但这也说明了，哪怕你不是顶尖的前沿 AI 公司，投资人依然愿意投你。

But again, a sign of even if you were not one of those, you know, top, top, top tier frontier AI companies, that investors are still excited to be backing you.

这又回到我们这周讨论的一个主题：人工智能里真正的钱到底在哪？

And again, just kind of coming back to that theme of like And think it's sort of, in some sense, a theme running through a lot of our discussion this week of sort of like, where is the actual money in AI?

大家普遍觉得，企业应用场景，尤其是编程相关领域，可能会是第一个金矿。

And it seems like there's a general sense that, okay, enterprise use cases, coding in particular, that's that's really gonna be maybe the first gold mine.

还有，如果你是个初创公司，被大公司收购了，那也是赚钱的好路子。

And you know where there's also money is if you're a startup and you get acquired.

那里也有钱可赚。

There's money there too.

还有另一笔交易，SAP花了大约十亿欧元收购了一家叫Prior Labs的AI初创公司。

There's another deal that happened, which is SAP spending, I think it was about a billion euros on an AI startup called Prior Labs.

所以，如果你的初创公司做的是企业级工具，那毫无疑问会成为被收购的目标。

So that's another space where if you are a startup that has enterprise tools, you are absolutely an acquisition target.

这对大公司来说未必是直接的收入来源，但充分说明了他们的兴趣所在。

And so that's not a revenue generator necessarily for the big companies, but it it shows the the interest.

我觉得也别忘了，我们预期OpenAI和Anthropic今年也会各自上市。

I think it's also good to remember the context of the fact that, like, we are expecting these companies, OpenAI and Anthropic, to also do their own IPOs this year.

在我看来，这更像是为可能的上市做准备，努力打造可持续的业务。

This is to me, this feels more buttoning up ahead of those possible public listings and trying to come up with durable businesses.

你退一步想想，过去几年这些公司获得的投资，全都是冲着爆炸式增长去的，我们常听到的一个词就是年经常性收入。

I mean, it's like, you take a step back and we think about the investment that's gone into these companies over the last couple years, it's all been about this meteoric growth And the the phrase we hear all the time is annual recurring revenue.

你知道谁其实根本不关心年经常性收入吗？

You know who doesn't really care about ARR?

这就像是公开市场的投资者。

It's like public markets investors.

ARR 这个指标，在你还是私营公司、财务数据不透明的时候，特别有用。

Like ARR is a thing that you can lean on when it's obscured by the fact that you're private.

但一旦上市，它就没那么重要了，因为你的财务数据全都公开了，大家能清清楚楚看到你到底赚了多少钱，哪些收入是靠得住的。

But when you go public, it's not really something that matters anymore because your numbers are out there and people can see what you're making, what you're actually making, and what you can bank on.

所以在我看来，这进一步说明这些公司已经被说服了，必须打造真正稳定可靠的收入来源——当然，很多人想买这些股票，是因为看好公司本身、讲的故事、造的神话，但也有大量指数基金想买，因为他们看中的是收入增长的确定性，愿意据此下注。

And so to me, this is more evidence that these companies have been had some sense talked into them that they need to come up with real reliable revenue that, you know, yes, of course, there's going to be a large component of people wanting to own these stocks because of, you know, the companies and and the the story and the myth making and everything, but there are also a lot of index funds and stuff out there that wanna own it because they're gonna see revenue rise and they're gonna be able to bet on it.

而且华尔街分析师还会告诉他们：没错，未来几年内，这项收入肯定会稳步增长。

And they're gonna get, you know, Wall Street Analysts telling them that, yes, you can rely on this revenue to grow in, you know, x amount over the next couple years.

让我特别明显感觉到、也很有意思的是，很多 SaaS 公司之所以纷纷推出企业级工具，尤其是智能代理类 AI 工具，是因为担心 AI 会侵蚀自己的主营业务。

What is really apparent to me and it's very interesting is that, like, so many SaaS companies got into launching enterprise tools, specifically agentic AI tools, because of concerns about the cannibalization of their business due to AI.

所以你看到这种爆发式的增长。

And so you see this massive rise.

我们在 Strictly VC 活动上，曾经采访过 Uber 的首席技术官。

I mean, we had an interview with the Uber CTO at our Strictly VC event.

我后来还给他打过电话，他说优步在AI方面，特别是智能代理AI上的投入，把IT预算花得差不多了——不是整个预算，只是IT这块。

And I also spoke to him on the phone, and he talked about how Uber, like, blew through their budget their IT budget, not their whole budget, because they have been really encouraging people to use AI, specifically, agentic AI.

他说，过去六个月里，这简直是一场革命。

And he said that that's a revolution that's happened over the last six months.

很明显，这里确实有实际应用，而且在优步这样的公司里，甚至开始放缓招聘速度了。

So there's clearly use happening there, and and it's causing actually some companies to slow in in Uber's case, actually slow the pace of hiring.

目前还没裁员，但为了抵消这笔开销，招聘节奏明显慢下来了。

So no layoffs yet, but slow the pace of hiring to offset the cost of this.

总得有人在这中间赚钱吧。

So someone's gotta be making money somewhere.

对吧？

Right?

优步确实在用Cursor，这我知道。

Like, Uber is using Cursor.

但问题是，到底谁在赚这笔钱呢？

But, you know, who who is making money here?

在企业级AI领域，尤其是代理式AI方面，有没有哪家公司脱颖而出？

And is there a winner on the enterprise AI side of things, particularly with AgenTic AI?

我注意到Uber的那些说法里，有一点挺引人注意的，他们说目前大约有10%的代码是某种形式由AI生成的。

I think one thing that was kinda caught my eye in the in the Uber comments was I think they were basically saying that about 10% of their code at this point is basically generated by AI in some form.

这个数字挺有意思的，因为考虑到Uber的规模和它运营的时间这么久，你就会觉得，这确实意义重大。

And that is an interesting number because it is both like, I mean, you think about Uber's scale and and how long it's been operating, you're like, wow, that's, like, both really significant.

同时，当你想到人们常说的‘AI编程’已经多么主流，几乎要取代传统编程方式时，就更觉得值得玩味。

And also when you think about people talking about how dominant vibe coding has become, how it's essentially replacing sort of traditional programming.

也就是说，这个比例还没达到我原本以为的水平，这其实反映了现在关于AI的很多讨论：一方面，增长确实非常显著，另一方面，炒作又太夸张，有点超前于实际进展。

Like, oh, well, that's still not as big as I thought it would be, which I think speaks to a lot of discussion about AI right now, is that both there is, like, really, really significant growth and also the hype is so incredible that it's slightly outpacing it.

这当然没错。

Well, sure.

不过我想稍微纠正一下，Uber其实直到去年年底才开始大量使用代理式AI。

But I would push back a little bit on the fact that, like, in Uber's case, they didn't seem to be using AgenTic AI much until late last year.

那么一年后，我们会处在什么位置？

So where are we gonna be a year from now?

我觉得这会非常有意思。

I think that's gonna be really interesting.

我不确定你们俩怎么看，我特别想听听你们的意见，因为《Semaphore》上有一篇里德·阿尔伯戈蒂写的文章。

I don't know if either of you I'm really curious about your opinions about this, which is there was an opinion piece in semaphore by, Reed Albergotti.

他 basically 说的，其实又回到了我之前那个问题：到底谁在靠这个赚钱。

And basically, he said that and it goes back to my point of who's making money on on this.

他的观点是，职场早就已经失去对员工使用哪些AI工具的控制了。

And his point is kinda like, the workplace has already completely lost control over dictating what AI tools people make.

换句话说，可能根本不会出现什么真正的赢家。

Therefore, you know, you could translate that to there isn't gonna be necessarily a winner.

他拿人们用手机和相关工具的情况作比，说企业早就放手了。

And he equated it to how people would use their phone and, like, some of the tools associated with with that and companies kinda giving up.

说白了，员工就是用自己觉得顺手的工具，然后通过大型合同之类的途径，把这种使用习惯带进公司，事情已经没法回头了。

So, basically, it's like the employees are just using the tools that they like best and, like, trying to dictate that within a corporate setting through these, like, very large contracts or, like, that it's over.

这事儿早就定下来了。

It's like it's already done.

我想知道你是否同意这个观点？

And I'm wondering if you agree with that at all.

作为现在在工作流程中用微软Word的人，员工有选择权确实不错。

I mean, as as someone who now uses Microsoft Word in his in his workflow, yeah, employee choice is great.

对。

Yeah.

我觉得这里面存在很多不同的压力。

I think that the there are, like, a lot of different pressures.

而且我觉得，这其实这些年已经被各种流行说法说过了。

And I think that it's certainly true that, I mean, think this has been went by a lot of different catchphrases over the years.

其中一个说法是IT的消费化，对吧？

I think one was consumerization of IT, right?

意思是你完全可以自己用个人设备，反正现在都是云服务，谁还在乎呢？

The idea that you could just use your own devices sort of everything's SaaS anyway, so who cares?

在很多环境下，这种说法确实成立，但在其他环境下就没那么明显了。

And that, I think, was true in a lot of environments and less true in in other environments.

我猜这里也会是同样的情况，因为当你开始谈论设备时，大家担心的隐私和安全问题，其实用在人工智能上也同样存在。

And I suspect that will also be the case here in the sense that I mean, a lot of the concerns that that are things that come up when you start talking about devices, like privacy and security, those are things that are also gonna be an issue when it comes to AI.

所以，如果你把大量公司信息输入到各种不同的AI模型里，你肯定得确保这些信息将来不会变成安全漏洞。

And so, like, hey, if you're, you know, putting a bunch of company information into a variety of different AI models, you wanna be really sure that that is not gonna become a security vulnerability in the future.

说实话，我觉得他既描述了当下的现实——没错，确实如此。

So it it's hard for me I mean, I think that he's both describing a reality now, which is that I think yes.

当然，公司无法完全掌控员工在用哪些AI工具，但这并不意味着就会完全失控。

Absolutely, companies do not have complete control over what AI tools their employees are using, but that doesn't mean it's just gonna be a complete free for all.

说得很好。

Well said.

我们时间不多了，不如顺便问问国防部那边现在是什么情况。

We have tiny bit of time, so I think we should do a little check-in with the Department of Defense and find out what's going on there.

这其实也不是什么新消息，不过我们以前聊过，Anthropic和国防部之间一直有拉锯。

And, you know, this isn't necessarily, you know, new news, but we've talked a little bit about in the past how there's been kind of a tug of war going on with Anthropic in DOD.

在我看来，硅谷的公司几乎都在争抢军方合同，尤其是在我所在的自动驾驶领域。

It seems like, you know, to me, everyone is going after you know, everyone, let's say, in Silicon Valley seems to be a dual use or is going after a defense contract of of some kind, particularly in my world in in autonomous vehicles.

我想知道你们两位观察到什么，是不是也觉得情况确实如此。

And I'm wondering what you two are seeing and if and if that tracks.

我很想知道你们怎么看，因为在我看来，这部分原因在于政府在国防科技上更愿意花钱。

I'd be curious what you guys think on this because to me, part of this seems to be driven by the fact that we're dealing with a government that is more willing to spend on new technology when it comes to defense.

他们把现在称为‘战争部’的这个部门推到了最前沿，这种做法是很久没见过了，我觉得这进一步迫使他们必须显得在技术应用上非常前沿。

They've put this department, which they now call the Department of War, like front and center in a way that we haven't seen an administration do in a long time, which I think even further puts the pressure on them to look like they're, you know, being very forward thinking with the technologies that they use.

而且这个部门的预算本来就远高于其他联邦机构，未来可能还会继续上涨。

And also we have, you know, a budget for this department, which has already always been really high compared to other federal offices, that is only probably going to get higher.

所以对我来说，这是推动这一现象的关键因素。

So like to me, that's a huge driver of this.

如果你在开发的东西对国防部有点用处，为什么不争取一下合作机会呢？

Like if you're building something that could be in some way useful to the defense department, like why wouldn't you try to line up some business?

但我相信，这不可能是唯一的原因。

But I'm sure that can't be the thing that's accounting for all of this.

是的。

Yeah.

我觉得有一件事值得提一下，五角大楼刚和英伟达、微软以及亚马逊云签了协议，要部署非机密的AI网络。

I mean, I think one thing that's worth mentioning is the Pentagon just inked a deal with Nvidia, Microsoft, and AWS to deploy AI unclassified networks.

这些可是大公司，肯定从中获益不少，我们看到的新闻也大多集中在这些巨头身上。

So these are really big companies that are certainly benefiting, and a lot of the news that we see are with these very large companies.

但还有一批规模小一些的初创公司，要么跟大公司合作，要么直接找现有的防务承包商，想分一杯羹，拿下这些利润丰厚的合同。

But then there's a second tier of, you know, smaller startups that either partnering with larger companies or maybe even existing sort of defense contractors to land some of these con some of these possibly lucrative contracts.

它们这么做的目的，某种程度上也是为了缓解资金压力。

And and they're sort of doing it in a way as, like, another another way to shore up capital.

我觉得有意思的是——或者说，我的看法是，美国政府投资初创企业和创新，本身未必是坏事。

What's what's kind of interesting to me is or maybe not in so much interesting, but, like, I guess my opinion here is that I don't necessarily think it's a bad thing that the US government is investing in startups and innovation.

我们在这方面本来就有很长的历史。

We have a long history of that.

只是现在所有资源都集中在一个方向上：国防和武器系统。

It's just that it's singularly focused through the, like, defense and weaponry sort of filter.

过去我们也有大量科学领域的资助，比如DARPA，但那时候并不全是围绕军事用途。

And we have a long history of grants on the science side or DARPA, and that those didn't necessarily have a defense focus.

所以现在的问题是，这种投入实在太集中在这一个方向上了。

So now it just seems like the risk here is that it's so focused on this.

创新的多样性明显不够。

There is not a lot of diversity in terms of innovation.

对我来说，这值得继续关注，但我不确定这是不是件好事。

And to me, that is I guess I'll be watching it, and I don't know if it's necessarily a good thing.

而且，考虑到我们目前这么关注这件事，很大程度上是因为Anthropic那场争议——五角大楼显然对Anthropic试图为AI使用设定底线感到不满，这引出了更多根本性的问题：当AI越来越多地参与决策时，道德责任、决策机制这些到底该怎样界定？

Well, and I think the other risk, of course, given that part of the reason we're paying so much attention to this right now is because of that whole Anthropic dispute and how it seemed like the Pentagon really did not like that Anthropic was trying to draw some red lines around how AI gets used, is that I think there are some big questions about what does, you know, moral responsibility, what is decision making, what do all those things look like once you incorporate more AI into these decisions?

我感觉现在掌权的人根本没打算认真探讨这些问题。

And I do not get the sense that the people in charge right now are interested in exploring those questions in any serious way.

不仅如此，就像我们这档节目之前聊过的，这一切最终都归结到词语和语言的定义上——这些公司可以说得天花乱坠，比如声称他们只同意做‘合法’的事，或者怎么包装都行。

Not only that, but as we've talked about on this show, this all comes down to, like, definitions of words and language where, like, you know, you these companies can say all they want that, like, you know, they're only agreeing to allow something legal to happen or whatever however they wanna, you know, call it.

但政府到底怎么理解这些说法？

But, you know, what is the government's interpretation of that?

他们又会怎么定义这些词？

How is it going to define it?

所以你就算划了红线，也完全可能因此惹上麻烦，就像Anthropic那样。

So you can draw red lines, and certainly that can be strong enough move to get you into some hot water, like, with Anthropic.

但说实话，这些红线的实际分量，没公司们说的那么大。

But also, I I feel like those red lines kind of don't mean as much as the companies would like you to believe.

是的。

Yeah.

当然。

Absolutely.

时间到了，不过下周一《Equity》还会回来。

Well, we are out of time, but Equity will be back next week.

在这之前，你可以关注我们在X和Threads上的账号@Equity Pod。

And until then, you can follow us at Equity Pod on x and threads.

《Equity》由TechCrunch的资深记者主持，制作由特蕾莎·洛孔佐拉负责，剪辑由Cal完成。

Equity is hosted by TechCrunch senior reporters and produced by Teresa Loconzola with editing by Cal.

在YouTube或你常用的播客平台订阅我们，并前往techcrunch.com/events了解最新动态。

Subscribe on YouTube or wherever you get your podcasts and find out what's next at techcrunch.com/events.

非常感谢收听，我们下次再聊。

Thanks so much for listening, and we'll talk to you next time.