Google Plans New ‘Frozen’ AI Chip, Preventing Scams in the AI Era, 160 Software Startups For Sale

欢迎各位收看《The Information》的TITV节目。

Welcome everyone to the Information's TITV.

我是阿卡什·帕什里卡。

My name is Akash Pasricha.

今天是7月20日，星期一。

It is Monday, July 20.

本期节目，谷歌正在研发一款新的服务器芯片，让它的Gemini人工智能模型运行得更高效。

Today on the show, Google is developing a new server chip that would allow its Gemini AI models to run more efficiently.

我们会采访《The Information》那位率先报道这一消息的谷歌记者。

We'll talk to the Information's Google reporter who broke that story.

我们还会看看人工智能是如何影响网络安全行业——无论是好是坏。

We'll also look at how AI is changing the cybersecurity sector for better or for worse.

我们也会对话Savi Security的首席执行官兼联合创始人。

We'll talk to the CEO and cofounder of Savi Security.

另外，上周《The Information》发布了一份清单，列出了超过160家可能在近期内被收购的企业级软件初创公司。

Plus, last week, the information published a list of more than 160 enterprise software startups that could be acquired in the near future.

我们来采访一下整理这份名单的记者。

We'll speak with the reporter who put together that list.

今天这期节目内容很精彩，咱们马上开始。

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

谷歌正在研发一款新的服务器芯片，名叫Frozen芯片，作为其TPU产品的补充。

Google is working on a new server chip called the Frozen chip, complementing its TPU offering.

我请来我们的谷歌记者 Erin Woo，她和亚洲记者刘畅一起独家报道了这个消息。

I want to bring on Erin Woo, our OpenAI Google reporter who broke that story with our Asia correspondent, Channel Liu.

Erin，欢迎来到节目。

Erin, welcome to the show.

很高兴你再次来做客。

It's great to have you back.

嗨，谢谢邀请我。

Hey, thanks for having me.

说说看，谷歌正在研发的这款新芯片是怎么回事？

Tell me about this new chip that Google is working on.

是的。

Yeah.

谷歌正在研发一款新芯片。

So Google is working on a new chip.

公司内部非正式地称它为Frozen V2。

It's informally dubbed Frozen V2 inside the company.

它的设计目标是，刚推出时就能比当时的TPU高效得多。

Essentially, idea is that it could be way more efficient even than the TPUs at the time of launch.

他们预计，它的效率将是当时TPU的六到十倍。

They're projecting it could be six to 10 times more efficient than the TPUs at the time of launch.

这是因为它的运行方式略有不同——直接把模型刻录在芯片上。

That's because it works slightly differently by etching the model itself into the chip.

这样，模型运行时就不需要做那么多决策了。

So fewer decisions have to be made at the time that the model is running.

把模型刻在芯片上。

Etching it into the chip.

那为什么叫‘Frozen’呢？

So why is it called Frozen?

这个名字是因为模型某种程度上被固化在芯片里了。

So the name Frozen has to do with the idea that the model is frozen into the chip to some degree.

一般来说，普通芯片是通用型的，能做很多不同的任务。

So essentially the way that a regular chip works, it's generalizable, which means that it can do a lot of different things.

但这也意味着芯片在运行时要处理大量决策点。

But that means that there are a lot of different decision points that the chip has to run.

这款芯片通过预先设定部分流程，来提升效率。

So this chip makes things more efficient by combining some or sorry, by presetting some of that essentially.

这样一来，就不必每次都做那么多决策了。

And so it doesn't have to make all those decision points.

那它和TPU有什么不同呢？

And so how would it be different than from the TPU?

是的。

Yeah.

TPU 更偏向于那种通用型芯片，而这些冻结芯片则是专门为运行这类模型架构设计的。

So the TPU is more of this generalizable kind of chip versus these Frozen chips, which would specifically run like this model architecture.

我明白了。

I see.

也就是说，如果这个冻结芯片真的只能运行 Gemini 模型，听起来是这样。

I mean, if this Frozen chip I mean, if it's really only I guess it's only meant to run Gemini models, sounds like.

那它是只供谷歌内部使用吗？

So is it only meant to be used internally by Google then?

是这个意思吗？

Is that the idea?

其实很多谷歌云的客户也在用 Gemini 模型，所以理论上外部客户也能用这种芯片。

I mean, so a lot of Google Cloud customers also run Gemini models, and so this is something that could theoretically also be run by external customers.

不过这还处于很早期的阶段。

But this is still something that's pretty early.

最早也要到 2028 年才会推出。

So it's not expected to launch until 2028 at the earliest.

所以我觉得，很多这类战略问题还在逐步明确中。

So I think a lot of these strategy questions might still be in the process of being worked out.

对。

Right.

我问这个问题是因为，我们都知道TPU以前主要供内部使用，但据我了解，现在他们已经开始为客户的数据中心部署这些TPU了。

Because I mean, the reason I'm asking this question is because we of course know that the TPU was mostly used internally, but now they're at the point where they're really trying to outfit customer data centers, if I recall correctly, with these TPUs.

你说得对，不只是谷歌内部在用Gemini，很多客户也在用，那这跟其他公司做推理芯片的相比，差距有多大呢？

I mean, your point is well taken that it's not just people at Google running Gemini, it's all sorts of customers, so that could very How much be the does this compare with other inference folks' chips?

我们之前请过Sama Nova的CEO来做节目。

We've had the CEO of Sama Nova on the show.

OpenAI也在研发自己的芯片。

OpenAI is working on their own chip.

那这个跟他们的相比怎么样？

How would this compare to that?

是的。

Yeah.

很多人都在做同样的事，就是想办法降低推理成本，因为现在大家都面临算力紧张的问题。

So a lot of people are trying to do the same thing, which is essentially driving down the cost of inference because everyone's in this big compute crunch.

谷歌的芯片因为采用固定设计，运作方式有点不一样。

So Google's chip works a little bit differently because of this frozen design.

这其实和一家加拿大芯片初创公司Thales的做法有点相似。

It's actually something somewhat similar to what this Canadian chip startup, Thales, is doing.

他们也在尝试把特定模型直接刻录到芯片里，从而提升效率。

So they're also trying to specifically etch a model into the chip so that it is more efficient.

谷歌之前是不是也受算力限制？

Has Google been compute constrained?

这是不是解决这个问题的一种方式？

Is this a way to get at that problem?

是的。

Yeah.

他们在财报电话会上经常提到，云业务一直受到算力不足的制约。

So they talk about that constantly on earnings calls, the idea that their cloud division is compute constrained.

谷歌这周要公布财报，我猜他们也会提到这个问题。

Google reports earnings this week, I would expect them to also talk about that this week.

跟我说说，台积电的产能，我猜这也是大家对这款新芯片最关心的事吧？

Tell me, TSMC capacity, that I imagine is also top of mind here for this new chip?

是的。

Yeah.

谷歌得为这款芯片找到代工厂的产能。

So Google is going to have to find fabrication space for this chip.

目前还不太清楚是不是一对一的置换关系，也就是说，为了生产这些芯片，腾出台积电用于TPU的产能，这确实有可能。

It's not totally clear to me that it's gonna be like a one to one exchange in the sense that in order to make these chips that seals from TSMC space for the TPUs, that's definitely possible.

但也有不少人告诉我，这个制造工艺差异挺大的，可能会用到不同的产线，所以对谷歌来说，未必是零和博弈。

But it's also been suggested to me that this is enough of a different fabrication process that it might use different space, so it might not be a zero sum game for them.

很好。

Great.

埃琳，非常感谢你今天来做客。

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

这位是 Erin Woo，我们《信息》杂志负责报道 OpenAI 和谷歌的记者。

That is Erin Woo, our OpenAI and Google reporter here at The Information.

人工智能让黑客和诈骗分子更容易欺骗普通人。

AI has made it much easier for hackers and scammers to deceive people.

我们下一位嘉宾写了一本关于这个话题的书。

Our next guest wrote a book on that topic.

Patrick Coughlin 是 Savi Security 的联合创始人兼首席执行官。

Patrick Coughlin is the co founder and CEO of Savi Security.

他撰写了《繁荣的黑暗面》一书。

He is the author of Dark Side of the Boom.

我们想请 Patrick 来聊聊他的研究。

Want to bring on Patrick to talk about his work.

Patrick，欢迎来到节目。

Patrick, welcome to the show.

很高兴你今天能来。

It's great to have you here.

谢谢邀请我，阿卡什。

Thanks for having me, Akash.

我想先聊聊整个网络安全的大环境，然后再具体谈谈你在Savi所做的工作。

So I want to talk a little bit about the broader cybersecurity landscape at large, and then we'll zoom into what you are doing at Savi.

我刚才一直在想这个问题。

I was thinking about this question here.

现在安全措施跟得上AI的发展速度吗？

Is safety keeping up with AI innovation right now?

你怎么看？

What do you think?

是的。

Yeah.

过去二十年里，网络罪犯和国家行为体一直把最复杂的攻击留给政府机构和大型企业，因为他们觉得这些地方能带来最大收益。

Look, for the last twenty years, cybercriminals, nation state actors, they saved their most sophisticated campaigns for exploiting government agencies and large enterprises where they thought they could extract the most value.

但最近几年情况变了，这些老对手，加上新冒出来的组织，正在把AI应用到他们运营的每一个环节，这让攻击成本更低、操作更简单，最终也让他们更有动力把目标转向普通老百姓，用越来越个性化、精致、逼真的骗局来行骗。

I think what's changed in the last few years is these same organizations and new ones that are popping up are deploying AI across every stage of their operations, which is making it cheaper and easier and ultimately more financially feasible for them to move down market and really lever increasingly personalized, polished, and convincing scams at everyday people.

我是说，你看看现在这一堆模型，我问这个问题是因为最近几天Kimi K3一直被热议。

I mean, tell me, when you look at the fleet of models and I ask this question because, of course, Kimi K three has been the topic of conversation for the last couple days.

很多人都说它是个非常出色的模型。

There have been a lot of observations that it's a really good model.

有人在问，它在网络安全和安全防护方面，真能跟Fable比吗？

Some questions around, hey, is it as good for cybersecurity and safety as Fable, for example?

你怎么看？

What's your view on that?

有没有哪些模型在安全方面特别突出，而其他模型就差一些？

Are there certain models that are better for safety than for others?

你看啊。

Look.

网络安全这事儿，就是一场猫鼠游戏。

I mean, cybersecurity is a cat and mouse game.

过去二十年里我们看得清清楚楚：坏人不断出新招，防御方也只能跟着不断升级。

And, you know, we've seen that for the better part of the last twenty years is bad guys innovate and then defenders try to innovate as well.

这就像一场不断追赶的猫鼠游戏。

And it's a cat and mouse game of trying to keep up.

我觉得现在不同的是，基础模型和KeyAI实验室推出的创新速度实在太快了，让人很难搞清楚我们到底处在猫鼠游戏的哪个阶段，谁又占了上风。

I think what's different now is this this relentless pace of innovation coming out of the the foundation model and the KeyAI labs that are creating these new innovations, and it's making hard to really understand where are we in this cat and mouse game and and who's winning.

我们清楚的是，针对普通消费者和家庭的创新投入远远不够，而数据已经清楚地证明了这一点。

What what we do know is that we haven't seen nearly enough investment in innovation deployed at the consumer and deployed at families because the data is really proving it out.

现在，普通美国成年人每天都要遭遇好几起诈骗，过去十二个月里，美国人因为这些AI驱动的骗局损失了近两千亿美元。

I mean, now the average American adult is encountering multiple scams a day, and Americans in this country have lost close to $200,000,000,000 in the last twelve months to these AI powered scams.

所以我们必须转变思路，把创新和讨论的重点，不只是放在企业或政府机构上，更要思考如何让这些技术让普通人的网络安全变得更简单、更易用。

And so we've got to shift our innovation and our effort and our conversation about the power of all these models and this technology, not just to include what we're doing for enterprises, what we're doing for government agencies, But how are we using that technology to make sophisticated cybersecurity a little bit more accessible for everyday people?

那么回到刚才的问题，模型之间真的有区别吗？

And so going back to it then, I mean, is there a difference then between the models?

OpenAI的模型在安全性上比Anthropic的更好，还是反过来？

Are OpenAI's models better than Anthropix for security or vice versa?

你怎么看？

How do you see that?

关键在于你怎么使用这些模型。

It's all about how you use the models.

它们之间的差别微乎其微，而且还在不断变化。

And they're so close to each other and they're changing all the time.

如今的骗子和网络罪犯，用的也是同样的AI图像生成工具。

And so, you know, scammers today, cybercriminals today can use the same AI image generators.

他们用的聊天机器人背后，和我们用的是同一套模型。

They could use the same models that are powering the same chatbots.

只要三秒钟的音频，就能克隆出声音。

They can do voice cloning off of three seconds of audio.

这些工具现在全都被骗子拿来用了。

All of these tools are available to scammers today.

你根本不需要最顶尖的模型，就能对普通用户造成这么大影响。

So you don't have to be running off of the most sophisticated models to have the impact on consumers that we're seeing.

说到这个，我 personally 有切身体会——我妈妈曾经被一个伪造绑架的骗局盯上，骗子冒充了我妹妹。

And and, you know, for me, this this came home personally when when my mom was targeted with one of these fake kidnapping scams where where cybercriminals were able to impersonate my sister.

他们伪造了她的电话号码。

They spoofed her phone number.

还可能克隆了她的声音，让我妈妈以为我妹妹被人绑架了。

They potentially cloned her voice and convinced my mom that she was had been taken hostage.

要放人的话，就得立刻通过 PayPal 转 1200 美元。

And the only way to set her free was if she, you know, sent $1,200 over PayPal immediately.

正是这件事让我决定写这本书。

And that's what sparked the interest in writing a book for you.

对吧？

Right?

是的。

Yeah.

对。

Yeah.

这件事发生后，我开始深入研究——我过去二十年都在从事国家安全和企业网络安全工作，但妈妈遭遇这起骗局后，我才意识到，这些网络犯罪组织背后的经济模式到底发生了什么变化，居然让他们能发动如此精准、复杂的攻击，而这种攻击以前只可能针对政府机构或财富五百强企业。

So, you know, that sent me after I spent the last twenty years working in national security and ultimately enterprise cybersecurity, when that incident happened with my mom, it sent me down the rabbit hole of looking at, really what had changed in the underlying economy of these these cybercriminal organizations that allowed them to deploy that kind of a sophisticated and personalized attack that we used to see for you know, saved for those three letter government agencies or for a Fortune 500 company.

但如今，这种级别的精密手段却被用在了堪萨斯州草原村一个叫贝丝·考夫林的普通女孩身上。

But now that kind of sophistication was being pointed at little Beth Coughlin in Prairie Village, Kansas.

这让我开始想，还有多少人正被这类骗局盯上？

And so it made me think, you know, how many more people were being targeted with these types of scams?

我们眼前这些AI实验室描绘的技术路线图，对那些利用技术剥削消费者的网络罪犯来说，究竟意味着什么？

And what would this technology roadmap that we're seeing laid out in front of us from the AI labs, what would that mean for the cyber criminals and the bad guys who are using this technology ultimately to exploit consumers?

你在写这本书的过程中，研究中最让你惊讶的发现是什么？

What was the most surprising thing you found in your research as you went to go write this book?

我觉得最让我意外的是，当我深入暗网购买诈骗工具、调查这些组织时，发现它们真的像正规公司一样运作。

I think it was, as I, you know, was buying scam kits on the dark web and looking into these organizations, and they really are organizations, what surprised me is how they are operating like businesses.

事实上，从东欧到东南亚、西非，甚至就在我们身边，这些诈骗团伙其实是最早、最赚钱的AI应用者。

In fact, you know, these scam syndicates from Eastern Europe to Southeast Asia to West Africa and even closer to home, you know, in some ways, they're the most profitable AI early adopters that we have.

研究显示，一旦这些组织采用AI，它们的诈骗量能增加九倍，收入能提升四倍，远超那些不用AI的对手。

Research shows that when these organizations deploy AI, they're able to generate nine times as much volume and four times as much revenue than their sort of less AI inclined peers.

所以虽然——

So while-

为什么呢？

Why is that?

但为什么会这样呢？

And why is that though?

你觉得这背后的原因是什么？

What do you chalk that up to?

这全靠个性化。

It's all about personalization.

我的意思是，你看，现在这种生成式AI的强大之处在于，它能说得跟我们一模一样。

I mean, look, the power of this generative AI trend that we have is that it can talk like us.

它能表现得跟我们一模一样。

It can behave like us.

它还能生成逼真的图片，人眼根本看不出破绽。

It can create pixel perfect images that are impossible for us to detect with the human eye.

同样的技术，一方面能大幅提升我们的效率，另一方面，对那些想骗我们的人来说，简直就是工业级的武器。

That same technology is really powerful for making us more productive, but it's also an industrial grade weapon for people who want to deceive us.

但我想问的是，他们是不是特别擅长使用这些工具？

But I guess what I'm getting at is, are they just really good at using the tools?

他们用这些工具干坏事，所以你开什么价都行，这很明显。

They're using it for nefarious purposes, so you could charge whatever you want, obviously.

我当然明白，在某些情况下，赚钱可能要容易得多。

So it's not lost on me that revenue might be easier to come by in some cases there.

但他们是不是在用这项技术的方式上，比正规企业更厉害？或者说，你觉得原因到底是什么？

But are they particularly talented at using this technology in ways that, I don't know, above board enterprises are not, or what do you chalk it up to?

我觉得有一部分确实是这样。

I think it's some of that.

这些组织一直善于搭上重大技术变革的顺风车，借此寻找新的方式来坑害普通人。

These organizations have always sort of piggybacked on big technology transformations and used those to find new ways to exploit everyday people.

但更重要的是，这些工具本身特别适合用于冒充和诈骗。

But I think what's more than that is the tools themselves are perfect for impersonation and impostor scams.

现在数字诈骗中，几乎有三分之一都是冒充类骗局。

Digital fraud now is made up almost a third of impostor scams.

所以，向联邦调查局或联邦贸易委员会报告的案件中，有三分之一都是冒充骗局——骗子假装成你可能会信任的人，比如你熟知并喜爱的品牌，像微软、苹果、贝宝，或者邮政服务，也可能是你亲近的人。

So thirty three percent of cases that are reported to the FBI or to the FTC are some sort of impersonation where a scammer pretends to be somebody you are likely to trust, whether that's a brand you know and love like Microsoft or Apple or PayPal or the postal service, or maybe it's somebody that you love.

这些工具让骗子很容易模仿我们的声音，让我们困惑，最终轻松骗走我们辛辛苦苦赚来的积蓄。

These tools make it very easy to sound like us and to confuse us and ultimately to make it easy for scammers to help us depart from our hard earned savings.

对。

Right.

跟你聊聊你创办的公司Savi Security吧，它具体是做什么的？

Tell me about the company, Savi Security, that you've started.

这家公司在这个领域里的目标是什么？

What is the ambition of this company in the context of what you're focused on?

是的。

Yeah.

就像我之前说的，过去几年，企业和政府在安全AI和开发防御工具来对抗AI黑客和国家行为体方面，已经取得了大量创新。

So like I said, you know, there's been a ton of innovation for enterprises and the government around, you know, safe AI and creating defensive tools to protect against AI powered hackers and nation state actors.

但我觉得，在过去二十年里，消费者这边却几乎没有看到同等程度的投资和创新。

I think what we haven't seen in the last twenty years is the same kind of investment and innovation on behalf of the consumer.

如果你看看消费级网络安全市场，以前在麦咖啡和诺顿那些年代，它的规模远超企业级市场。

If you look at the consumer cybersecurity market, you know, it used to dwarf the enterprise cybersecurity market, if you go back to the times of the McAfee's and the Nortons.

但过去二十年，这个市场基本没什么增长。

But for the last twenty years, that market has stayed relatively flat.

与此同时，攻击者却把目标转向了普通消费者，因为在这里更容易赚到钱。

And meanwhile, the threat actors are shifting their crosshairs down market to focus on the consumers more because they can profitably extract value there.

所以我们需要为消费者带来一波全新的创新。

So we need a whole new wave of innovation on behalf of the consumer.

而我们正从攻击者最得手的地方开始着手。

And we're starting with the places where the threat actors are having the most success.

短信已经取代邮件，成为消费者遭遇诈骗最常见的渠道。

So text just surpassed email as the most common, vector for fraud for consumers.

邮件可是长期占据主导地位的。

It's been email forever.

是的。

Yeah.

现在短信和文本消息，我们都深有体会。

Now SMS and text, and we all feel it.

每天都能收到那些骗人的垃圾短信，谁都没法忽视。

We feel it every day and are the scammy and spammy texts that we get.

语音通话在大额诈骗损失方面是个大问题。

And voice calls are a huge issue when it comes to the big ticket losses.

你看那些五位数、六位数的损失，骗局不一定从电话开始，但最终受害者总会跟骗子通上电话，结果损失的钱就多得多了。

If you look at the five and six figure losses, the scams don't always start with call, but at some point, you know, a victim gets on the phone with a scammer and ultimately they lose significantly more money.

所以我们现在先从怎么过滤垃圾短信入手，再给用户提供更多工具，帮他们识别是不是在接诈骗电话或骚扰电话。

And so we're starting with how do we filter out scammy texts, how do we provide more tools to consumers to understand, you know, when they're on a scam call or a spam call?

那我是不是只要注册一下，把手机号和邮箱给你，你们就能帮我拦截这些诈骗电话？

So do do I just do I sign up and and and give you my phone number, my email, and then you basically protect you deflect these these scam calls?

这具体是怎么运作的？

How does it work?

是的。

Yeah.

我们最初推出了一款免费工具。

So we we we started with a a free tool.

你可以在 scamwise.com 上找到它。

It's available at scamwise.com.

你可以提交一条短信的截图。

You can you can submit a screenshot of a text message.

也可以拍下你不确定真假的纸质信件。

You could take a picture of physical mail that you don't know whether that's real or not.

还可以把邮件转发给我们，或者直接跟我们对话。

You could forward an email to it, and you can just talk to it.

几个月前我们上线这个工具后，用户反馈说他们希望能在手机App里使用。

And what we heard from users when we launched that a few months ago is that they wanted to see that in an app.

现在，我们每个月通过 Scamwise 处理超过十万条可疑信息，最近我们推出了 Savi 应用，内置了更主动的防护功能。

So we're now processing, you know, over 100,000 suspicious messages every month through Scamwise, and we just launched our Savi app where we've built in more proactive protection.

你不需要再特意上传信息来检测了。

So you don't have to actually come and submit the message.

我们实际上可以在您的设备上部署工具，在这些信息到达您之前就将其过滤掉。

We can actually deploy things on your devices to filter out those messages before they even reach you.

您觉得呢？我在想，再回到这些模型上，政府目前的立场是，他们说：我们已经有了一些自愿性框架。

Do you think you know, I'm thinking about the, again, going back to the models, the position that the government has taken, that the government has said, Hey, we have these voluntary frameworks in place.

他们希望与AI实验室合作，共同规范这些模型的发布方式。

We would like to work with the AI labs on how they release these models.

我们当然也注意到，现在这种分阶段发布的方式已经越来越普遍了。

We of course know that these staggered releases have become a little bit more common now.

您认为，在涉及模型发布这一块，政府是否应该在AI安全方面发挥作用？

Do you think that the government should have a role in AI safety as it relates specifically to the models that are released?

说实话，我觉得我们需要多管齐下。

Look, I think we need a whole host of things.

我们需要更多的创新。

We need more innovation.

这正是我们在Savi所做的，但同时我们也需要更明智的政策。

That's what we're doing at Savi, but we also need smarter policies.

我们家庭里需要更多仪式感，用不同的方式来讨论这些问题。

We need more rituals in our families to talk about these things in different ways.

当然，这也会是其中一部分。

And of course, is going be a part of that.

不管这是来自行业还是政府，我认为都有多种方式可以帮助解决问题。

Now, whether it comes from the industry or whether it comes from the government, I think there's different ways that you can help solve the problem.

但说实话，我们现在落后太多了。

But, you know, right now, we're way behind.

我们在制定应对这项技术的政策上，已经远远落后了——这技术正落入坏人手中。

We are so far behind in the kinds of policies that we need to keep up with this technology that is getting into the hands of bad guys.

这并不是第一种具有双重用途的技术，它既能为个人、企业、机构和政府带来巨大好处，也可能被不良分子利用，造成可怕后果。

This isn't the first technology we've created that has dual use, that can do wonderful things for people, for companies, for enterprises, for governments, but can also do scary things when they're in the hands of bad actors.

现在，在这个国家，买伪麻黄碱比获取最先进模型的API密钥还要难，而后者转手就能变成诈骗分子的完美工具。

Right now, it's harder to buy Sudafed in this country than it is to get an API key for the most advanced model and to turn it into the perfect weapon for scammers.

所以我觉得，我们完全可以做更多，增加使用门槛，让这些组织更难拿到最前沿的技术，再拿去对付普通用户。

And so I think we can do more to increase the friction so that it's harder for these organizations to get their hands on the most cutting edge technology and then turn that against consumers.

没错。

Right.

太好了。

Great.

帕特里克，非常感谢你今天来做客。

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

这位是萨维安全公司的联合创始人兼首席执行官帕特里克·考夫林，欢迎来到TITV。

That is Patrick Coughlin, the co founder and CEO of Savi Security here on TITV.

本月早些时候，《信息》发布了一份特别报告，梳理了我们认为最有可能被收购的160家企业软件公司。

Earlier this month, the information put out a special report covering the 160 enterprise software companies that we think are likely to be acquisition targets.

这是我们连续几年一直在做的年度项目。

This is an annual project that we've been working on for a few years now.

这份报告生动地反映了当前软件行业的现状。

It is a telling snapshot of the software sector.

接下来，我们邀请到这篇报道的记者阿里克斯·库图尔。

I want to bring on Alix Coutures, the reporter behind that story.

Alix，欢迎来到节目。

Alix, welcome to the show.

很高兴你能来。

It's great to have you here.

你好，Akash。

Hi, Akash.

Alix，跟我说说，你是怎么整理出这份名单的？

Tell me, Alix, So how did you put this list

我先以PitchBook提供的一份名单作为基础。

I worked as a starting point with a list provided by PitchBook.

那是一份包含300家初创公司的名单。

It was a 300 startup list.

然后我查阅了最新的动态，并对所有初创公司进行了筛选。

And then I looked for any updates or really screened all the startups.

我重点查看了它们的融资情况。

I looked for updates in fundraising.

我还关注了估值方面的最新动态。

I looked for updates in valuation.

我只筛选了过去两年没有融资、属于私营企业且估值超过十亿美元的初创公司。

I looked only at the startups that haven't raised money in the past two years that are privately owned and that are valued more than $1,000,000,000.

之后，我重点研究了那些处于消费市场与企业市场交叉点的公司，并逐一筛选出来。

So then after that, I looked at a lot of companies that set at the intersection of the consumer market and the enterprise market, and I took those one off.

然后我联系了所有这些公司，核实最新情况，确保它们确实符合标准。

And then I reached out to all the companies to look for any updates and make sure that they really fitted the criterias.

最终我整理出了一份包含160家初创公司的名单，数量比去年翻了一倍。

And then I ended up with a list of 160 startups, which is a lot because it doubles since last year.

这是因为许多原本有意收购的企业现在都撤回了对软件公司的兴趣，转而把重心放在了人工智能领域。

And that's because a lot of would be buyers have backed away from software companies because they redirected their attention to AI.

同时，很多这类公司本身已经盈利，根本不需要再融资。

And also because a lot of those companies are profitable and they actually don't need to raise money.

所以它们完全符合我的筛选条件。

So they do fit the criteria.

我想聊聊名单上的这些公司，待会儿我们会具体讲几个例子。

So I want to talk about the companies that are on this list, and we'll go through a couple names in particular in a minute here.

不过总的来说，你提到的这160家公司，从整体上看，有哪些共同的特点或聚焦方向让你印象深刻？

But just broadly speaking, across these 160 names that you identified, what were the themes that stood out to you about who these companies are and what they focus on?

其中一个反复出现的趋势，无论是在和投行、投资人还是创始人交流时，还是在名单上，都指向了数据类初创企业。

Well, one theme that surfaced a lot, both in conversations with bankers, investors, and founders, and on the list, our data startups.

很多初创公司都在做和数据相关的事情。

A lot of startups are working with data.

它们不仅开发数据软件，还拥有自己的专有数据集。

They put together data softwares, but also have proprietary datasets.

随着人工智能成为焦点，它们认为数据的价值比以往任何时候都高得多。

And they think that data, as AI is taking the center stage, they think that data is dramatically more valuable.

因此，有不少数据类公司非常值得关注，对软件公司、私募股权机构和大型科技企业来说都极具价值。

So, there's a lot of data companies that are really interesting and that are very valuable to software companies, private equity firms, and big tech companies.

此外，还有很多公司在生产力工具、销售和营销领域也很活跃。

There's also a lot of companies both in productivity and sales and marketing.

这是因为进入门槛低，产品容易开发也容易分发，但要增长就难多了。

That's because there's low barriers to entry and it's easy to build and easy to distribute, but then it's more difficult to grow.

所以对这些公司来说，只有那些在AI领域真正加速的，才可能成功并找到退出路径。

So for those companies, only those that are really accelerating in AI will really succeed and find an exit.

数据这个类别挺有意思，因为我们之前在节目里聊过，像前端公司、应用层公司、CRM公司这些，其实更容易被 vibe coding 取代。

And, you know, the data category is interesting because we've been talking on the show about how front end companies, application layer companies, CRM companies, for example, those are the companies that can probably conceivably be replaced a little bit easier with vibe coding.

数据公司则不同，它们的基础设施很有价值，很适合被收购。

Data companies, I mean, you could say that they could be ripe for acquisition because we could use that infrastructure.

这也意味着它们可能还会再融资，所以绝对值得关注。

It also could mean that they could raise again, and so they certainly are a category to watch.

在这些数据公司里，有没有哪家特别让你觉得会是收购目标？

Tell me, in this data category of companies, was there any one particular company that stood out to you as a takeover target?

有的。

Yeah.

在数据领域，有一家叫 Project 44 的公司。

In the data category, there's one that's called Project 44.

这公司不是数据公司，而是做物流的，只不过用到了数据。

So it's not data company, but it's a logistics company that works with data.

它从成千上万的运输商那里收集货运数据，既有基于算法的数据网络，也有海量的货运数据集。

So it pulls shipment data from thousands of carriers and it has both data graphs, so a data network powered by an algorithm, and gigantic datasets of shipment data.

这家公司对大企业、软件公司或者私募股权机构来说非常有价值，因为它掌握着独家数据。

And that company is extremely valuable to big companies or to software companies or private equity firms because it has proprietary data.

对AI公司来说尤其有意思，比如那些想训练和运行模型的公司，特别需要独家数据集。

So, it's really interesting for AI companies, for example, that wanna train and run their AI models to have proprietary datasets.

这家公司的CEO告诉我，这几年已经有不少人来接洽，今年的问询量是去年的十倍。

So that company, the CEO told me that it's been approached several times and that the number of inquiries has has been multiplied by 10 since last year.

所以确实有很多人来找他们。

So they are getting inquiries then.

要说起这份名单，得先说明，这不光是我们从PitchBook数据里筛选出来的。

Mean, to start us on this list, we should point out, this is not just based on our filtering of PitchBook data.

而是基于我们实际做的调研和报道。

I mean, is based on real reporting that we are doing.

说说名单上最大的公司 Miro，它怎么样？

Tell me about the biggest company on this list, Miro.

关于这家公司，我们了解多少？还有哪些关键信息？

What do we know and need to know about that company?

Miro 是一家生产力工具公司，估值高达 170 亿美元，是我们名单上估值最高的公司之一。

Well, Miro is a productivity company, and it's valued $17,000,000,000 So that's a lot, and it's one with the highest valuation on our list.

它是一家生产力领域的初创公司，特别吸引大科技公司和私募股权机构的关注。

It's a productivity startup that really draws interest from both big tech companies and private equity firms.

简单说，它是一个可视化协作平台，帮助混合团队一起工作、制定策略、头脑风暴，而且它本身也是一个收购方。

So basically, it's a visual workspace that helps hybrid teams work together, plan strategies, brainstorm, and it's really an example of a startup that has also been a buyer.

今年三月，它收购了 AI 生产力初创公司 Reforged。

So it acquired the AI productivity startup Reforged in March.

它的增长速度非常快。

So they're really growing very quickly.

而且利润相当可观。

They're very profitable.

他们没透露是否考虑过退出，所以拒绝置评。但这家公司盈利良好，增长迅速，而且确实吸引了不少关注。

They haven't said whether or not they were interested in an exit, so they declined to comment, But they are profitable and growing quickly, and we know that they are drawing interest.

不过，以这么高的估值来说，退出并不容易，毕竟谁会愿意花一千七百亿美元去买一家公司呢？

However, at this high valuation, an exit would not be easy, because who would want to buy a company for $17,000,000,000

我的意思是，当然总有被收购的可能性，但价格肯定不会是一千七百亿美元，更可能是比上一轮融资估值打个折——这种事以前也发生过。

Well, I mean, look, there is always the option that you get bought, but it's not for $17,000,000,000 and maybe it's a discount to your last funding round, which, look, we've seen it happen before.

你之前说，Miro的创始团队没有回应你的提问，我没记错吧？

You said that Miro's founders, I believe you said that they did not respond to a comment to you, if I'm not mistaken?

他们没对这个问题做出回应。

They didn't respond to the question.

他们其实是有意退出的。

They were interested in an exit.

我本来想问的是，在你接触过的所有创始人里，有没有谁直接说过：‘对，我们愿意卖’？

I mean, the question I was going to ask is, of all the founders that you did talk to, was there any founder that was open in saying, Yeah, I mean, we'll sell.

你知道的，就是表示愿意被收购的那种。

You know, we'd be open to an acquisition.

很多人说他们没兴趣。

Well, a lot of them said they weren't interested.

有些人直接说，我们不感兴趣。

Some of them said, We're not interested.

我们现在只专注于发展。

We're just focused on growing.

但我们从不直接拒绝，这要看价格。

But we would never say no, it depends on the price.

所以很多人其实留着余地，没把话说死。

So a lot of them say, well, they leave the door open.

有家公司的态度特别明确，他们确实对Nexa感兴趣。

One in particular has said and confirmed that they were interested in Nexa.

就是那家人力资源公司，Papaya Global。

It's the HR company, Papaya Global.

他们估值37亿美元，已经从一家人力服务公司转型成了做支付科技的金融科技企业。

It's valued $3,700,000,000 and it really transitioned from being an HR firm to being a fintech firm that helps with payments.

这家公司很有价值，而且特别突出，因为它成功地从人力资源公司转型成了金融科技公司。

And that company is valuable and really stands out because it did a real transition from HR to FinTech.

这家公司现在引起了PEO公司和软件公司的兴趣。

And that one is drawing interest from PEO firms and software companies.

它愿意被收购，目前正处在谈判阶段，今年就有可能完成出售。

And it is open to an exit, and it is in talks right now to sell the company, and the sell could happen this year.

有些公司虽然没明说，但像这家就是这种情况。

So some companies haven't admitted it this, such as this one.

你刚才提到了软件公司。

You mentioned software companies.

那大型上市公司、科技巨头，是不是最有可能的买家？

So are big public software companies, big tech companies, I mean, that the most logical buyer here?

你觉得谁会是这些公司的收购方？

Who could you see as being an acquirer for these businesses?

那些运营人工智能模型的大型科技公司，可能性最大。

Well, they would be the big tech companies that run AI models.

所以他们主要对数据公司感兴趣，因为他们需要数据来训练和运行自己的AI模型，也要用这些数据服务客户。

So, they would be mostly interested in data companies because they need data to train and run their AI models, and customers their AI models.

因此，专注数据的软件公司，最大的买家就是那些大型科技AI公司。

So for data software companies focusing on data, it would be big tech AI companies.

至于一般的软件公司，买家则是大型软件企业。

Then for software companies, it would be large software companies.

但像施耐德电气这样的工业巨头也想加快布局AI，他们同样会关注那些在AI领域进展迅速的软件公司。

But industrials like Schneider Electric also want to accelerate in AI, and they would also have an interest in buying software companies that have accelerated in AI.

另外还有私募股权公司。

And then there's private equity firms.

今年26年初，并购交易数量确实放缓了，但他们仍在积极寻找，目标同样是那些在AI方面表现突出的软件公司。

So the number of deals have slowed down this year in '26 at the beginning of this year, but they're also hunting and they're also looking to acquire software companies that have accelerated in AI.

不过对私募股权公司来说，他们更倾向于收购那些AI转型不够快的初创企业，这类公司估值会打折扣，因为对大多数创始人来说，这不是最理想的退出方式。

However, for those companies, so for private equity firms, it would mostly be startups that haven't accelerated fast enough in AI, and they would sell at a discount because it's not the most attractive exits for most founders.

对。

Right.

对于所有这些买家来说，他们最感兴趣的是那些在AI领域真正加速发展的软件公司。

And for all those buyers, they would be interested in software companies that really have accelerated in AI.

因为现在光是做普通软件的公司，对这些买家来说未必有吸引力。

Because as it is, just software companies doing software is not necessarily attractive to those companies.

他们真正关注的是那些已经全力投入AI、并取得显著进展的公司。

And they're really looking into those that have accelerated and doubled down on AI.

艾丽克斯，你接触了这么多与这些公司相关的人——投资者、创始人、分析师，甚至客户，你从他们那里了解到，AI在企业软件领域到底带来了怎样的格局变化？

Tell me, Alix, as you talk to so many people in the orbit of these companies, investors, founders, analysts, customers even, what did you learn about the state of the war on enterprise software that AI in some cases, has positioned itself as?

人们说，我们自己也能开发出这些应用。

People say, We could just make these applications on our own.

软件股到现在还没涨起来。

Software stocks have yet to come up.

我们上周还和瑞银的卡尔·克尔斯塔德聊过这个话题。

We were talking to Carl Kerstad from UBS about this last week.

对于这些规模较小、未上市的公司来说，现在SaaS的寒冬到底是个什么状况？

What is the state of the SaaS apocalypse for these smaller, non public companies right now?

这得看具体情况。

Well, it really depends.

对一些公司来说，情况非常艰难，因为很多投资者已经把注意力从软件公司转向了AI公司。

So, for some of them, it's been very hard because a lot of investors have shifted their attention away from software companies to AI companies.

所以对他们而言，融资变得极其困难，心理上也受到了很大打击。

So, So, for some of them, it's been very difficult to raise money and it really hit them hard psychologically.

但对另一些公司，比如数据公司，他们认为AI让数据的价值大幅提升。

For others, such as data companies, they think that AI is making data dramatically more valuable.

因此他们觉得这是个巨大的机会，而且现在确实吸引了所有关注。

So they see this as a huge opportunity, and they think that they're getting all the attention.

对大多数软件公司来说，目前他们只想稳住阵脚。

For most software companies, they just want to hold right now.

他们觉得公司不太可能被收购。

They're thinking we're not gonna be acquired.

融资确实不容易，但至少他们还在盈利。

It's not easy to raise money, but they're still profitable.

很多公司其实都在盈利运营。

A lot of them are actually operating profitable.

他们在2021年融了不少钱。

They raised a lot of money in '21.

花得不多，经营得不错，还在稳步增长。

They didn't burn that much and they're doing pretty well and growing.

所以他们正抓住这个时机，全力投入AI发展，好为将来提升吸引力。

So what they're doing is they're using that moment as an opportunity to really grow and accelerate in AI so that they can be more attractive down the road.

这正是大多数公司的策略。

And that's the strategy of most of them.

他们根本没打算退出，尤其是不想低价套现。

Most of them are not gonna be looking for an exit, especially an exit as at a discount.

他们就等着，同时加速布局AI。

They're just gonna wait, and accelerate in AI.

即使被SaaS寒冬的说法影响过情绪，他们依然把这看作一个机会。

And they see this as an opportunity, even if they've been psychologically impacted by the SaaS apocalypse narrative.

Alix，这个项目非常棒，我建议大家都去了解一下，里面有不少知名公司。

Well, Alix, it is a great project, and I encourage everyone to go check it out because there are some bigger well, not bigger company, well known company.

Notion 就在名单上。

Notion is on the list.

Snyk 也在名单里。

Snyk is on the list.

这些公司覆盖了软件行业的各个领域。

There's companies from every corner of the software sector.

这是《The Information》的记者 Alix Coutures，她负责报道企业级软件初创公司。

That is Alix Coutures, our reporter covering enterprise software startups here at The Information.

今天的节目就到这里。

That does it for today's show.

提醒一下，我们每周一到周五早上10点（太平洋时间），下午1点（东部时间）直播。

A reminder, we are on this stream Monday through Friday at 10AM Pacific, 1PM Eastern.

如果错过直播，节目回放可以在 theinformation.com、我们的YouTube频道，或者你常用的播客平台收听。

If you can't make it then, episodes are available on theinformation.com, on our YouTube channel, or wherever you get your podcasts.

别忘了在社交媒体上关注我们，包括 X、Instagram、TikTok 和 LinkedIn。

Make sure to follow us on social media, on X, on Instagram, on TikTok, and on LinkedIn.

我已经开始期待明天的节目了。

I am already excited for our next show tomorrow.

祝你周一剩下的时间愉快。

Have a great rest of your Monday.

那我们先告一段落，再见。

Bye bye for now.