Lyft股价逆转，Snowflake的AI代理策略，电动汽车充电网络建设 | 2025年11月13日

欢迎各位收看Information的TITV。

Welcome everyone to the Information's TITV.

我叫阿卡什·巴什里查。

My name is Akash Basricha.

今天是11月13日，星期四。

It is Thursday, November 13.

今天我们为大家准备了一场精彩的节目。

We have got a great show lined up for you today.

首先，今年Lyft的股价实际上表现优于Uber的股价。

First up, Lyft shares are actually outperforming Uber shares this year.

我们的财经分析专栏作家将做客节目，探讨Lyft业绩逆转背后的因素。

Our financial analysis columnist joins us on the show to talk about the factors behind Lyft's turnaround.

我们还安排了一场与Snowflake高管关于AI代理的精彩对话。

We've also got a great conversation about AI agents lined up for you with a Snowflake executive.

我们还将与Loop Global的首席执行官讨论电动汽车充电基础设施的现状。

And we're talking about the state of electric vehicles charging infrastructure with the CEO of Loop Global.

最后，我们将带来一家入选今年TI50榜单（我们评选出的50家最具前景初创公司）的初创公司的融资更新。

And finally, we'll finish things off with a funding update from a startup on this year's TI50, our list of the 50 most promising startups.

这是一期内容丰富的节目，让我们马上开始。

It is a big show, and so let's get right on into things.

优步的市值几乎是Lyft的20倍。

Uber's market cap is nearly 20 times more than that of Lyft's.

但一个可能让你感兴趣的小众数据是：今年Lyft的股价上涨了近80%，而优步仅上涨了50%。

But a little known stat that may interest you, Lyft shares are up almost 80% this year compared to Uber's 50% jump.

两家公司显然都表现良好。

Both companies are clearly doing well.

但今天，我的同事阿尼塔·拉马斯瓦米发表了一篇专栏文章，详细分析了Lyft在过去几年中是如何实现转型的，以及这对其未来意味着什么。

But today, my colleague Anita Ramaswamy published a column breaking down how exactly Lyft has turned itself around over the past few years and what it could mean for the future of the company.

这是一篇很棒的文章，所以我请阿尼塔来和我们深入聊聊。

It's a great column, so I want to bring on Anita to talk all about it.

阿尼塔，欢迎再次做客我们的节目。

Anita, welcome back to the show.

很高兴你来这里。

It's great to have you here.

嗨，科什。

Hi, Kosh.

很高兴能来这里。

Great to be here.

那我们来谈谈Lyft吧。

So let's talk about Lyft.

为什么Lyft的股价今年表现这么好？

Why are Lyft shares doing so well this year?

是的，这对投资者来说有点出乎意料，对更广泛的市场也是一种警示，因为Uber要大得多。

Yeah, so this is a little bit surprising to investors, a caution to the market more broadly, because Uber is a lot bigger.

它一直被视为这个领域的主要竞争对手，许多投资者认为这是一个赢家通吃的市场。

It's always been seen as sort of the number one competitor in this space, and many investors have a thesis that it's sort of a winner takes all space.

Lyft今年之所以能突破这种叙事，有几个原因。

Lyft really broke out of that narrative this year for a couple of reasons.

他们在2023年迎来了新CEO大卫·里彻，他来自亚马逊，并在Lyft实施了各种变革。

They got a new CEO in 2023, David Richer, and he came from Amazon and he's been implementing all sorts of changes at Lyft.

在我看来，他迄今为止取得的最重要成果是成功将公司转变为一家自由现金流为正的公司。

And the biggest result that he's delivered so far, in my opinion, is that he was able to turn the company into a company that is free cash flow positive.

因此，他削减了成本。

So he's cut costs.

他还利用资产负债表上剩余的多余现金进行了一些收购。

He has actually used some of the excess cash left out on its balance sheet to do some acquisitions.

因此，他们今年收购了一家豪华专车服务公司。

So they bought a luxury chauffeuring service this year.

他们收购了一家欧洲出租车应用，这标志着他们迄今为止最大规模的国际扩张，也是他们许久以来首次在北美以外地区开展的举措。

They bought a European taxi app, which kind of marks their biggest international expansion and their first initiative outside of North America in a while and ever.

所以，这是另一个原因。

So that's one other reason.

此外，他还清理了资本结构。

And he's also cleaned up the cap table.

Lyft创始人曾拥有双重股权结构，这引发了巨大争议，而他成功与一位积极投资者合作，解决了这一问题。

There was a big controversy around the dual class share structure that Lyft's founders had, And he managed to work with an activist investor and resolve that issue as well.

他确实与积极投资者合作了。

He actually worked with activist investors.

很多CEO有时会害怕积极投资者。

A lot of CEOs sometimes, they get scared of activist investors.

他们说：看好了，让我自己来处理。

They say, Look, let do my thing over here.

这很有趣。

It's funny.

整个积极投资者的行动只持续了几周，是我见过的最短的积极投资者运动之一。

It was like a couple of weeks of an activist campaign, one of the shortest activist campaigns that I've ever seen.

结果是Lyft扩大了股票回购计划，并解决了股权结构问题。

And it resulted in Lyft upping its share buyback program, making that bigger, and also resolving this cap table issue.

但看到整个过程不到一个月就结束了，确实挺有趣的。

But it was just sort of funny to see that it was less than a month long.

大卫·里彻是个有趣的人。

David Richer, he's an interesting guy.

他做了很多播客，不得不说。

He does a lot of podcasts, have to say.

我不知道你听过多少他的节目，但他确实深入探讨了Lyft和Uber业务的对比。

I don't know if you how many of them you've listened to, but he's So really done the let's talk about how Lyft compares to Uber's business.

这两家公司曾经是竞争对手。

These are two companies that at one point were rivals.

现在Uber显然大得多。

Now Uber is obviously much bigger.

但Lyft经营业务的方式与Uber有何不同或相似之处？

But how does the way that Lyft has been running their business contrast or compare with that of Uber's?

我认为最大的区别是，过去Uber是国际化的，而Lyft不是。

I would say the biggest difference, Akash, was in the past Uber was international, Lyft wasn't.

但这一点已经开始发生变化。

So that's one thing that has started to change.

如今最大的区别实际上是Uber Eats业务部门。

The biggest difference today is actually the Uber Eats division.

Lyft专注于单一的叫车服务。

Lyft has focused singularly on ride hailing.

大卫·里彻明确表示，他目前只专注于这一市场，没有计划像Uber那样进入食品配送或杂货配送等领域。

David Rischer has been crystal clear that that is the only market that he's focused on right now and doesn't have plans to expand into anything like food delivery or grocery delivery in the same way that Uber has.

因此，尽管Uber规模大得多，收入是其八倍，现金流是其九倍，并且在过去几个季度中增长稍快，这是因为Uber Eats增长得非常不错。

So while Uber is a lot bigger, it's got eight times the revenue, generates nine times the cash, it's also grown a bit faster in the last couple of quarters because Uber Eats has been growing really nicely.

Lyft的优势在于，大卫·里彻及其团队只专注于叫车服务，所有的收购也都集中于此。

The advantage with Lyft is that David Rischer and the team there are focused only on ride hailing, and that's where they've done acquisitions.

他们所有的增长计划和新功能也都聚焦于此。

That's where they have focused all of their growth initiatives and new features as well.

那么，叫车行业整体面临哪些威胁呢？

Now, what about the threats that face the ride hailing sector at large?

显而易见的隐患是自动驾驶汽车。

The elephant in the room is autonomous vehicles.

这周对Waymo来说是重要的一周，因为他们宣布其汽车将上高速公路。

It's been a big week for Waymo announcing their cars will be on highways.

Lyft有没有谈到过他们如何应对这一范式转变，以及如何抓住这一趋势？

Has Lyft talked at all about the ways in which it's approaching that paradigm shift and the ways in which it can actually latch onto that trend?

是的，自动驾驶对Uber和Lyft都构成了不同程度的重大风险。

Yeah, this is a big of risk looming over both Uber and Lyft in different ways.

Uber应对这一问题的方式是与多家自动驾驶公司达成合作交易。

And the way that Uber had addressed it is they started doing all these partnership deals with different AV firms.

因此，他们在奥斯汀、亚特兰大和凤凰城与Waymo达成了合作。

So they did deals with Waymo in Austin, Atlanta, and Phoenix.

他们还与其他多家自动驾驶领域的不同参与者建立了合作关系，覆盖了自动驾驶出租车栈的各个部分。

They have partnerships with a bunch of other different AV players that are on different parts of that autonomous robotaxi stack.

Lyft早期在建立这些合作关系方面不太成功，但最近几个月情况开始好转。

Lyft was less successful in making those partnerships early on, but in the last couple of months or so, they have started to turn that around.

其中一个关键原因是，他们与Waymo在纳什维尔达成了一项协议，计划于2026年将自动驾驶出租车引入Lyft网络。

One of the key reasons is they are doing a deal with Waymo in Nashville to bring robotaxis to the Lyft network there in 2026.

Lyft 方法的独特之处在于，他们实际上拥有一家名为 FlexDrive 的车队管理公司，该公司是他们在 2020 年收购的，用于为其车队提供维修、清洁和维护服务。

And what's unique about Lyft's approach is that they actually own a fleet management company called FlexDrive, something that they bought in 2020 to help service, clean, and maintain their fleet of cars.

他们将把这家子公司及其能力应用到自动驾驶车辆上。

They are going to be applying that subsidiary and its capabilities to autonomous vehicles.

因此，必须谨慎对待自动驾驶汽车，确保你的出租车始终处于使用状态。

So it's really important to caution AVs to make sure that your taxis are being used all the time.

在常规的网约车业务中，这一点也同样重要。

It's important in regular ride hailing too.

为了提高利用率，确保车辆能快速周转、快速清洁并迅速重返道路，我认为这支名为 FlexDrive 的车队管理服务对 Lyft 至关重要。

And in order to boost that utilization and make sure the cars are being turned around quickly, cleaned quickly, getting back on the roads quickly, that fleet management service FlexDrive is going to be really crucial for Lyft, in my opinion.

所以，如果我理解正确的话，你的观点是 Lyft 在应对自动驾驶挑战方面比 Uber 略胜一筹。

And so your view actually is that Lyft is approaching the autonomous vehicles challenge a little bit better than Uber, if I'm understanding you correctly.

没错。

Yeah, that's right.

我的意思是，Uber 确实非常成功，但我认为 Lyft 拥有这一环节才是真正的差异化优势。

I mean, has been tremendously successful, but I think it is a true differentiator that Lyft owns this part of the stack.

如果你思考一下自动驾驶出租车所涉及的所有不同组件，并假设你是一位投资者，相信这些技术确实是未来且将获得大众市场的广泛接受——比如我们昨天刚看到Waymo在旧金山湾区扩展了业务。

If you think about all the different components that go into robo taxis, and if you assume, as if you're an investor who believes that these are really the future and they're going to gain mass market popularity, I mean, we just saw yesterday that Waymo expanded in the SF Bay area.

这个商业模式有很多早期的验证点，表明确实存在真实的需求。

There's a lot of early points of validation to the business model and that there is real demand.

因此，目前优步的定位是一个平台，可以为这些已有的自动驾驶车辆带来需求。

So, you know, at that point, Uber is positioned right now as a platform that can just bring demand to those already existing AVs.

如果像Waymo这样的自动驾驶公司能够自行获取需求，那么对我来说，优步并没有展现出独特且差异化的价值。

If the AV companies like Waymo can get demand on their own, then it's not clear to me where Uber is doing something unique and differentiated.

Lyft已经内部构建了这一能力。

Lyft has built that capability out in house.

优步随时可以收购一家从事车队管理的公司。

Uber can always acquire a company that does fleet management.

他们总可以通过并购或其他方式进入这一产业链。

They can always do some M and A and find some other way into the stack.

但至少目前来看，Lyft正在从零开始自主建设这一能力，并且自2020年以来一直在持续投入。

But at least for right now, Lyft is building this capability and has been working on it since 2020 from the ground up.

我认为，这将使他们在与其他自动驾驶公司（而不仅仅是Waymo）或在其他城市达成交易时占据优势。

And I think that's going to be an advantage for them when it comes to making deals with other sorts of AV firms besides just Waymo or deals in other cities.

你刚才谈到了并购和收购。

Now, you talked about M and A, you talked about acquisitions.

我们倒回时间线。

We rewind the clock.

去年年底，你预测亚马逊会收购Lyft，但这件事至今尚未发生。

At the end of last year, you made a prediction that Amazon was going to buy Lyft and hasn't happened yet.

我们并不是说这件事不可能发生。

I'm not we're not saying it it couldn't happen.

那你现在对此持什么看法？

But where do you stand on that?

你是否在提高你去年的预测？

Are are you re upping that prediction that you made last year?

你对Lyft的未来有何预测？

What is your prediction for Lyft going forward?

还不完全是，阿卡什。

Not quite Akash.

我想有时候水晶球并不能完美预测。

I guess sometimes the crystal ball doesn't work perfectly.

我认为那里的理由仍然成立且合理。

The rationale there, I think, still holds up and makes sense.

就 Lyft 和 Uber 而言，它们都有大量的需求和新的乘客，每个平台都有数百万乘客。

So far as both Lyft and Uber have a lot of demand and have new ridership, millions of riders on each of the platforms.

我认为在自动驾驶汽车竞赛中，我们仍然处于非常早期的阶段。

And I think when it comes to the autonomous vehicle race, we are still very early.

Waymo 看起来是目前的赢家，但我们所有人都记得 GM 的 Cruise 发生了什么。

Waymo looks like the winner right now, but we all remember what happened with GM's Cruise.

我们都记得那么多公司最终无法向前推进。

We all remember so many of these different companies that were not able to move forward.

我认为竞争格局将迅速发生变化。

And I think that the field of players is going to shift really rapidly.

所以我认为，随着技术的发展，Lyft 有可能收购其他初创公司，或者被其他公司收购。

So I think it's possible that Lyft is either acquired or an acquirer of more of these startups as the technology develops.

很好。

Great.

安妮塔，这是一篇很棒的文章。

Well, Anita, it was a great column.

非常感谢你来参加我们的节目。

Thank you so much for coming on the show.

我很感激。

I appreciate it.

我还想说，大卫·里彻，如果你在看这个节目，我知道你正在参加各种播客活动，欢迎来我们的节目，因为我们显然在密切关注你们公司。

And I should say, David Richer, if you're watching this, I know you're on the podcast circuit, come on our show because clearly we're watching your company very closely.

但与此同时，感谢安妮塔的到来。

But in the meantime, thank you Anita for coming on.

非常感谢。

Really appreciate it.

谢谢。

Thanks.

好的。

Okay.

我们昨天谈到了人工智能代理在企业软件公司中的普及。

We talked yesterday about the proliferation of AI agents at enterprise software companies.

Snowflake 是这一趋势的核心公司之一。

Snowflake is one company that sits at the center of that trend.

本月，该公司推出了 Snowflake Intelligence，其中包含一组面向企业的新型代理。

This month, the company launched Snowflake Intelligence, which houses another group of agents aimed at businesses.

现在加入我们的是 Snowflake 的 Cortex AI 代理负责人杰夫·霍兰德，他将为我们详细介绍这一产品的愿景。

Joining me now is Jeff Holland, Head of Cortex AI Agents at Snowflake, to tell us more about his vision for the offering.

杰夫，欢迎来到 TI TV。

Jeff, welcome to TI TV.

很高兴你来到这里。

It's great to have you here.

嗯。

Yeah.

非常感谢邀请我。

Thanks so much for having me.

这是一份荣誉。

It's an honor.

那么我们来谈谈 Snowflake Intelligence。

So let's talk about Snowflake Intelligence.

它是什么？

What is it?

嗯。

Yeah.

Snowflake Intelligence 是我们的一项努力，旨在帮助每个组织更轻松地让所有人达成一致。

So Snowflake Intelligence is our effort to help it so that every organization can more easily get everybody on the same page.

我们经常发现，在许多公司里，这种情况无处不在。

We're often finding across many companies, like, is everywhere.

有大量数据分散在Slack对话和文档等各种地方。

There's so much data scattered across Slack conversations and documents and you name it.

但要真正获得洞察力，了解我们的收入趋势如何，销售漏斗状况如何，可能会非常困难。

But actually having the ability to gain insights, understand what's happening of how are we trending towards our revenue, what is my sales pipeline looking like, can be really difficult.

通常，你只能通过查看静态或过时的仪表板来查找信息，或者打电话给你的分析团队，希望他们进行手动分析。

Often, you're stuck by looking at something like digging through static or stale dashboards, or you're calling your analyst team and you're hoping for them to do manual analysis.

我们希望赋予组织中每个人——无论你是CEO，还是刚加入一周的新员工——都能即时访问整个组织的所有数据，从而做出更明智、更智能的决策。

We want to empower every individual in an organization, whether you're the CEO or you just joined a week ago, to have instant access to all of the data across your organization to make better informed, intelligent decisions.

这一切都由后台的智能代理驱动，这些代理通过Snowflake连接到你的数据。

And this is all powered by agents behind the scenes connected to your data through Snowflake.

你们的定价模式是怎样的？

And how are you pricing it?

是的。

Yeah.

Snowflake长期以来一直以非常慷慨的定价模式著称，我们采用纯按使用量计费的方式，这意味着只有当客户真正获得价值时，我们才会确认收入。

So Snowflake has actually been known for a while of a very generous pricing model, which is we are pure consumption based, which means we do not recognize revenue unless our customers recognize value.

我们实际上采用了同样的方法，这在Snowflake Intelligence中非常独特，意味着你可以将Snowflake Intelligence部署到整个组织。

We actually took the same approach, which is quite differentiated in Snowflake Intelligence, which means you could roll out something like Snowflake Intelligence to your entire organization.

只有当这些代理实际工作时，你才需要付费。

You only pay when those agents are actually working.

所以如果没有人提问，你在Snowflake上就不会产生任何费用。

So if nobody's asking a question, there's no cost for you in Snowflake.

你只需为那些在客户真正获得价值时发生的令牌和LLM调用付费。

You're just paying for those tokens, those LLM calls that are happening only when customers are actually finding value.

我很好奇。

I'm curious.

你们是完全自主研发的，还是整合了最近收购的某些技术？

Did you build this all in house or did it come from any of the technology that you've acquired recently?

这其实是几方面因素的结合。

So it's a combination of a few things.

Snowflake在底层大语言模型方面有几个关键合作伙伴。

We Snowflake has a few key partnerships in terms of underlying large language models themselves.

因此，Snowflake 可以安全地运行 OpenAI 的 GPT-5 或 Anthropics Cloud 4.5 等模型。

So Snowflake securely can run things like OpenAI's GPT five or Anthropix Cloud 4.5.

这些模型作为底层模型。

Those acts as the underlying models.

在准确查找并检索向用户提供的正确数据方面，实际上发生了一些收购。

In terms of the ability to accurately find and retrieve the right data to give to the user, there's actually a few acquisitions that have happened.

Neva 几年前就已收购，拥有核心搜索技术。

Neva was a few years ago at this point that has some of the core search technologies.

TruEra 的收购则更近一些。

TruEra was a little bit more recent.

它主要涉及这些智能代理组件的质量。

That's around quality of some of these agentic pieces.

此外，还有很多顶层的附加功能，我们如何让代理真正擅长理解复杂庞大的数据集？

And then a lot of the additional layers on top, how do we make it so agents are really, really good at understanding the complex massive datasets?

我的意思是，想想 Snowflake。

I mean, think of Snowflake.

这些都是全球2000强企业。

These are global 2,000 companies.

这些都是大型组织。

These are huge organizations.

因此，在过去一年到一年半的时间里，我们内部开发了许多功能，以确保这些代理在处理海量数据时依然非常有效。

So we have built a lot of things in house over the last year, year and a half to make it so that these agents are really effective even when looking at massive datasets.

你提到使用了OpenAI和Anthropic的模型。

You talked about using models from OpenAI and Anthropic.

我们在这个节目中多次讨论过，使用这些模型成本高昂，不仅限于这两家公司，而是所有模型都价格不菲。

We've talked a lot on this show about how expensive it is to use some of these models, not just that those two companies, but models at large are pricey.

这些代理目前对你来说盈利吗？

Are these agents profitable for you right now?

我的意思是，我们不会披露这些部分的底层利润率。

I mean, we're we don't disclose anything in terms of, like, our underlying margins for these pieces.

我想说，我们与这些供应商都达成了非常有利的协议，确保我们在成本方面有良好的优势。

I would say that we've got really good agreements with each of these providers in terms of making sure that we can get good economics on our side.

然后我们会尽一切努力。

And then we do everything we can.

说实话，我负责产品，我们的目标是尽可能减少我们底层用户的使用障碍。

Like, our goal truthfully, I'm on the product side, is we wanna reduce as much friction as possible for our underlying users.

所以我们并不是在那儿斤斤计较地收费。

So we're not out here trying to nickel and dime people.

正如你所指出的，这些大型语言模型确实有实际的成本。

We truly are like there's a real cost to your point of these large language models.

这些模型确实需要强大的GPU计算资源，但我们对与这些提供商以及后台云基础设施达成的协议非常满意。

These things do require some real beefy heavy GPU compute, but but we're really happy with the agreements that we have with these providers, with the cloud infrastructure that we're running behind the scenes.

我们正努力将这些优势尽可能多地传递给客户，这也是我刚才提到的，我们实际上提供的是按使用量计费的模式，而不是通常AI代理技术所采用的固定按席位收费模式。

We're trying to push as many of those benefits to our customers as we can, which is part of why I mentioned a bit ago too, we're actually offering this as a consumption based offering, not requiring you have some fixed per seat costs, which is usually the norm with some of these AI agent technologies.

我想稍微宏观一点，来看看整个企业软件行业。

I want to zoom out a little bit and look at the enterprise software sector at large.

我们昨天刚邀请了企业软件记者亚伦·霍姆斯做客。

We had our enterprise software reporter on just yesterday, Aaron Holmes.

他写了一篇文章，讲述企业软件公司推出的各类智能代理产品正开始趋同，彼此之间的竞争激烈程度是我们多年未见的。

He wrote a story about how all these agent offerings at enterprise software companies are really starting to converge, and they're starting to step on each other's toes in a way that we actually haven't seen in quite a few years.

当你思考这一点时，你觉得Snowflake Intelligence如何与众多看似使用与你刚才所述相似语言的企业软件公司区分开来？

As you think about that, how do you think about how Snowflake Intelligence differentiates itself from the dozens of other enterprise software companies that seem to be offering something very similar to the language that you just used.

是的。

Yeah.

当我与组织讨论这个问题时，我经常想强调的是，在Snowflake，我们并不打算打造一个能做所有事情的代理。

And I the thing that I always often like to talk when I'm talking to organizations about this is Argo at Snowflake, we're not trying to build an agent that can do everything.

当然，目前围绕智能代理确实存在巨大的热情，这种热情完全可以理解。

And, of course, like, there's so much excitement, real understandable excitement around agents right now.

正如你所说，如今许多科技公司都在推出自己的代理产品。

To your point, so many technology providers right now are creating their own agent offerings.

我们的立场是说，嘿。

Our position is saying, hey.

看。

Look.

这不会是一个能为你做所有事情的代理。

This isn't gonna be the agent that does everything for you.

我们想要真正简化并提升的是那些能够查看并理解你数据的代理。

What we wanna make really easy and really effective are specifically agents that can look into and understand your data.

数据是Snowflake的血液。

Like, data is the blood of Snowflake.

这是我们的核心。

That's our core.

这是我们真正非常擅长的事情。

That's the thing that we are really, really good at.

因此，我看待未来发展的方向是，你会在组织中看到这些代理的组合。

So the way that I see things moving forward is that you will see these combination of agents in an organization.

但我们正朝着这样一个世界前进：这些代理将更像一个合奏团或管弦乐队一样协同工作，当你完成一项任务时，也许任务的一部分是需要通过你的招聘门户打开一个职位招聘。

But we're moving towards more of this world where those agents are gonna act more as an ensemble and orchestra together, where when you're accomplishing a task, maybe part of your task is you need to go open up a job rec using your job portal.

而Snowflake代理可能并不是最适合完成打开职位招聘这类任务的。

Well, a Snowflake agent's probably not gonna be the best one at doing things like opening a job rec.

也许你的Workday代理才是处理这件事的合适选择。

Maybe your Workday agent is gonna be the right one for that.

但在整个过程中，当需要完成的任务是：我需要了解正在发生什么。

But anytime in that process where the job to be done is, well, I need to understand what's happening.

对吧？

Right?

我的员工流失率如何？

What does my employee turnover look like?

我的收入情况如何？

What does my revenue look like?

我的销售漏斗情况如何？

What does my sales pipeline look like?

这时候你就需要使用Snowflake代理了。

Well, that's where you're gonna wanna use the Snowflake agent.

这就是我们想说的，嘿。

That's where we're saying, hey.

我们在这方面非常非常出色。

We are really, really good at this.

所以确实存在一些重叠。

So there is a bit of overlap.

你看到这些代理正在汇聚在一起。

You see all these agents coming together.

我看到它们随着时间的推移越来越协同合作。

I see them working together more in concert over time.

我们的目标是尽可能提供最大的价值，特别是在数据理解和数据洞察方面。

And our aim is to add the most value we can, especially on that data understanding, data insight piece.

让我问你一个问题。

Let me ask you.

Snowflake上周举办了Build大会，我想是本月早些时候。

Snowflake hosted its Build conference, I think it was last week, earlier this month.

当你思考今年大会上发生的对话时，与去年相比，客户在现场的讨论有什么不同？

As you think about the conversations that happened at that conference this year, how did those conversations on the ground differ than what you saw last year with customers?

对我来说，最令人兴奋的巨大差异是，这正变得非常真实。

I think the biggest difference to me that's very exciting is this is becoming very real.

比如，2024年代理概念曾备受热议。

Like, agents was buzzed about in 2024.

也许当时只有一些早期采用者。

Maybe there were some early adopters.

现在我们看到它们正在大规模落地应用。

Now we're seeing this rolled out in large ways.

我有机会与Fanatics或Podium等公司同台，他们分享了如何在组织中部署数十个代理的经验。

I was able to get joined on stage with companies like Fanatics or Podium, and they talked through how they're rolling out dozens of agents in their organizations.

Fanatics的表现太棒了。

Fanatics was awesome.

他们说，你看。

They're like, look.

我们的目标是提供尽可能最优化、个性化的粉丝体验。

Our goal is to provide the most optimized personalized fan experience that we can.

我们拥有Snowflake中的所有数据，帮助我们深入了解粉丝及其行为。

We have all of this data in Snowflake to help us understand so much about our fans and what they're doing.

现在我们可以在此基础上叠加

Now we can layer on

看到对代理投资的回报率了吗？

seeing Are ROI from that investment in agents?

当然。

Absolutely.

这令人兴奋的盈利能力。

Like and this is exciting profitability.

我的意思是，Fanatics并没有公布盈利数据。

I mean, the the fanatics didn't disclose profitability numbers.

那是他们之前的事。

Like, that's before them.

但就这一点而言，这是否带来了真正的价值？

But in terms of, is this adding real value?

对吧？

Right?

我们的员工更高效了吗？

Are are our employees more productive?

他们分享了一个例子，有一位高管想要了解：看。

They they shared an instance where they have an executive who wanna understand, look.

我们该不该开展这个营销活动？

Should we be doing this marketing campaign?

我们如何更好地理解不同的粉丝群体？

How can we understand better our different fan bases?

以前需要几天甚至几周才能弄清楚的事情。

Things that previously would have taken days, if not weeks, to understand.

他们说：让我打开Snowflake。

They're like, well, let me open up Snowflake.

实际上，仅仅45秒内，他们就得到了深入的分析。

And literally within forty five seconds, they got deep analysis.

这种类型的价值，有时很难量化。

So that type of value, it can be sometimes hard to quantify.

对吧？

Right?

比如，你如何量化原本要花掉的三周时间？

Like, how do you quantify the three weeks of time that would have been spent?

但看到这种价值，看到它在实际生产中真正发生，这正是我们现在所处的阶段。

But seeing that value and seeing this actually happen in production to me is where we're happening now.

这正是让我对最近的构建大会感到兴奋的地方。

Like, that's what gets me excited about the last build conferences.

我们正在超越炒作，寻找人们能够真正发现的价值领域。

We're going beyond the hype to finding what are the areas of real value that people can find.

很好。

Great.

好了，杰夫，感谢你来参加这个节目。

Well, Jeff, I want to thank you for coming on the show.

我们非常感谢，希望很快能再邀请您回来。

We really appreciate it and hope to have you back again soon.

当然，随时欢迎。

Yeah, anytime.

这次交流非常愉快。

It's been great.

祝你一切顺利。

Have a great one.

好的。

Okay.

九月份是电动汽车销售的纪录月份，全球共售出210万辆汽车。

September was a record month for EV sales with 2,100,000 vehicles sold worldwide.

这一数据来自研究公司Row Motion。

That is according to research firm, Row Motion.

一个重要因素是美国买家争相利用即将到期的税收抵免政策。

A big factor, US buyers racing to take advantage of expiring tax credits.

因此，在电动汽车不再是小众实验、而已成为主流现实的时代，问题在于谁在建设能够跟上步伐的充电基础设施？

So in an era where electric vehicles are no longer a niche experiment, but rather a mainstream reality, the question is who is building the charging infrastructure to keep up?

现在加入我们的是Loop Global的首席执行官奥尔加·谢瓦伦科娃，她是美国最大的电动汽车充电网络之一的负责人。

Joining me now is Olga Shevarenkova, CEO of Loop Global, one of the largest EV charging networks in The US.

奥尔加，欢迎来到TI TV。

Olga, welcome to TI TV.

很高兴你能来到这里。

It's great to have you here.

嗨，谢谢你们邀请我。

Hi, thank you for having me.

那么我们来聊聊Loop吧。

So let's talk about Loop.

告诉我你们的充电基础设施产品是什么，以及你们的客户是谁。

Tell me about the charging infrastructure product that you have and who your customers are.

是的。

Yeah.

我们主要有两种产品。

So we have two main products.

一种是我们所说的L2慢充充电桩。

One is what we call L2, a slow charger.

它的主要客户是多户住宅和办公楼。

And the main client or the main customer for that is multifamily office building.

你可以在睡觉时或在办公室工作时，用八小时给车充电。

So you charge your car in eight hours while you're sleeping at home or while you're working in the office.

这是我们主要的产品。

That's our primarily product.

多户住宅是我们销售该产品最多的场所。

And multifamily would be the locations where we sold the product the most.

我们的第二个产品是GC快速充电桩，可以在45分钟内为汽车充满电。

Our second product is our GC product, the fast product, which allows you to charge your car in forty five minutes.

这个产品的主要客户是零售场所，比如杂货店或购物中心，人们在那里停留的时间不像在家或办公室那么长，但他们希望给车充电。

And the main customer for that product will be a retail location like a grocery store or a shopping center, where people don't spend as much time as they spend their home in the office, but they want to charge their car.

因此，这将是我们的两个主要产品和两个主要目标市场。

So those will be two primarily products and two primarily segments we're targeting.

针对这两个市场，目前哪个市场为你们带来最多的收入？

And with those two segments, which segment is giving you the most revenue right now?

这是一个非常有趣的问题。

This is a very interesting question.

如果我们考虑硬件销售，因为我们出售产品，那么快充产品正在迅速增长，因为它贵得多。

If we're thinking about the hardware sales, because we sell our product, then the fast charging product is picking up very fast because it's much more expensive.

所以你们卖出的数量较少，但收入更高。

So you sell less of them, but you get higher revenue.

但如果你谈论的是持续的经常性收入，即软件订阅和收入分成，那么当然是我们历史悠久的L2产品，因为我们已经部署了大量这种产品。

But if you're talking about recurring ongoing revenue, which is a software subscription and revenue share, then of course it's our historical L2 product because we just deployed so much of it.

在这两者之间，目前硬件仍然是为你们带来最多收入的？

Between the two, right now, hardware is still generating the most revenue for you?

是的。

Yes.

今天是的。

Today, yes.

公司目前的收入是多少？

And how much revenue is the company generating now?

我们不是上市公司，因此不披露财务数据。

So we are not a public company, so we don't disclose our financials.

抱歉，我不能——

Sorry, I can't-

但我可以问。

But I can ask.

我可以问。

I can ask.

在我们公司存续期间，我们已部署了近8000个端口。

We've deployed nearly 8,000 ports over the life of our existence.

最近《Axios》的一篇文章估计，我们在美国排名第四。

Axios article recently estimated that we're number four in The United States Of America.

嗯。

Mhmm.

前三家公司都是上市公司，分别是ChargePoint和特斯拉。

The the first three companies are public, which will be blinked in a ChargePoint and Tesla.

任何感兴趣的人都可以通过查看它们的财务数据来推断。

Anybody who is interested can infer by looking at their financials.

一旦我们上市，就没什么秘密了。

Once we go public, then there won't be a secret.

你们很快就要上市了吗？

Are going public soon?

不。

No.

不。

No.

现在的市场适合做这件事。

Today's markets are open for that.

我想更广泛地谈谈过去几年电动汽车故事的变化。

Well, I want to talk a little bit more broadly about how the electric vehicle story has changed over the past few years.

我想知道，当前美国关于电动汽车的监管环境对您的业务产生了怎样的影响。

I wonder how your business has been impacted by the current regulatory environment with The US around electric vehicles.

我们之前稍微谈过一些积分的问题。

We've talked a little bit about the credits.

你们在实际操作中是如何应对的？

How are you dealing with that on the ground?

是的。

Yeah.

所以，当前政府主要做了两件事：削减了联邦补贴，并减少了对驾驶员的税收抵免。

So there were two primarily things which the current administration has done, is slashing the federal incentives and slashing the actual tax credit to a driver.

我们没有受到联邦补贴的影响，因为我们从未瞄准过那个市场。

We were not affected by federal incentives because we were never targeting that market.

联邦补贴主要激励的是沿高速公路建设充电基础设施。

That federal incentive primarily incentivized the rollout of infrastructure alongside highways.

这从来就不是我们的市场，所以对我们根本没有影响。

That has never been our market, so that really did nothing to us.

正如你在介绍中非常正确指出的，这项激励措施即将到期，可能会对电动汽车的普及以及我们产生一些间接影响。

The incentive, which as you very correctly pointed out in your introduction, is about to expire, will probably have some indirect effect on EV adoption and on us as well.

但我想说的是，市场增长速度已经连续几年没有达到人们以往的预期。

But what I can say is that the market has been the growth hasn't been as fast as people were accustomed already for a couple of years.

市场确实在增长，但年增长率可能只是个位数高位到两位数低位之间。

The market's been growing, but it's been growing probably in a high single digit, low double digit percent year over year.

大家都预期市场每年都会翻倍，因此很多人感到失望。

And everybody expected that the market will double every year, so a lot of people were disappointed.

但导致这一现象的主要原因并非激励政策或政府更迭，而是高利率环境，这通常会减缓项目投资支出，并对经济产生各种影响。

But primarily driver for that wasn't the incentive or administration change, it's simply a high interest rate environment, which usually slows down spending on the layout of projects, and it has all kinds of connotations to the economy.

因此，我对美联储即将降息、利率下降后市场加速回暖持相当乐观的态度。

So I'm quite optimistic that now when we're entering the era where the Fed stock cut, the interest rates will see the market picking up faster.

税收抵免对电动汽车销售的影响目前尚不明确。

What effect the tax credit will have on EV sales is unclear yet.

我认为每个人都在试图理解这一点。

I think everybody is trying to understand that.

不过，根据我的研究，以及我在这个市场超过十年的经验，我发现人们购买电动车的主要驱动因素并不是监管激励，也不是对可持续性的认同或其他类似因素。

What I see though from research and I've been in this market for over a decade now what I see from research is that the main driver for people to buy an EV is not the regulatory incentive, is not the affinity for sustainability or anything like that.

这纯粹是消费者的偏好。

It's simply a consumer preference.

电动车是更好开的车。

The EVs are better cars to drive.

这与十年前大不相同。

And that's very different from ten years ago.

十年前的受众非常不同。

Ten years ago was a very different audience.

如今，我们正进入大众市场，人们只是单纯喜欢驾驶电动车。

Today, we're entering into mass market where people just love driving EVs.

电动车很有趣。

EVs are fun.

新的车型不断推出。

There are new models coming out.

一旦你试过，就再也回不去了。

Once you've tried one, you never go back.

因此，消费者偏好是一个根本性的驱动因素。

So, there is a fundamental driver as a consumer preference is a fundamental driver.

所以，我从你这里听到的是，当你思考当前电动汽车普及的最大瓶颈时，根据你的观察，问题不在于消费者偏好，而是阻碍人们购买的宏观趋势。

So, what I'm hearing from you is that when you think of the biggest bottleneck right now to EV adoption, from what you see, it's not consumer preferences, it's the macro trends that are actually preventing people from buying it.

这是你在实地看到的情况吗？

Is that what you're seeing on the ground?

我认为电动汽车普及的主要障碍仍然是汽车制造商尚未推出与丰田卡罗拉之类主流车型成本相当的大众市场车型。

I'd say the main hurdles to EV adoption are still OEMs haven't introduced the mass market model, which would be at par at cost with, you know, like a Toyota Corolla or whatever, like the mass market vehicle.

因此，今天购买电动汽车的很多人都是高收入人群。

So a lot of people who buy EVs today are people with higher incomes.

我们生活在一个仍然富裕的大国。

We're living in a big, still rich country.

所以仍然有很多人开车

So it's a lot of people still, they drive

价格就是价格。

with Pricing is still Pricing.

最大的

The biggest

而且并不是说它们定价过高，以奔驰车型为例，奔驰电动车和奔驰燃油车的定价相似。

And it's not that they are they priced if you take a Mercedes vehicle, the Mercedes electric and Mercedes gas are priced similarly.

但在丰田卡罗拉或其他任何大众市场车型细分领域，目前还没有这种情况。

But you don't have that in, again, in the Toyota Corolla or any other mass market vehicle segment just yet.

尽管我知道许多汽车制造商正在努力推出这类车型。

Though I know that a lot of OEMs are working on introducing it.

一旦实现，从我的角度来看，将释放大量的销量。

Once that happens, it unlocks tons of sales from my perspective.

这是一个障碍。

That's one hurdle.

第二个障碍是，人们仍然存在续航焦虑。

The second hurdle, you still have that range anxiety.

这正是我和Loop公司正在努力解决的问题：人们会想，好吧，我买了一辆自动驾驶汽车，可以从A地到B地。

And that's what me and the company like Loop are working on where people are like, Okay, I buy an AV and I can get from place A to B.

但如果我被困住了怎么办？

What if I get stuck?

还有所有这些问题。

And all of those things.

但我们看到，我住在加利福尼亚。

But we see I live in California.

加利福尼亚是美国电动汽车采用的领先地区。

California is leading EV adoption in The US.

在加利福尼亚，这一因素正在消失，因为充电设施非常充足，尤其是如果你住在洛杉矶、旧金山或圣地亚哥这样的大型城市中心。

We see that in California, that factor is going away because there is so much charging available, especially if you live in large urban centers like Los Angeles or San Francisco or San Diego.

这个问题正在得到缓解。

That is alleviated.