贝莱德Tony Kim谈AI交易、高盛经济学家AI报告及AI会计 | 2025年11月4日

欢迎各位收看资讯TITV。我是主持人Akash Pesritcha。今天是11月4日星期二。今天的节目内容非常丰富。首先我们将带来财报快讯。

Welcome everyone to the Informations TITV. My name is Akash Pesritcha. It is Tuesday, November 4. We have got a jam packed show for you today. First up, we've got our earnings blitz.

优步和Shopify今早公布了财报。Palantir则是昨晚公布的。我们的记者将协助解读这些财报。我们还刚刚发布了Anthropic大幅上调增长预测的重磅消息，稍后会为您带来详细报道。

Uber and Shopify reported this morning. Palantir reported last night. We've got our reporters coming on to help us break them all down. We've also just published some big news that Anthropic is dramatically raising its growth forecasts. We'll tell you more about that shortly.

之后我们将邀请主编Jessica Lesson与来自贝莱德的朋友们共同探讨当前人工智能的发展态势。高盛集团也将做客节目，分享他们关于企业应用AI的最新调研数据。节目最后将推出特别环节——正式发布'资讯50强'最具潜力初创企业榜单（简称TI50）。订阅用户可在官网查看完整名单，本周我们将通过创始人访谈为您提前揭秘部分上榜企业。今天的节目非常紧凑，让我们马上开始吧。

After that, we're bringing on our editor in chief, Jessica Lesson, for some broader reflections on AI in this moment with our friends over at BlackRock. We've also got Goldman Sachs coming on the show to talk about some new data they have on how companies are using AI. We're going to close things out with a special segment, the launch of the information's 50 most promising startups known as TI50. Subscribers can get the full list on our website, but all week, we're going be giving you a preview with conversations with some of those founders. It's going to be a busy show, so let's get right on into things.

Shopify今早公布了财报。公司营收增长32%，但更有趣的其实是各业务线的增长率。下面有请我们的电商记者Anne Gian为您全面解析。Anne，欢迎来到节目，很高兴今早能邀请到您。

Shopify reported earnings this morning. The company reported 32% revenue growth, though the more interesting story is, of course, the growth rates of the different business lines. Here to break it all down is our e commerce reporter, Anne Gian. Anne, welcome to the show. It's great to have you this morning.

嗨，Akash。很高兴见到你。

Hi, Akash. Great to see you.

那我们聊聊Shopify吧。今早的财报中哪些亮点最吸引你？

So let's talk Shopify. What stood out for you from the earnings this morning?

是的。我认为最值得注意的是他们的营收增长持续保持稳定，过去几个季度一直保持着30%左右的增长率。投资者对此非常满意，而且他们并未看到关税增加带来的任何放缓或影响。公司还表示没有观察到消费者支出出现明显放缓——这是当前许多投资者密切关注的指标。因此这份营收数据整体反响不错，基本符合市场预期。

Yeah. So I think one of the most notable things is their revenue growth keeps chugging right along that roughly 30% growth rate has been what they've been posting for the past several quarters. So investors really like to see that and they haven't really seen any slowdown or any dent from increased tariffs. They also say they haven't really seen a meaningful slowdown in consumer spending, which is something that a lot of investors are tracking really closely right now. So I think that top line number was pretty well received and I think in line with what people were expecting.

Shopify一直在积极拓展其非传统核心业务领域。他们与欧洲商户的合作大幅增长，同时也在那些实体店销售额高于线上销售的商户中取得显著进展。这持续强化了他们不仅限于线上销售、而是能与任何零售商合作的品牌叙事——这正是投资者希望看到的。今早财报中值得注意的一点是，他们本季度的成本略有上升。

Shopify has been really pushing to expand in what hasn't historically been their core categories. So they've been growing a lot with merchants that are based in Europe. They've been growing a lot with merchants that do more of their sales in brick and mortar stores versus online. So they're really continuing to sell this story that they're for more than just online sales and they can really work with any retailer, which is what investors want to see. I think the one interesting thing to call out from the earnings this morning were that their costs rose a little bit this quarter.

公司表示这主要源于营销费用增加及部分研发成本，同时也加大了对AI的投入——这是Shopify近期重点发力的另一个方向。

And they said that was mostly due to, increased marketing expenses as well as some R and D costs, but also, increased spending on AI, which is another story that Shopify has been looking to really lean into lately.

从你观察来看，他们是如何全面布局AI的？

How are they leaning into AI broadly from what you've seen?

这主要体现在两个层面：对内用于加速公司运营，他们频繁强调AI原生理念，将AI应用于产品开发、用户反馈分析以及理解商户需求；对外则整合进产品线，比如几周前宣布与OpenAI合作，让现有Shopify商户能更便捷地通过OpenAI新电商功能销售商品——这一举措引发了广泛热情。

So it's kind of twofold. It's both internally, you know, within the company to speed up their operations. They talk a lot about being very AI native, using AI for product development, product feedback, understanding what merchants want from Shopify, but then also, you know, incorporating it in into their products as well. So a couple of weeks ago, they announced a partnership with OpenAI to make it easier for all of the merchants that are already using Shopify's software to sell products through OpenAI's new ecommerce features. That's something that a lot of people are really excited about.

我认为这种策略与Shopify历史上开拓新销售渠道的方式一脉相承。回想TikTok小店推出时，Shopify就是最早与其合作的平台之一。这再次证明他们正全力帮助商户实现全渠道销售，而抢先布局AI自然令投资者倍感振奋。

But I think that approach is pretty in line with how Shopify has historically approached some of these newer sales channels or places to sell. If you think back to the launch of something like TikTok shop, Shopify was very early to partner with them. So, I think it's just kind of a continued, you know, another example of Shopify really making a making a big effort to help their merchants kind of sell anywhere they possibly can. And obviously being early to AI is something that investors are really excited about.

好的。安妮，感谢你的分享。需要说明的是今早股价下跌，部分原因可能如你所说与公司预期有关——当前超30%的增速预计将回落至25%-30%区间，这将成为本季度关注重点。非常感谢你的解读。

Great. Well, Anne, I want to thank you for coming on. We should say that shares were down this morning, and I think part of that might have to do with the guidance as well that the company gave that, as you said, growth is north of 30% right now. They're thinking it's going to come down into the mid to high 20s, and so that will be something to watch in the quarter. But I want to thank you for coming on.

这位是我们The Information的电商记者Ann Gein。今早Uber和Palantir股价双双下跌，Uber今早刚发布财报，Palantir则是昨晚公布的——两家公司的业绩表现都相当强劲。

That is Ann Gein, our e commerce reporter here at The Information. Uber shares were falling this morning, as were Palantir shares. Uber reported this morning. Palantir reported last night. Both companies reported pretty strong results.

当然，Palantir目前正处于飞速发展阶段。我想请我们的财经专栏作家Anita Ramaswamy来帮我们全面分析。Anita，早上好，很高兴见到你。

Palantir, of course, is on a bit of a rocket ship right now. I want to bring on our financial analysis columnist Anita Ramaswamy to help us break it all down. Anita, good morning to you. It's great to see you.

早上好，Akash。很荣幸来到这里。

Good morning, Akash. Good to be here.

那么我们来聊聊Uber和Palantir。我想先从Uber开始，因为他们今早发布了财报。从结果中你注意到哪些亮点？

So let's talk Uber and let's talk Palantir. I want to start with Uber because they reported this morning. What stood out to you from the results?

最引人注目的头条新闻是Uber的出行量同比增长了22%，这是该公司过去几年来最快的增长速度。我认为这非常重要，因为他们近期成功的关键在于外卖业务的快速增长，包括Uber Eats和杂货配送服务的扩展。同时他们的网约车业务也正处于关键转型期，正加大对自动驾驶车辆的投入。Dara在电话会议中也对此进行了大量讨论。

So the biggest thing was the top line news, which is that Uber actually grew its trips by 22% year over year, which is the fastest growth rate that we've seen from this company in the last couple of years. I think that's really significant because a big part of their success recently has been the rapid growth in their delivery business, which is the Uber Eats and grocery delivery that they're growing out. They're also at a pivotal time in their ride hailing business as they invest more in autonomous vehicles. So there was a lot of chatter about that on the call from Dara as well.

没错。我今早也听了电话会议，有趣的是虽然整体收入加速增长，但移动出行和外卖这两项业务都实现了加速发展，尤其是外卖业务简直像坐上了火箭。你能谈谈对此的看法吗？外卖业务会成为未来主力吗？

Right. And I should say, you know, I was listening to the call this morning. The interesting thing was revenue overall accelerated, but both of those business lines, mobility and delivery accelerated, and delivery was on a bit of a rocket ship. Can you talk a little bit about what you make of that? Is delivery the future here?

我在听电话会议时注意到讨论主要集中在自动驾驶车辆和欧洲市场。能否总体谈谈你对公司发展方向的看法？

I mean, was listening to the call, there was a lot of talk about autonomous vehicles. There was a lot of talk about Europe. Just talk to me broadly about where you see the business going.

确实。纵观两项业务，Uber在美国市场的增长速度远快于国际市场。本季度其国际业务增长基本停滞，这可能是股价小幅下跌的原因之一。但就外卖和移动出行业务而言，虽然外卖规模较小，但增长速度要快得多。

Yeah. I mean, taking a step back in both businesses, Uber has grown a lot faster in The US than it's been able to grow internationally. The growth rate in its international business actually stayed pretty much stagnant this quarter. That was one potential reason why we saw the stock fall a little bit. But when it comes to the delivery business and the mobility business, delivery is a lot smaller, but it's growing a lot faster.

这对他们来说是一项新业务。我认为这份财报中最令人印象深刻的数字或指标实际上是调整后的EBITDA在配送业务中增长了约30%。这值得注意，因为优步外卖目前正大力进军杂货配送领域，而杂货配送难度很大。像亚马逊这样的公司在这一领域历来举步维艰，其盈利能力通常不如餐厅配送。

It's a newer line of business for them. And I thought the most impressive number or metric in this earnings report was actually that adjusted EBITDA grew around 30% over that actually in the delivery business specifically. And that's notable Akash because right now Uber Eats is making a big push into grocery delivery and grocery delivery is tough. It's an area where companies like Amazon have historically struggled. It's not as profitable as restaurant delivery typically.

因此有可能我们看到优步外卖目前增速加快，但随着业务结构越来越多地向杂货端倾斜，这种增长可能无法长期持续。优步CEO达拉·科斯罗萨西在财报电话会上表示，优步外卖部门的杂货配送增速实际上远快于传统配送业务。

And so it's possible that, you know, we've seen this acceleration in Uber Eats now, but that may not be as sustainable in the long term as, you know, the mix of business shifts more and more towards the grocery side. I mean, Uber CEO, Dara Khosrashahi, said on the earnings call that grocery delivery is actually growing a lot faster than traditional delivery on the Uber Eats side of the house.

我们显然知道优步直到去年最近几个季度才实现盈利。具体时间我记不清了，但记得他们终于盈利时引起了很大轰动，人们都在感叹这段历程多么不易。今早电话会上讨论的重点是，正如你所说，我们正在向自动驾驶汽车和杂货领域投资，同时坚持盈利目标。你提到的调整后EBITDA确实值得关注。现在让我们快速转向讨论Palantir。

And we obviously know that Uber has just come into profitability as of the last couple quarters, last year. I can't remember when it was, but I remember it was a big deal when they finally turned a profit and everyone was sort of saying, Oh my gosh, what a story it's been. And the talk on the call this morning was we were making these investments into autonomous vehicles, into grocery, as you say, we are staying committed to profitability. And so as you said, that adjusted EBITDA number is certainly something to pay attention to. Let's pivot quickly to talk about Palantir.

这完全是另一回事了。哪些数据让你印象深刻？天啊，很难单独挑出一个数字，因为Palantir的各项数据都大得惊人。

This is a very different story. What stood out to you there? Gosh, it's hard to pick one number because the numbers are all so big when it comes to Palantir.

确实很难选择，但阿卡什，我认为Palantir最震撼的数字是其美国商业业务的增长——该业务收入增长了121%。

It is tough, but I think the biggest, most eye popping number Akash, that I was really thinking about when it comes to Palantir is the growth in their US commercial business. They grew revenue in that business by 121%.

翻了一倍多。确实翻了一倍多。

More than doubled. More than doubled. Yeah.

最近几个季度这项业务增长非常迅猛。传统上很多人认为Palantir是一家以政府收入为主的企业。

It's been growing really rapidly in the last couple of quarters. Palantir is traditionally known, I think to a lot of folks as a business that focuses on government revenue.

那是

That is

他们业务中的一大块，但其商业平台——人工智能平台AIP，已向众多企业销售，帮助这些公司整理数据以运行AI应用，取得了巨大成功，甚至开始超越美国政府业务的增长势头。

a big chunk of their business, but their commercial platform, the artificial intelligence platform, AIP, which they've been selling to a lot of different enterprises to help these companies get their data houses in order to run AI apps, has been wildly successful and has started to actually overshadow growth in The US government business.

不过这挺有意思的。商业业务正在蓬勃发展，规模增长了一倍多。我认为收入增长仍和上季度一样强劲。然而今早股价却下跌了。

So this is kind of interesting, though. The commercial business is booming. It's more than doubled. Revenue growth was still just as strong, I think, as last quarter. And yet shares are down this morning.

有人说这可能只是因为这只股票的估值已达到未来十二个月收入的约100倍，需要降温。你认为今天股价下跌说明了什么？

People are saying maybe it's just kind of a cool off of a stock that is trading at around 100 times next twelve months revenue. What do you think the stock being down today says?

我认为这恰恰印证了你所说的，市场对这家公司的期望值极高。投资者在做预测时，如果用贴现现金流法看未来一两年，可以说Palantir将继续快速增长，我们看到了强劲势头。但再往后就完全不确定了。股价的大部分价值来自第一年和第二年之后的年份，而这些我们根本看不清。

I think it just goes to show exactly what you said, that expectations are sky high for this company. When investors are making projections, if you take discounted cash flow method and you look at, you know, the next one or two years, you can say, okay, Palantir is going to continue to grow really fast. We've seen a lot of momentum. But beyond that, it's completely uncertain. Most of the value of that share price is coming from the years past year one and year two, which we don't have any visibility into.

我的意思是，我们知道他们的商业平台现在销售良好，但归根结底，这不是他们的核心业务。这部分收入非常新，尚未经过充分验证。所以即使假设Palantir未来五年、十年甚至永久保持30%以上的增长，也远不足以证明每股200美元的股价合理性。按这个算法，股价可能更接近每股50美元。

I mean, we know that their commercial platform is selling well now, but at the end of the day, that's not been their bread and butter. It's a very new and sort of not tried, not tested portion of their revenue. And so even if you assume Palantir is going to grow upwards of 30% for the next, let's say five years, ten years into perpetuity, that still doesn't come anywhere close to justifying a $200 per share price. I mean, the price at that point would be somewhere maybe closer to $50 per share.

你写过关于Palantir真实价值的专栏文章吗？你确实写过一篇对吧？

Have you written a true value column yet on Palantir? You did write one, right?

我记得当时股价确实更接近每股50美元时写过一篇分析

I did write one when the price, I believe, was trading a bit closer to that 50 per share

比...好吧，我

than Well, I

我们会把链接放在节目备注里。我需要重新梳理一下，但我记得你写过相关文章。谢谢你来参加节目，安妮塔。现在正值财报季忙碌时期，我们就不多耽误你了。还有其他多家公司即将发布财报。

we'll link it in the show notes. I have to refresh myself, but I I I remember you having written about it. Thank you, Anita, for coming on. It is a busy time for earnings, so we'll get you back to it. We've got a number of other companies reporting.

相信这周晚些时候还会再见到你。这位是我们《The Information》的金融分析师专栏作家安妮塔·拉玛斯瓦米。虽然OpenAI的财务状况备受关注，但今早我们独家披露的Anthropic最新财务数据更令人瞩目。不出所料，该公司预计营收将呈惊人增长，某些业务板块的表现甚至优于OpenAI。下面有请今晨这篇报道的作者斯里·穆皮提为我们深入解析。

And so I'm sure we will see you later on this week. That is Anita Ramaswamy, our financial analyst columnist here at The Information. Much has been made about OpenAI's financials, but a new exclusive story in The Information this morning reveals the latest picture of Anthropix financials. The company is projecting it will grow its top line at a staggering pace unsurprisingly, and in some categories of its business, it's actually doing even better than OpenAI. I want to bring on Sri Mupiti, who wrote that story this morning to help us break it all down.

斯里，很高兴见到你。最近怎么样？

Sri, it's great to see you. How are you doing?

我很好。很荣幸来参加节目。

I'm doing well. Great to be here.

那我们来聊聊今早这篇重磅独家报道吧。给我们详细说说这些数字——Anthropic的发展势头简直像火箭般迅猛。

So let's talk about this big scoop that you had this morning. Lay out the numbers for us. Anthropic is also on a bit of a rocket ship.

确实如此。Anthropic正在飞速发展。我们今天早些时候的报道显示，Anthropic预计到2028年将实现700亿美元收入和170亿美元利润——这还是最乐观的预测。基本上，Anthropic已将收入预期上调至高于上次预测的水平。这些数据早在2024年3月融资前就与投资者共享过。

Exactly. Anthropic is growing really quickly. What we reported earlier today is that Anthropic expects to generate 70,000,000,000 in revenue and 17,000,000,000 in profits in 2028, and that's in its most optimistic projections. Basically, Anthropic had raised its revenue projections to be higher than before when they had last made projection. They're shared with investors back before its March raise, which was done in late twenty twenty four.

正如我所说，我们看到Anthropic的发展速度远超预期。它预计最早在2027年就能实现正向现金流，当年利润将达到30亿美元。相比之下，OpenAI同年预计会有350亿美元的现金消耗，直到2030年才能实现盈利。

And what we see is that, as I said, that Anthropica is growing much faster than expected. It actually expects to be cash flow positive as soon as 2027 when it will generate 3,000,000,000 in profits that year. And comparing it to OpenAI that year, OpenAI expects to have $35,000,000,000 in cash burn and won't actually be profitable till 2030.

那我们来聊聊收入来源，因为与OpenAI相比，收入结构略有不同。

So let's talk about where the revenue is coming from, because the breakdown is slightly different than compared with OpenAI.

没错。Anthropic有两大收入支柱：首先是核心B2B业务，即通过应用程序接口（API）向企业客户销售Anthropix模型。这部分业务到2028年将占公司总业务的80%。例如今年，Anthropic预计通过API及相关销售将产生38亿美元收入。

Exactly. Anthropic has two main revenue drivers. The first is its main B2B business, which is selling Anthropix models to business customers via an application programming interface, also known as an API. And that makes up roughly 80% of Anthropic's business through 2028. So for this year, for example, Anthropic expects to generate 3,800,000,000.0 from API and related sales.

这大约是OpenAI预期API收入的两倍。不过OpenAI绝大部分收入来自ChatGPT，这里只是做个对比。Anthropix的第二大业务是专业消费者业务，即向自费订阅Anthropix云服务高级功能的个人用户或企业员工销售服务。这包括热门编程工具Cloud Code等，订阅费每月17至150美元不等。

That's roughly about double what OpenAI expects to generate from API sales. Although, of course, OpenAI generates a vast majority of its revenue from ChatGPT. That's just a comparison point. The second line of Anthropix business is its prosumer business, which is basically selling consumers or employees who pay with their own credit card access to Anthropix Cloud subscription premium features. So this includes, for example, the popular coding tool, Cloud Code, and the subscriptions for access to these premium features range from $17 to $150 per month.

我们此前报道的另一大收入来源是Cloud Code，它在开发者和数据科学家中迅速走红，这些人希望快速生成内容。预计这部分业务年化收入也将接近10亿美元。

And the other piece that we had reported, and that's a big revenue driver for Anthropic as well, is Cloud Code has just grown really quickly amongst both developers and even data scientists and other folks that wanna generate something really quickly. And they're expected to generate close to 1,000,000,000 in annualized revenue as well.

你报道中最让我印象深刻的是之前提到的观点：收入是一回事，但盈利能力才是当前AI公司最关注的焦点。正如你所说，Anthropic计划比OpenAI更早实现自由现金流。总体来看，其利润率似乎远高于OpenAI。我们是否了解Anthropic为何——或者说如何——比OpenAI运营得更高效？

So the thing that stood out to me from your story is what you were mentioning earlier. Revenue is one thing, but profitability is the thing that is very much in vogue right now for AI companies. And like you said, Anthropic is planning to get to generating free cash flow much sooner than OpenAI. Broadly speaking, the margins feel like they're a lot better than OpenAI. Do we have any sense for why Anthropic- how Anthropic is being more efficient than OpenAI?

这一切的根源是什么？

What's at the root of all this?

实际上，Anthropic去年利润率是负的。他们去年的毛利率大约是负94%，而OpenAI去年约为43%。但Anthropic预计今年及未来的利润率将略高于OpenAI。例如，Anthropic预计今年将达到50%，到2028年约为77%。而OpenAI今年是46%，2028年为67%。

So actually, Anthropic had negative margins last year. They had roughly negative 94% gross margins last year, comparing that to OpenAI's 43 or so percent for last year. But Anthropic expects to have slightly higher margins for this year as well as into the future as compared to OpenAI. For example, Anthropic expects 50% this year and then roughly 77% in 2028. OpenAI, meanwhile, has 46 this year and 67% in 2028.

我认为细微差异可能在于：Anthropic只计算付费用户使用其AI模型和产品的成本，不包括非付费用户，而OpenAI则涵盖两者。另外，Anthropic更专注于服务企业客户，这部分占其收入的80%，而OpenAI的产品类型非常多元化。因此利润率方面也可能存在差异。

I think what the slight differences might be is that Anthropic, for example, only includes the cost of running its AI models and products for its paying users, but not paying users, while OpenAI includes the cost for both. The other piece is that Anthropic just has more focus in terms of serving just, for example, business customers, which makes up 80% of its revenue versus OpenAI is very diversified across different types of products. And so there might be variants in terms of the margins as well there.

没错。我正想说，所有这些都指向一个核心问题——人们都在猜测Anthropic是否会进行新一轮融资。他们现在显然有更好的数据支撑。你觉得我们何时能看到进展？目前听到什么风声？

Right. And I was going to say that all of this is leading up to, of course, the anticipation of whether or not Anthropic is going to fundraise. They obviously have better numbers now to do it. When do we think we can expect that? And what are you hearing?

我听到风声说投资者很期待追加投资。上轮融资很多投资者没能参与，认购严重超额。因此如果如我们今早报道的那样启动融资，Anthropic估值可能达到3000亿到4000亿美元。相较之下，OpenAI近期员工股票出售的估值是5000亿美元，两者可谓针锋相对。我很期待未来几个月事态的发展。

I'm hearing chatter that investors are excited to pour more money. There were a lot of investors that weren't able to get into the last round, and it was really oversubscribed. And so I believe that if it were to raise, which we've reported this morning, Anthropic could probably expect a valuation between 300,000,000,000 to 400,000,000,000. And so comparing that to OpenAI's recent 500,000,000,000 in its employee tender, they're really going head to head. And so I'm excited to see the next couple of months play out as well.

太棒了。Sri，这是个精彩的报道，期待未来几天和几周内能继续带来更多消息。感谢你的分享。这位是The Information负责OpenAI和Anthropic报道的Sri MooPD。

Great. Well, Sri, it was a great story, and I anticipate that we're going to have more reporting from you in the days and weeks to come. Thank you for coming on. That is Sri MooPD who covers all things OpenAI and Anthropic here at The Information. Okay.

今天才周二，本周两大新闻——OpenAI与AWS的380亿美元合作，以及我们今早独家报道的Anthropic财务飙升——已经展现了AI领域此刻的庞大规模。我想请主编Jessica Lesson和贝莱德全球科技板块负责人Tony Kim从更宏观角度解读这些新闻。Jessica、Tony，欢迎二位。很高兴见到你们。

It is only Tuesday and already the two big news stories from the week, OpenAI's $38,000,000,000 deal with AWS and our scoop this morning about Anthropic's booming financials show the sheer scale of this moment in AI. I want to bring on our editor in chief Jessica Lesson and Tony Kim, BlackRock's head of global tech within the firm's equities practice, for a bit of a broader view on what this news all means. Jessica and Tony, it's great to have you. Good to see you.

谢谢，阿卡什。嘿，托尼。

Thanks, Akash. Hey, Tony.

嗨，各位。这很棒

Hi, guys. It's great

能来到这里。

to be here.

所以我对这次讨论很兴奋。杰西卡，看，这周真的非常忙碌。你对OpenAI最近接二连三的交易怎么看？昨天的交易，今早的新闻？你对这一切有什么看法？

So I'm excited for this discussion. Jessica, look, it's been a really busy week. What do you make of of the flurry of the OpenAI deals that have going on, the deal yesterday, the news this morning? What's your view on it all?

你知道，显然OpenAI与AWS的交易非常引人注目。虽然金额不大，OpenAI花费了3800万美元用于计算资源。但这个联盟说实话在两年前——甚至两个月前——我们都没想到会在AI领域看到，对吧？OpenAI原本是和微软合作的。

You know, obviously the OpenAI AWS deal is really notable. It's a teeny amount of money, OpenAI spending, you know, dollars 38,000,000,000 relative to its compute. But it's an alliance that honestly was not clear we were going to see in AI two years ago, maybe two months ago. Right? OpenAI was partnered with Microsoft.

显然，它的计算需求改变了这一点。但看到他们现在与AWS成为合作伙伴，我觉得非常有趣，也让我对这两家公司更广泛的关系产生了很多疑问。我们知道采购是OpenAI的重点任务之一。根据我长期科技报道的经验，交易背后总是有很多谈判。所以我认为这笔交易之所以耐人寻味，在于它可能预示着两家公司未来的走向。

Obviously, its compute needs have changed that. But seeing them as now a partner with AWS, I think, is very interesting and raises a lot of questions for me about the broader relationship between these two companies. We know shopping is a big priority for OpenAI. So I think what I've learned in a long time of tech reporting is there's always a lot of negotiation behind the deal. So I think it's an intriguing deal for what it might portend in the future for the companies.

当你说预示着未来时，这对OpenAI的未来以及它可能寻求达成的交易引发了哪些疑问？

And when you say portend for the future, what questions does raise for you about the future of OpenAI and the deals that it might seek to strike in the future?

嗯，我认为如果从宏观角度看，阿卡什，我们看到的是——我称之为多边主义。这是个非常华丽的词汇，描述的是AI领域以及这些科技巨头之间前所未有的合作态势，这种景象我们已经很久没见过了。OpenAI还与谷歌达成了协议，尽管谷歌云显然是它的主要竞争对手。我们都知道谷歌正在大力推广TPU。

Well, I I think and if you zoom out, Akash, we're seeing I'll I'll call it a multilateralism. That's a very fancy word that's happening in AI and sort of these big tech giants that we haven't seen in a while. OpenAI also has a deal with Google. Clearly, it's chief rival in Google Cloud. We know that Google is pushing TPUs.

而且我持续听到人们对TPU的进展感到非常兴奋，并对其寄予厚望。这是谷歌与英伟达竞争的芯片。突然间，这些原本各守赛道的竞争对手仍在正面交锋，但由于前所未有的算力需求，以及构建下一代AI的现实要求——无论是涉及跨多平台的智能体——都导致了联盟关系的转变。作为一个热爱商业报道的人，我认为通过理解这些公司之间的关系来预判产品走向是很有价值的，这确实是个前所未有的时刻：企业不再固守原有赛道，而是将'亦敌亦友'的关系提升到了全新高度。

And I think and continue to hear that people are very excited about the progress in TPUs and expecting big things for them. This is Google's chip that competes with NVIDIA. And so all of a sudden, these companies that were kind of in their lanes as rivals are still competing head to head, but the unprecedented compute needs and just the realities of building this next layer of AI, whether it involves agents that span multiple sites, is is leading to shifting alliances. And so I I think it's exciting as someone who loves business reporting and and just kinda believes you can learn a lot by understanding the relationships between these companies in terms of what products are coming. It it is kind of an unprecedented moment where people are no longer in their lanes and they're kind of frenemies to a whole new level.

托尼，想听听你的看法。杰西卡关于多边主义的观点中，有个有趣之处在于OpenAI与多家Mag7企业的合作。我们已经看到七家Mag7企业中有六家公布了财报，英伟达即将发布。但我想请教你的是：这如何影响估值？我们都知道这个套路——OpenAI发布合作公告后，相关公司股价就会上涨，进而影响整个市场。在你看来，当前科技巨头的整体估值处于什么阶段？

Tony, want to come to you because one of the interesting points here Jessica is making about multilateralism is the number of Mag7 companies, for example, that OpenAI is doing deals with. We obviously saw that I think six of the seven Mag7 companies have reported now NVIDIA is coming up. But what I want to get your sense on is how this is affecting valuations. We already know the story once OpenAI issues a press release, they've done a deal with a company, we can expect the stocks to come up, and it ends up influencing the market more broadly. But what do you make for this moment for big tech valuations more broadly as you see it?

作为估值风险，我认为情况瞬息万变。杰西卡提到的多边主义现象和巨大算力需求，没有单一实体能完全满足。这导致了亦敌亦友的合作与竞合关系，

As a risk to valuations, think it's constantly changing. Think Jessica said a lot of this multilateralism, the amount of compute that is necessary. No one entity can fulfill it all. So this is causing frenemies and cooperation and coopetition,

而且

And

这种趋势已扩散到所有Mag7+企业，或者说前十的科技公司。最初当这些多年期、数千亿美元规模的合作协议公布时，市场反应剧烈，人们开始消化这些承诺的深层含义。但随着财报季推进，某些公告的反响却不尽相同——有时强烈有时平淡。现在市场正在消化的是：这些巨型承诺究竟意味着什么。

it's diffusing across all the Mac seven plus, you know, in general, the top 10 tech companies. In the beginning, I think when these deals were announced, these multiyear, hundreds of billions of dollar commitments, you you're seeing large stock reactions as people start to digest and decipher what the implications of these commitments are. And then, you know, as we continue going through this earnings season, certain announcements don't have the same kind of reaction. In some cases they do, some cases they don't. And what is now being digested is, okay, well, these are some mega commitments.

接下来需要考量的是这些承诺对资产负债表、现金流、债务等方面的影响。因此市场反应出现分化。我认为现在进入新阶段了——我们必须审视资产负债表和自由现金流能否支撑这些承诺。那些资金储备更雄厚、融资能力更强的公司，往往会比资金紧张的公司获得更积极的市场反应。这就是我对本次财报季市场反应演变过程的观察。

What is then the implications of those commitments in terms of balance sheet, cash flow, debt, etcetera. And so you do see different reactions. And I think it's now we're going to this next level and looking at we have to look at the balance sheets, have to look at the free cash flow to fund these commitments. And in some cases, those companies that have higher firepower, if you will, capacity to fund this might get a better reaction than those that are a little more strained in terms of ability to fund this. So that is kind of how I see the evolution of the reaction for earnings season.

但最初几个月前，你会说这难以置信，前所未有的多年期承诺数量。但现在我们正在重新审视这一点。

But in the beginning, a few months ago, you would say this is unbelievable, unprecedented number of multi year commitments. But now there is a reexamination of that.

杰西卡，托尼在这里提出了一个很好的观点，关于这些公司如何为所有投资融资。我们上周看到资本支出数字在上升，我想知道你是否能帮我们把最新的资本支出热潮放在背景中来看。过去几年我们见过资本支出的增长。你对当前资本支出的时刻怎么看？

Jessica, Tony's making a good point here about how how these companies fund all this investment. We saw the CapEx numbers last week ticking up, and I wondered if you could help us put the the the latest CapEx boom into context. We've seen upticks in CapEx over the past couple of years. What you think of this moment for CapEx?

我认为没有背景可言，尽管托尼在这方面更有权威。但我们确实处于未知领域。显然，市场对Meta上周消息的反应值得注意，因为这不仅仅是继续加大支出。你知道，我们更乐观了，对吧？

I don't think there is a context, although Tony would be much great authority on this. But we're certainly in really uncharted territory. I mean, obviously, the market's reaction to Meta's news last week was notable because it wasn't just the continuation of great spending more. You know, we're more optimistic. Right?

所以这里有些变化，但大方向是投资者某种程度上相信了这个故事。我认为托尼提到的自由现金流是更高层次的问题。我也很好奇，我们很看重托尼对此的看法。我想知道公开市场对这些快速增长但亏损的AI公司的胃口如何。我不是在说，是的，那会很棒或很有趣，从商业新闻角度看，看到OpenAI或Anthropic的IPO会很棒。

So there there there's some shifts there, but but the big picture is that investors are kind of buying the story. I think Tony, you know, bringing up free cash flow is is the next level. I I'm also curious, and and we love Tony's opinion on this. You know, we're I'm curious about the public market's appetite for these fast growing but money losing AI companies. I'm I'm not talking about, you know, yes, it would be awesome and or so interesting, awesome in a journalist business news sense, you know, to see an open AI anthropic IPO.

我不认为它们近在眼前，因为它们还有很多选择。但展望明年，我确实认为，我们的团队不断听到，可能有更多人会测试公开市场，阿卡什，关于钱从哪里来这个问题？但对我来说投资者最终会如何选择仍是个开放性问题。

I I don't think they're around the corner per se because they still have a lot of options. But as we look into next year, I I do think, and and our team continues to hear, you know, more people may be testing the public markets to that point, Akash, of where's this money coming from? But that's still a very open question for me where investors are gonna net out on that.

托尼，接着杰西卡说的继续谈谈。

Tony, pick up on what Jessica's saying here.

当然，绝对可以。让我提供些数据，或许能构建这个框架。杰西卡说得对。我们可能从未见过如此庞大的金额，从未见过这种规模的支出总量，我们现在谈论的是数万亿美元。但占全球GDP的百分比来看，电力时代初期或工业革命时期的支出比例更高。

Sure, absolutely. Let me give some numbers, maybe frame this. And Jessica's right. We might not have seen the quantum of dollars that we've never seen the quantum, the sheer aggregate number of spend, and we're talking trillions now. But as a percentage of global GDP, we've had higher spending in the beginning of electricity or the industrial revolution.

但就目前的美元支出来说，如果看包括大型基础模型公司在内的十大科技公司，它们的EBITDA大约有1万亿美元。它们的净债务基本为零，所以没有杠杆。资产负债表与现金储备大致相当，EBITDA达到1万亿美元。

But in terms of the dollar spend right now, so if you take the 10 biggest tech companies, including the big foundation model companies, they have roughly a trillion dollars of EBITDA. They also have roughly net debt of zero. So they're not levered. The balance sheet and the cash are roughly equivalent. And there's a trillion dollars of EBITDA.

明年在约6000亿或更多资本支出后，还将产生数千亿美元的自由现金流。正如我们所说，有1万亿美元我称之为EBITDA的利润，且没有实际净债务。对比其他行业可承受的债务水平，这些企业能否将净债务与EBITDA比率提升至1倍、2倍甚至3倍？虽然科技公司历来不习惯加杠杆，但面对未来可能数万亿美元的资本支出需求，必要的杠杆化将不可避免。若采用1倍净债务与EBITDA比率，就能新增1万亿美元融资空间——在我看来其中大部分将通过杠杆实现。

And there's a couple $100,000,000,000 of free cash flow after roughly about a $600,000,000,000 or more of CapEx next year. So as we said, there's a trillion dollars of what I call EBITDA and no debt, no effective net debt. When you start looking at other industries and the amount of debt that can be sustained, could you lever these things up one, two, three times net debt to EBITDA? So I don't think generally tech companies have historically not added leverage, but I think there is no doubt that with the amount of the trillions of dollars of CapEx that are potentially coming, there will be leverage that will be necessary. So if you went to one times net debt to EBITDA, you'd be at a trillion dollars of additional capacity, of which most of that will come through leverage, in my opinion.

若采用2倍EBITDA比率，则能达到2万亿美元规模——这正是健康基金即将达到的量级。十大科技公司完全具备这种融资能力。接下来还会产生我所说的下游连锁反应：信用评级最高的巨头企业将主导投资方向。据我观察，这个趋势仍有发展空间。

And then if you did the two times EBITDA, it'd be $2,000,000,000,000 And that's what is coming in terms of the health fund. So there is capacity to fund this from the 10 biggest tech companies. And then there is what I call a cascade downstream, right? You're going to have the biggest spenders with the best credit ratings to drive the spending. From what I see, I still think there's room to run.

正如杰西卡提到的，你们会看到大量数据中心供应商、电力地产等配套产业的扩散，它们将通过自建或租赁方式服务这些巨头企业。届时将出现五花八门的融资方案来支撑这个体系。

Then you're seeing what you're seeing, as I think Jessica mentioned, you're seeing a lot of the diffusion of other data center providers and power real estate and all these things that will then serve these big companies because you're doing some of it, you're building yourself, and sometimes you're leasing to others. And so there's going be a whole myriad of financing options to finance this.

确实。托尼，正如你所说，观察事态发展及市场对潜在债务的反应会非常有趣。感谢二位参与——这位是我们的主编杰西卡·莱森，这位是来自贝莱德的托尼·金。在TI TV的节目中，我们试图探究企业应用AI的实际情况。

Right. Well, think it's certainly an interesting thing to see how it plays out and certainly to see how the market reacts to that debt, as you say, Tony, if it's coming. I want to thank you both for coming on. That is Jessica Lesson, our editor in chief, and Tony Kim from BlackRock here on TI TV. Okay, we've tried on this show to get a sense for the extent to which enterprises are using AI.

通常我们只能从高管那里获得零星的案例数据。但高盛新发布的报告用具体数字表明：虽然很多企业正在使用AI，但其应用场景与大众想象可能大相径庭。现在有请高盛全球高级经济学家、全球经济研究团队联席主管约瑟夫·布里格斯为我们解读报告发现。约瑟夫，欢迎来到CITV。

Typically, the data points we get are anecdotal from executives. But a new report out from Goldman Sachs put some numbers to that problem and shows that while many corporations might be using AI, they might not be using it for the tasks that we think they are. Joining me now to discuss the firm's findings are Joseph Briggs, a senior global economist and co lead of the global economics research team at Goldman Sachs. Joseph, it's great to see you. Welcome into CITV.

很高兴今天能参与节目。

Great to be joining you today.

那么我们来谈谈你们发布的这份报告。对你来说最关键的统计数字是什么？我看到37%的客户在日常生产中应用AI。请帮我们解析这个数据背后的原因。为什么比例没有更高？

So let's talk about this report that you put out. What is the headline number for you? I see 37% of clients using AI in regular production. Help us break down what's behind that. Why isn't it higher?

是的。我们重点关注AI在日常生产中的应用概念，因为我们认为这更能体现全经济范围内的生产力提升。当前美国企业界确实存在大量AI的零散应用案例，但我们认为只有将AI深度整合到日常商品和服务的生产中，才能真正推动劳动生产率和GDP的大幅提升——这正是我们相信可能实现的。我们调查中发现的37%这一数据，基本能代表美国企业日常生产的AI应用水平，远高于人口普查局《商业趋势与展望调查》报告的10%门槛值，也明显超过普查局统计的250人以上企业中14%的应用比例。

Yeah. So we focus on the concept of use of AI for regular production because we think that it's a measure that is more closely aligned with economy wide productivity gains. I mean, there's lot of anecdotal use of AI occurring across The US corporate sphere today, But it's really the close integration into the regular production of goods and services that we think is going be necessary to drive the significant uplifts to labor productivity and GDP that ultimately we think are possible. The 37% number that we found in our survey, which we think is broadly representative of use of regular production across corporate America, is much larger than the 10% threshold that's reported in the Census Bureau's Business Trend and Outlook survey. It's also much higher than the 14% number that the Census Bureau finds for companies with more than two fifty employees.

因此我们的结论是：美国企业实际AI应用程度远超其他数据源反映的水平。

And so our takeaway was that there's a lot of use of AI occurring in corporate America above and beyond what other data sources were telling us.

请简要谈谈不同行业间的应用差异。之后我想深入探讨裁员问题，以及企业使用AI是出于降低成本还是提升效率的不同动机。现在先说说各行业的应用情况，稍后再讨论那些话题。

Talk to me a little bit about how it differs across industries. And then I do want to get into the headcount reductions and the mix between whether people are using it to reduce costs or improve productivity. But talk to us about adoption across industries, and then we'll get there.

确实，科技行业在应用上处于领先地位。我们调查中63%的技术、媒体和电信领域银行家表示，他们的客户目前已将AI用于日常生产。预计这一比例未来三年将升至90%。全经济范围内，调查显示总体应用率预计将在三年内达到74%。显然，科技行业已通过已验证的使用案例推动了AI的快速普及。

Yeah, it's definitely the case that tech is leading the way in adoption. 63% of technology, media, and telecom bankers in our survey reported that their clients were already using AI for regular production today. That number is expected to rise to 90% over the next three years. Economy wide, our survey suggested that overall adoption is expected to increase to 74% over the next three years. And so it's definitely the case that we are seeing tech where we do have proven use cases already driving adoption.

我还想强调，这与宏观经济学数据相吻合。观察科技行业就业占比，已低于长期趋势线。该领域净招聘人数比预期低5%到10%。可见招聘遇冷、技术应用和生产力提升等现象在科技行业确实以更快的速度显现。

I'd also emphasize that this corresponds to what we see in the broader macroeconomic data. If we look at the tech sector's share of employment, it's fallen below its long run trend. We're seeing net hiring in the tech sector underperform by five to 10%. And so the hiring headwinds, adoption, the productivity gains are definitely playing out in a more accelerated fashion in tech.

那么裁员情况如何？有多少比例的企业用AI来缩减人力？你们的发现让我很意外——实际采用这种方式的企业数量比我预期的要少得多。

What about headcount reduction? And what proportion of companies are using it to reduce headcount? I was surprised by what you found. It actually wasn't as many companies as I would have expected.

是的。企业部署人工智能有两种方式，两种用途。一种是提高生产力和收入，另一种是降低成本并减少员工人数。有趣的是，根据我们调查中银行家的反馈，47%的公司目前主要利用AI来提升效率和增加收入。

Yeah. So there's two ways that companies can deploy AI, two uses. One is to increase productivity and revenue. The other is to reduce costs and reduce headcount. What was interesting is if we look at the skew across these two different use cases, 47% of companies in our survey, at least according to our bankers, are mostly using AI to drive efficiency gains and increase revenue today.

仅有20%的受访者表示其客户主要将AI用于削减成本。此外，仅11%的银行家报告其客户实际出现了裁员情况。再次强调这是全经济范围的数据。若聚焦科技行业，报告客户裁员的银行家比例超过30%。但整体信息明确显示：截至目前我们尚未看到AI对就业岗位产生重大影响。

Only 20% reported that they were mostly seeing their clients use AI to cut costs. Furthermore, only 11% of our bankers reported that their clients were actually seeing headcount reductions. Again, this is an economy wide number. If we look at the tech sector, the share of bankers who reported headcount reductions among their clients is above 30%. But the overall message was very much one that we're not seeing significant impacts on headcount as of today.

这种情况在未来几年可能会改变，我认为已有迹象表明这一点，但现在断言就业市场会受到显著影响还为时过早。

Now, that could change over the next few years, and I think there's signs that it will, but it is a little bit too early to see significant employment effects.

我想深入探讨这些就业影响。数据本身说明了一切，但您认为根本原因是什么？是技术尚未成熟？还是实际投资回报未达预期？请帮我们理解背后的原因。

And so I want to zone in on those employment effects. The numbers are what they are, but what is your best guess as to why? Is this just the technology not being good enough? People not seeing the ROI that they thought they would? Help us understand the why here.

没错。我们在调查中专门提出了这个问题。银行家反馈客户仍认为AI技术处于早期阶段。因此企业普遍持观望态度，等待技术和应用生态更加成熟。另一个突出问题是许多公司表示缺乏开发必要AI应用的内部专业能力。

Yeah. We actually asked exactly that question in our survey. And our bankers reported their clients were still viewing AI as a bit too early of a technology. And so I think a little bit of hesitancy to dive in full force until the technology matures and the landscape matures a bit. The other thing that really stood out is that a lot of companies were reporting that they didn't have the in house expertise to develop the necessary AI applications.

实际上我认为这对AI整体应用前景是个利好信号。如果当前障碍仅是技术早期性和应用不完善，那么随着时间推移这些都会改变。我们看到应用层正在爆发大量创新，如今许多企业级解决方案已通过AI实现。随着这些要素成熟，我们将见证更广泛的应用落地和更深远的影响。

I actually think this is a bullish story for the overall AI adoption outlook, because if it is the case that it's just too early and the applications aren't quite there yet, then these things will change over time. We are seeing a lot of activity in the application layer. A lot of enterprise solutions are being solved with AI today. As these dynamics mature, we should see adoption occur and we should see more significant impacts.

基于您对宏观趋势的研判，关于AI与劳动力的问题——如果企业试图通过AI减员或替代人力（如您所言虽未成主流），您如何看待这个局面？许多公司视此为获取投资回报的明确途径。对于那些可能被自动化取代的岗位人员，您认为他们将何去何从？

And talk to me about, given your macro view on all this, where do you sit on the labor question in AI if people are looking to reduce headcount or maybe replace workers with AI? Like you said, it's not happening right now, but I think a lot of companies are saying, well, that is one clear way to get ROI. What is your sense for what all these people will do they're not doing the jobs that conceivably can be automated?

是的。所以两方面都会有一点。会出现一定程度的劳动力替代，即因AI而消失的工作岗位。同时也会有大量劳动力增强，因为工人无需再做那些目前必须完成却又非其核心职能的繁琐任务，从而提升效率。我们在八月份撰写过一份关于劳动力替代问题的报告。

Yeah. So there will be a little bit of both. There will be some amount of labor displacement, jobs that are displaced because of AI. There's also going to be a lot of labor augmentation where workers become more efficient because they don't have to do some of the more tedious tasks that they currently have to do and aren't necessarily the main function of their jobs. We wrote a report in August where we looked into the labor displacement question.

从历史数据来看，我们发现：当整体劳动生产率提升15%（这是我们预期AI全面普及后对整体经济的影响幅度），对应的劳动力替代率约为5.6%至7%。如果这一过程如预期在十年内逐步发生，对劳动力市场的冲击应该不会太大——年均失业率增幅约0.5个百分点。真正的风险在于，如果AI应用进程过于集中在前端，失业率上升幅度会更大，对劳动力市场的影响也将更显著。但总体而言，基于现有证据，尤其是考虑到经济层面的渐进式普及，我们对劳动力市场的平稳过渡仍持乐观态度。

And if we look at historical evidence, what we came up with was that for our 15% increase in overall labor productivity, which is what we're expecting on an economy wide basis following full adoption of AI, This should correspond to roughly a 56%, 7% of labor displacement. If this happens over a ten year period the way that we're expecting, it's probably not going to be that disruptive to the labor market. You're looking at a half point boost to the unemployment rate in any given year. The big risk is that if we do see adoption occur in a much more front loaded manner, then the rise in the unemployment rate would be larger and the impacts on labor markets would be more significant. For the most part, we still remain fairly optimistic based on the evidence that we're seeing, and especially on an economy wide basis, a relatively slow adoption, that the labor market transition will be manageable.

太好了。约瑟夫，感谢你参与节目。很高兴能邀请到你。

Great. Well, Joseph, I want to thank you coming on. It's great to have you here on the show.

很荣幸参加。

Great to join you.

好的。今天我们正式发布《信息》2025年度最具潜力初创企业榜单TI50。长期订阅用户都知道，我们每年都会精选50家最被看好的跨领域企业，涵盖AI、加密技术、能源、消费、电商等行业。这些未必是估值最高的公司——入选前提条件包括估值低于10亿美元且年收入不足1亿美元。

Okay. Today, we are launching our 2025 edition of the information's 50 most promising startups, also known as TI50. And if you're a longtime subscriber, you'll know that every year we put together this list, which consists of the 50 companies that our sources are most bullish on across all sectors, including AI, crypto, energy, consumer, e commerce, and more. Now, these aren't necessarily the companies with the highest valuations. In fact, one of the precursors to being on this list is that companies should be valued at less than $1,000,000,000 and also need to be generating less than $100,000,000 in annual revenue.

举个例子，往年上榜企业中不乏后来成为行业巨头的名字：2021年我们选中了AI爆发前的Hugging Face，当时估值45亿美元；2022年入选的Tabular后来被Databricks以20亿美元收购；2023年入选的Perplexity近期正以200亿美元估值敲定融资轮。直白地说，这份榜单有着惊人的精准度。本周我们将每天重点介绍2025榜单中不同领域的代表企业。

But for context, some companies on this list in years past have gone on to become tremendous companies, names that you might recognize. In 2021, we picked Hugging Face before the AI boom. The company was reportedly then valued at $4,500,000,000 two years ago. In 2022, we picked Tabular, a company that was then acquired by Databricks later for a reported $2,000,000,000 And in 2023, we picked Perplexity, which we reported weeks ago was finalizing commitments for a funding round, valuing it at $20,000,000,000 To put it bluntly, this list has an extraordinary track record. And so every day this week, we are going to be featuring one company from our 2025 edition of the list across several of our categories.

今天登场的是Replt公司。这家由安德森·霍洛维茨和红杉等顶级机构投资的企业，正致力于打造未来原生的AI企业资源规划工具。有请CEO尼古拉斯·科普为我们详解他的创业蓝图及ERP系统的本质。尼克，欢迎来到节目。

Today, the company that we're bringing on the show is Replt. The company is backed by big names like Andreessen Horowitz and Sequoia, and is trying to build the AI native enterprise resource planning tool of the future. Want I to bring on CEO Nicholas Kopp, tell us more about what he's building and what ERPs really are. Nick, welcome to the show. It's great to have you.

那么我们来谈谈Rilla。恭喜你上榜了，进入了前50名。请介绍一下你们的工作内容，以及你们如何从整体上优化ERP系统。

So let's talk about Rilla. Congratulations. You're on the list. You made the top 50. Tell us about what you do and how you're looking to overall the ERP system broadly.

是的。我们正在构建一个原生AI的ERP系统，主要为增长最快的公司提供核心财务运营支持，包括Relit上许多即将IPO的企业。对于外行来说，ERP本质上是一个将所有财务会计信息整合到一个平台的系统。它帮助会计师和财务专业人员核算过去发生的费用、收入等各类事项，同时也为预测和规划奠定基础，这在临近2026年的当下尤为重要。

Yeah. So we are building an AI native ERP, which essentially powers core financial operations for the fastest growing companies out there, including a lot of pre IPO companies on Relit. An ERP for the layman maybe is basically a system that strings together all financial accounting information into one platform. It helps accountants, finance professionals account for what's happened in the past, including expenses, revenue, all these types of things, as well as helps build the foundation for forecasting and planning, which is particularly relevant at this time of year, given it's 2026 soon.

好的。那么谈谈目前的业务表现吧。据报道，你们的年化收入已达到1000万美元。这些收入主要来自哪些方面？

Right. So let's talk about the business performance right now. We've reported that you guys are doing $10,000,000 on annualized revenue. Where is that coming from?

是来自大公司还是初创企业？目前我们主要服务的是中等规模的初创企业和中小企业市场。比如最近因Decacon融资新闻而备受关注的Mercor，他们是最快达到5亿美元ARR的公司之一，即将突破10亿美元大关。

Is that big companies? Is that startups? Yeah. So we are serving today the mid market of startups, lower enterprise segment. So that is companies like Mercor, who was recently in the news with their recent Decacon, a fundraise, a fastest growing company to get to half a billion dollar of ARR out there, on track to hit a billion here shortly.

这类公司就是Rilla服务的典型案例。我们还为Windsurf整个收购过程提供了支持。我们服务的客户主要是软件公司、专业服务机构和正在使用Unreal而非传统系统的金融科技企业——那些传统系统你可能听说过，比如NetSuite、Workday和Sage Intacct。

And that's sort of a great example of a company on Rilla. We were also famously powering Windsurf throughout all their acquisition situations. And so that's sort of the type of company we power. A lot of software businesses, professional services, fintech that is running Unreal today, instead of The legacy systems are called sort of NetSuite, Workday, Sage Intacct,

与这些知名ERP巨头相比，你们的竞争优势是什么？

that you may have heard of. And what is the value prop against using some of these bigger ERP giants that we've heard of?

有几个方面。这取决于使用团队的具体需求。现有系统缺乏某些功能——我们利用AI自动化了许多会计人员在月末进行的繁琐手工操作，比如对账和记录交易。

Yep. So it's a few things. It depends also on the team that is using it. There is a functionality that current systems don't have. So we use AI to automate a lot of the tedious manual tasks that accountants do at month end to reconcile their financial statements, to book transactions.

这极大地减少了我们Relit团队结账和生成财务报表所需的时间，这是一个实实在在的优势。同时也提升了报告质量。您在我们应用中获得的精细化财务数据，在其他许多系统中是难以获取的。最后同样重要的是，实施过程会快得多。

And that just reduces massively the time spent for our teams in Relit to close their books and produce these financials. So that is sort of a hard benefit there. It increases the quality of the reporting as well. So the granular financial information that you get in our application, you can't get out of a lot of other systems out there. And then lastly, and not least actually, is implementations are much faster and quicker.

您可能听说过Rillet那些大型传统厂商需要六到十二个月实施的恐怖故事。而我们能让这个过程对您来说轻松得多——

With Rillet, you will maybe be familiar with some six to twelve months implementation horror stories from the big guys, legacy guys out there. So, we make that process a lot easier for you to

轻松多少？用你们公司方案实施需要多长时间？

offer. A lot easier, meaning how long does it take then to implement it with your company?

好问题。平均需要四到八周。四到——

Yep, great question. Four to eight weeks is the average. Four to

八周？也就是说我能在四到八周内获得端到端的ERP会计软件和全套财务系统？整个转型只需要这么久？

eight weeks, so I can get an end to end ERP accounting software, all my finance systems, four to eight weeks. That's all the transformation will take.

没错。难以置信对吧？是的，这太棒了。

Yep. It's unbelievable. Okay. Yes. It's awesome.

能说说为什么您对这个课题如此热忱吗？是什么让您愿意毕生致力于解决这个问题？

And talk to me about why you're so passionate about why is this problem the thing that you have dedicated your life to?

是啊是啊，没人一觉醒来就想开发ERP系统或会计软件。

Yeah, yeah, yeah. Nobody jumps out of bed wants to build an ERP or accounting software.

显然你是例外。

Apparently you do.

其实不是。我认为推动我们的主要有两点。第一，这里存在一个亟待解决的大问题。对于不在CFO办公室工作的人来说，很难相信这些流程和系统至今仍如此落后且依赖人工操作——它们实际上已有二三十年的历史了。

Yeah, no. So I think there's two things that really drive us. So A, there is a huge problem that can be solved here. I think it's hard to believe for people not operating in the office of the CFO, how backward and manual a lot of these processes and systems still are. They're literally 20 to 30 years old.

我不确定你Akash上次使用这么古老的软件是什么时候。这些系统严重拖累效率，而我们能对众多企业和个人的决策产生深远影响。这是其一。其二，从个人角度说，我们对自己做的事都充满极客精神——一群会计金融怪才与顶尖工程师共同打造会计软件。

So, I'm not sure when you Akash have last used a software that's that old. And it really drags people down and you can have a meaningful impact on so many people's lives and companies and their decision making. So that's one. And then two, just on a more personal note, we're all like nerds about what we're doing. So a bunch of accounting and finance geeks building accounting software with the best engineers.

就个人而言，这让我感到无比充实。我们的产品开发团队包括注册会计师、财务总监和审计师，每个人都有自己故事：可能是系统上线毁了孩子的生日派对，或是让他们连续加班数周。因此我们都致力于从根本上解决这个问题。

And so on a personal level, it's been extremely fulfilling. Building this product, we have a team of CPAs, controllers, auditors that helps build that product. And everybody has their own little story as to how maybe a system implementation or a system usage has ruined the personal birthday party of a kid, or has personally put them through weekends and weekends of work. So we're all very passionate about solving this problem at its core.

太棒了。Nick，感谢你参与节目，恭喜上榜，期待很快再见。感谢！今天的节目就到这里。

Great. Well, Nick, I want to thank you for coming on. Congrats on making the list, and we'll hope to see you soon. Thank you. Well, that does it for today's show.

提醒大家，我们每周一到周五太平洋时间上午10点/东部时间下午1点在此直播。感谢本节目冠名赞助商亚马逊云科技，也感谢各位观众的收看。我们非常重视您的关注，已经迫不及待期待明天的节目了。

A reminder, we are on this stream Monday through Friday at 10AM Pacific, 1PM Eastern. I want to thank Amazon Web Services, who is our presenting sponsor for this production. And I want to thank you for tuning in. We really do appreciate your viewership. I'm already excited for our next show tomorrow.

祝你周二剩下的时间愉快。现在先再见啦。

Have a great rest of your Tuesday. Bye bye for now.