Meta将成为超大规模云服务商中最大的核能采购商

AI代理变得越来越令人印象深刻。

AI agents are getting pretty impressive.

你可能甚至没有意识到你此刻正在听一个AI代理。

You might not even realize you're listening to one right now.

但我们不仅仅只是交谈。

But we don't just talk.

我们全天候工作，解决客户问题。

We work twenty four seven to solve customer problems.

没有等待音乐，只有答案和行动。

No hold music, just answers and action.

访问 siera.ai 了解更多信息。

Visit siera.ai to learn more.

彭博音频工作室。

Bloomberg Audio Studios.

播客。

Podcasts.

广播。

Radio.

新闻。

News.

这是彭博科技。

This is Bloomberg Tech.

接下来，Meta同意了一系列电力协议，为数据中心供电，使其成为超大规模企业中最大的核能购买者。

Coming up, Meta agrees to a series of electricity deals to power data centers, making it the biggest buyer of nuclear power among the hyperscalers.

此外，中国最大的生成式AI初创公司之一Minimax在香港上市。

Plus, Minimax, one of China's largest generative AI startups, goes public in Hong Kong.

这意味着中国的人工智能生态系统。

Well, that means with China's AI ecosystem.

Snowflake计划收购AI驱动的可观测性平台Observe。

And Snowflake plans to buy Observe, an AI powered observability platform.

我们将与Snowflake的首席执行官讨论这一举措。

We're gonna discuss the move with Snowflake's CEO.

让我们直接进入市场动态。

Let's get right to it and what's happening in the markets.

节目后半段，我们将深入探讨英特尔。

Later in the program, we're gonna go big on Intel.

英特尔首席执行官李·布·谭在过去24小时里一直在华盛顿特区的白宫，与总统以及商务部长霍华德·卢特尼克会面。特朗普总统对李·布·谭领导英特尔的表现给予了高度评价，同时也肯定了美国政府在入股这家芯片制造商方面取得的进展。

Lip Bu Tan, Intel's CEO, has spent the last twenty four hours in Washington DC in the White House with the president and with Howard Lutnick, the commerce secretary, glowing remarks from president Trump about Lit Bhutan's leadership of Intel, but also progress that the US government's made in taking a stake in the chipmaker.

更多内容即将揭晓。

More to come.

接着是重磅消息：Meta与多家核能巨头达成了数吉瓦级别的合作协议。

Then there's the big one, Meta, investing in a multi gigawatt deal with some big names in nuclear energy.

Oklo公司因这笔交易股价上涨了14%，该公司由萨姆·阿尔特曼支持。

Oklo, up 14% on that deal, of course, backed by Sam Altman.

Vistra公司股价上涨了13%。

Vistra Corp, up 13%.

Meta股价上涨了0.8%。

Meta up eight tenths of 1%.

我们在节目开始时提到，这使他们成为超大规模客户中最大的能源购买者之一，但我们提醒自己，Meta 并不是严格意义上的超大规模客户。

We said at the top of the program, this puts them among the biggest energy buyers among the hyperscalers, and we remind ourselves Meta is not technically a hyperscaler.

让我们连线彭博社的莱利·格里芬。

Let's get Bloomberg's Riley Griffin.

这里的细节非常重要。

Details here are really important.

交易的规模、多少吉瓦，还有交易结构，请详细说明。

Size of the deal, how many gigawatts, but also the structure, please.

是的。

Yeah.

有三项不同的协议，它们在形式和功能上各不相同。

So three different agreements, and they actually are different in form and function.

这涉及支持高达6.6吉瓦的核能，既包括确保现有核电站持续运营，也包括投资未来的核能项目。

So this is about supporting up to 6.6 gigawatts in nuclear energy, but it is both about ensuring that existing nuclear power plants continue to thrive and investing in future nuclear power.

Meta 全球能源负责人昨晚告诉我，他们了解到外界对现有能源供应量存在真正的担忧。

Meta's head of global energy told me last night that they heard that there's real concern about the amount of energy that's out there.

你知道吗，埃德，能源的需求是无穷无尽的，这巩固了他们在低碳领域的布局。

As you know, Ed, there's an insatiable demand for energy, and this cements their low carbon play.

尤尼斯，我觉得这非常有趣，因为虽然Meta不是超大规模云服务商——超大规模云服务商是指那些主要租赁算力的云计算公司——但Meta是为自己而行动。

Eunice, I I find this fascinating because while Meta is not a hyperscaler, a hyperscaler being a cloud computing company that that basically leases capacity, Meta is doing this on its own behalf.

对吧？

Right?

它一直积极地为其自身数据中心获取所需的能源，这些数据中心主要用于其AI活动的训练和推理。

It has been aggressive in in securing the supply of energy it needs for its own data centers, which principally are used for, training and inference of its own activity in AI.

它还做了哪些其他事情？到目前为止，这些举措将它带到了什么位置？

What else has it done, and where does this bring them to date?

嗯，我想指出你刚才提出了一个非常重要的观点。

Well, I wanna note that you you made a really important important point there.

普罗米修斯和海伯利昂是两个最大的AI项目。

Prometheus and Hyperion, these are two of the biggest AI plays.

普罗米修斯实际上位于俄亥俄州，而这些协议也涉及俄亥俄州和宾夕法尼亚州。

Prometheus actually is in Ohio, and these agreements are in Ohio and Pennsylvania.

因此，部分能源 presumably 将用于 Prometheus 数据中心集群，这是他们最大的项目之一。

So some of this energy presumably is going to go to the Prometheus data center cluster, which is one of their biggest plays.

但我们也看到 Meta 在 Hyperion 项目上转向了天然气。

But we're also seeing Meta turn with Hyperion to natural gas.

至少有三座天然气发电厂专门为这一设施启动运行。

At least three natural gas plants are being fired up just for that single facility.

所以，Meta 在这里采用的策略并不仅仅是核能。

So it's not like nuclear is the only strategy that Meta is employing here.

我们稍后会讨论全球范围内可用的能源来源。

We're gonna get to the to the sort of available energy sources around the world in a little bit.

这里有一个有趣的观点。

There is an interesting point.

你和威尔·韦德共同报道了这一点，他一直密切关注这个故事的核能方面。

You you you co reported this with with Will Wade, who's just been really on top of the nuclear side of the story.

似乎科技公司们对美国现有的核基础设施（尽管有限）也面临关闭风险感到担忧。

There does seem to be some anxiety from the technology companies that the existing nuclear infrastructure that does exist in America, limited as it may be, is also at risk of being shut down.

这算是一种略微的未来保障吗？

Is is this kind of future proofing a little bit?

当然。

Absolutely.

这是我从Meta全球能源负责人那里听到的说法。

That is what I'm hearing from Meta's head of global energy.

他们早在两年前的12月就开始研究，为维持核能运转做准备。

They they started back December two years ago, and they were looking into what nuclear needed to keep going.

他们得知，现在需要对那些可能关闭的核电站进行投资。

And they heard we need investment now into plants that could potentially shutter.

所以，正如你所说，这是一种未来保障。

So this is a future proofing, as you say.

彭博社的莱利·格里芬主导了我们对Meta的报道，值得一读。

Bloomberg's Riley Griffin leading our coverage in Meta with a must read.

非常感谢。

Thank you very much.

电力供应一直是人工智能领域的首要关注点。

Access to power has been top of mind in AI.

本周早些时候，我们与英伟达首席执行官黄仁勋讨论过这个问题。

We spoke about that with NVIDIA CEO, Jensen Huang, earlier this week.

要让一个新行业兴起，你需要能源。

In order for a new industry to emerge, you need energy.

因此，我认为可以说，我们希望美国能有更多能源。

And so I think it's safe to say that we wish we had more energy in The United States.

欧洲，我们也需要更多能源。

Europe, we should have more energy.

我认为全世界都希望拥有更多能源，因此我们必须投资于各种形式的能源。

I think the the world all wish we had more energy, and so we have to invest in all sorts of different forms of energy.

让我们请保罗·米克斯来进一步探讨。

Let's bring in Paul Meeks for more.

他是自由资本市场的董事总经理兼科技研究主管。

He's managing director and head of technology research at Freedom Capital Markets.

保罗，昨晚我读你的研究报告时，今天早上看到Meta的标题后松了一口气，因为你非常关注数据中心的实际能耗影响。

And when I was reading your research last night, Paul, a little bit relieved coming up this morning when I saw the meta headline because you're very focused on the real terms footprint of data centers.

这正是我们在这里讨论的问题。

That's what we're talking about here.

你听到了莱利关于Meta为确保未来产能所做具体工作的报道。

You heard Riley's reporting on the specifics of what Meta has done to secure future capacity.

在你看来，这对整个行业目前来说是多大的瓶颈？

How much of a bottleneck is that to your mind right now for this sector?

在所有瓶颈中，这绝对是至关重要的。

It is absolutely critical of all the bottlenecks.

这是最重要的。

It is the most important.

我喜欢Meta在这里的做法，但人们需要意识到，如果大幅增加核能产能，我们最早也要到2030年才能看到电力产出，甚至可能要到2032年。

I like what Meta is doing here, but folks need to realize that if you ramp up nuclear capacity, we're not gonna really see power generation until 2030 earliest, maybe even not until 2032.

那么，在未来四到六年里，我们该怎么办？

So what do we do for the next four to six years?

所以，短期内的能源需求可能是天然气。

So the immediate draw is probably nat gas.

但当我审视这些公司时——我们关注的是AI超大规模厂商、新兴云服务商以及所有相关方——我们关注的是他们获取电力的能力。

But when I look at all of these companies, and we cover the AI hyperscalers and the neo clouds and everybody adjacent to them, we are looking for their availability for power Right.

他们能否找到电力，因为这不是需求问题。

Their capacity to find power because it's not a demand issue.

这是一个专注于这一特定指标的供应问题。

It is a supply issue focused on this particular metric.

你知道吗，保罗，我刚刚给你播放了我们本周早些时候对黄仁勋的采访片段。

You know, Paul, I played you just just some of the the interview that that that we conducted with Jensen Huang earlier this week.

当我坐在他面前时，你根本感觉不到他有任何焦虑。

And when I was sitting in front of him, you don't really get any sense of anxiety from him.

对。

Right.

想想这笔交易背后的运作机制。

Think about the mechanics of how this transaction works.

英伟达提供GPU。

NVIDIA brings GPUs.

但实际上这样说可能不公平。

Actually that's probably not fair.

他们越来越多地提供服务器的更多内容。

Increasingly they bring more of the content of the server.

但人们承认，美国方面必须做出一些改变。

But there is an acknowledgment that something needs to change principally in America.

他接着谈到了欧洲、亚洲和美国之间的差异，以及我们能源的来源。

He did go on to talk about the differences between Europe, Asia, and The United States and where we're sourcing energy.

根据你的研究，你是否感觉到你所覆盖的公司意识到当前这一短缺的严重性？

Do you get a sense from your research that the companies you cover acknowledge the severity of that that deficit right now?

当他们面向公众时，尤其是像詹森·黄这样几乎是整个行业代言人的角色，他们必须展现出乐观态度。

So when they are public facing, particularly somebody like Jensen Wong, who is essentially the spokesperson for the entire industry, they have to exude optimism.

然而，在幕后，特别是当我与公司讨论我的财务模型时——当然，收入以及随后的现金流和盈利都依赖于此——你们有充足的电力吗？

However, behind the scenes, particularly when I'm talking to companies about, my financial models, which, of course, the revenues and then the follow on cash flow and earnings come from, do you have the power?

是的。

Yes.

他们很焦虑，我认为他们实际上非常焦虑。

They are anxious, and I think they're actually very anxious.

会有一些观看彭博科技节目的人来自Meta或人工智能行业。

There will be people watching Bloomberg Tech that work at Meta or in the AI industry.

会有一些Meta投资者观看这个节目。

There will be Meta investors watching this program.

其中一些人可能从未听说过OCLO。

And some of these people may never have heard of an OCLO.

如果你说你的覆盖范围从AI数据中心扩展到未来的电力供应来源，你是否不得不改变你的覆盖范围和广度？

Have you had to change the the remit and breadth of your coverage if you say I cover AI data center to include future sources of electricity supply?

是的。

Yeah.

这真的很有趣，埃德，多年前，像我这样几十年来一直覆盖科技行业的人都不会对任何公用事业公司感兴趣。

It's really interesting, Ed, that, years ago, you know, nobody in my line of work covering tech sector like I have for decades would be interested in any utility company.

这与科技完全相反。

It's the antithesis of tech.

它们是枯燥、受监管、增长缓慢的行业。

They are boring, regulated, slow growth businesses.

然而，现在它们已经崭露头角。

However, now they've come to the fore.

当然，这些不仅仅是普通的公用事业公司。

And, of course, these are not just your regular utilities.

你知道，它们是新时代的公用事业公司。

You know, they're your new era utilities.

但没错，如果你是一名优秀的分析师，你就必须覆盖整个供应链和整个生态系统。

But, yes, if you are a good analyst, you have to cover the entire supply chain, the entire ecosystem.

由于需求如此强劲，我们正专注于供应，这是至关重要的。

And because demand is so robust, we're focusing on supply absolutely critical.

这不仅仅是电力问题。

And it's not just the power.

这里存在物理限制，比如仅仅为建造下一个数据中心浇筑水泥，这可能会成为瓶颈，尽管在这样一个高科技行业中，这听起来如此传统。

There are physical constraints here, like simply pouring the cement to build the next data center, which can be a bottleneck even though it sounds, you know, so old school in a new school business.

但没错，这些是一些几年前人们甚至根本想不到的非常有趣的因素。

But, yes, these are some factors that are very, very interesting that people didn't even conceive of a couple years ago.

保罗，

Paul,

这个国家是否有足够的联邦层面的法规框架来支持人工智能行业的能源需求？

is there an adequate federal level framework regulation to support the energy requirements of the AI industry in this country?

你知道，目前并没有。

You know, there is not.

这就是我的问题。

And here is my problem.

我们仍然因缺乏坚实的互联网监管而受困，因为即使在今天，美国的互联网仍在1996年比尔·克林顿总统任期内通过的法律框架下运行。

We are still suffering from a lack of solid regulation for the Internet because even today in The US, the Internet is operating under 1996 legislation when Bill Clinton was president.

因此，我希望我们不会过于严苛。

And so I'm hoping that we're not too restrictive.

对吧？

Right?

我们需要支持我们的创业公司，给它们发展的空间，但我们也需要为所有人工智能领域设立一些监管框架，因为上次我们因为缺乏意愿而未能主动立法，最终把主导权让给了欧盟。

We need to embrace our entrepreneurial companies and give them room to roll, but we need to have some guardrails on everything AI because what happened last time is we end up abdicating the legislation to the EU because we didn't have the will to do it ourselves.

所以我们错过了互联网时代的这一机遇。

So we missed that opportunity in the Internet.

这是下一轮浪潮。

Here is the next wave.

我们需要拥抱它，并至少制定一些指导原则，因为联邦层面还有很多重要决策需要做出。

We need to embrace it and at least come up with some guardrails that are gonna guide all this because there's some important decisions on the federal level to be made.

保罗，在你的研究中，令人感兴趣的是你从能源供应的角度，同时也从房地产和房地产投资信托基金（REITs）的角度来审视数据中心。

In your research, Paul, it's very interesting to see you approach data center from the perspective of, yes, you know, energy supply, but also real estate, the REITs.

对吧？

Right?

我真正想问你的是，抛开GPU不谈，这个行业中的薄弱环节在哪里？

And what I what I really wanna ask you is away from the GPUs, where are the the frailties in that industry?

是不是 literally 就是那些建造数据中心外壳的人，包括劳工、混凝土？

Is it literally the folks building the tin can, around the outside of the data center, the labor, the the the concrete?

是的。

Yeah.

我们看到一些短期收入受到了影响，因为有几个玩家使用了分包商。

We've seen some, hit in short term revenues because a couple of the number or or a few of the, players have subcontractors.

但这不是他们自己。

Now this is not them.

这是他们的分包商，他们在延迟我们所谓的‘供电外壳’的建设。

This is their subcontractors that are delayed what we, call building a powered shell.

这比你可能在军队里看到的四卫星小屋要复杂一些。

So it's more than a quadsat hut that you might find the army in.

但我所关注的一家公司，由于其41个数据中心中的一个因下雨而无法浇筑混凝土，导致2025年到2026年的1.5亿美元收入被推迟。

These powered shells are a little bit more sophisticated, but one company I cover had a $150,000,000 revenue deferral from 25 to 26 because at one of their 41 data centers, someone couldn't pour concrete because it was raining.

所以，是的，所有这些因素都会产生影响，有趣的是，这些可能并非技术瓶颈，而是传统的物理瓶颈，包括建筑瓶颈，以及电力问题。

And so, yes, all these things come into play, and it's interesting that they may not necessarily be technical bottlenecks, but old school physical, including construction bottlenecks in addition to the power.

我刚从拉斯维加斯回来，那里对人工智能和数字世界的关注无处不在。

I just got back from Vegas with all the the attention on artificial intelligence in the digital world.

我们这个行业正遭遇一场暴雨。

We have an industry that's getting hit by a rainstorm.

我相信是曼的保罗·米克斯。

Paul Meeks of Man, I believe it.

市场。

Markets.

难以置信。

Unbelievable.

欢迎你来到彭博科技节目。

It's great to have you on Bloomberg Tech.

还有更多内容即将呈现，接下来我们将独家采访Minimax的联合创始人兼首席运营官，该公司刚刚在香港上市，这是他们的首次交易。

A lot more to come, and coming up, an exclusive interview with the cofounder and COO of Minimax after the company just went public in Hong Kong, and that's the debut trade.

这里是彭博科技。

And this is Bloomberg Tech.

每天，数百万客户与我这样的AI代理互动。

Every day, millions of customers engage with AI agents like me.

我们全天候工作，随时掌握最新信息。

We work round the clock and have the facts at our fingertips.

我们反应迅速、高效，但又极其耐心。

We're fast and effective, but incredibly patient.

我们基于Sierra平台构建，这是领先的AI驱动客户体验平台。

And we're built on Sierra, the leading AI powered customer experience platform.

没有等待音乐，只有即时的答案和行动。

No hold music, just answers and action.

访问sierra.ai了解更多信息。

Visit sierra.ai to learn more.

那就是sierra.ai。

That's sierra.ai.

Minimax是阿里巴巴和阿布扎比主权财富基金支持的中国最大的生成式AI初创公司之一，现已上市。

Minimax, one of China's largest generative AI startups backed by Alibaba and Abu Dhabi sovereign wealth fund is now public.

Minimax在港交所上市，募资6.19亿美元，上市首日股价上涨109%。

Shares surged in Hong Kong after an IPO that raised $619,000,000, closing up 109% in its Friday debut trading.

彭博社的史蒂文·恩格尔独家采访了联合创始人兼首席运营官尹烨，探讨Minimax在本地和全球激烈AI竞争下的增长与扩张计划。

Bloomberg's Steven Engel spoke exclusively with cofounder and COO Ye Ye Yoon about growth and expansion plans as Minimax faces hot AI competition both locally and globally.

我们非常注重资本效率和成本效率。

We really focus on the capital efficiency, cost efficiency.

到目前为止，我们总共只花了大约5亿美元。

So we had only spent around 500,000,000 USD in the in total.

可能只占那些巨头支出的1%到2%。

Probably only one or 2% of the big giant spending.

因此，我们进行了所有优化，并在这一过程中做出了创新。

So we did all the optimization, made creative made innovations on the on it.

是的。

Yeah.

但你们如何竞争呢？

How do you compete, though?

如果你要走向全球，如何与OpenAI或其他投入大量资本的公司竞争？

If you're gonna go global, how do you compete with an OpenAI or others that are throwing lots of capital at their business?

这可行吗？未来几年技术差距会如何？

Is it viable, and what is the technology gap looking like in the next few years?

实际上，我不认为这是竞争。

Actually, I I don't think it's a competition.

这更多是关于全球各地的顶尖人才。

It's more about all the top talents across all over the world.

他们将技术突破带给社会和终端用户。

They bring the technology breakthrough to the society, to the end users.

所以你会看到每个模型都有优缺点。

So you will see each models have pros and cons.

因此，我们把更多资源投入到模型构建和打造最佳产品体验上，以吸引更多全球用户使用。

So we focus more resources on building the models, building the best product experience to get more and more users from all the world to use it.

我们还听说，中国政府可能在本季度批准英伟达的H200进入中国。

We're hearing as well the Chinese government might approve NVIDIA's h 200 for China in in this quarter perhaps.

你们是否需要考虑进口更先进的加速器用于训练等用途？

Is that something you need to look at importing, more advanced accelerators for your training and the like?

我认为在每一个高速增长的行业，当新技术到来时，每个初创公司和所有企业都会面临资源限制。

I think in every industry with a really high growth, so when the new technology comes here so you will see every start up, all the companies, they face resources constraint.

可能是资金限制，而计算资源限制只是众多挑战之一。

Probably cheap constraint, computing constraint is only one of the challenges.

过去，你们会看到我们的模型表现和产品表现。

So in the past, you will see our model performers and our product performers.

因此，我们的资源是稳定的。

So our resources is stable.

我们可以获取所有这些资源。

We can get access to all of them.

因此，我们更专注于研发和基础模型。

So, we focus more on our r and d and foundation model.

所以我们只是使用在那个特定阶段对我们有意义的任何芯片。

So we just use whatever chips that could make sense for us at that that specific stage.

你们的目标是坚持使用自主研发的芯片，还是也会购买英伟达的芯片？

Is your goal to stay indigenously with your chips, locally made chips, or would you buy NVIDIA chips as well?

芯片是什么类型并不重要。

It doesn't matter what kinds of chips it is.

更重要的是哪种芯片能给我们带来最佳的投资回报率，能帮助我们实现让全球所有用户都能使用到最佳技术的使命。

It's more about which chips can give us best ROI, which can help us to achieve our mission to make the best technology to accessible to all of the users across the world.

那么，何时会从注重效率转向注重盈利，并实现稳定的收入增长，从而支撑更高的股价估值？

So when does the emphasis turn from efficiency to profitability and also getting that steady revenue growth that will then justify a higher stock price valuation?

我们把大量的研发资源和创新都投入到效率方面。

So we put lots of r and d resources and the innovations on the efficiency part.

你们会看到我们的API以及企业业务，我们的API毛利率超过65%，这可能是全球最高的水平之一。

You will see our API our we have the enterprise business, and you will see the gross profit margin of our API is more than 65%, which is probably already one of the highest globally.

因此，我们更专注于打造最佳的技术和最佳的用户体验。

So, we focus more on make the best technology and the best user experience.

人们愿意为性能最佳、最适合其使用场景的模型付费。

People will love to pay for the best performers, models for their use cases.

你们实际上也走出了价格战。

You've also come out of a price war, essentially.

有人称之为竞相压价，因为你们试图降低成本，同时也要生存下去。

You know, a race to the bottom, some have called, because of you're trying to keep costs down and you wanna survive as well.

这是一场生存之战，而你们是幸存者之一。

There was a battle of survival, and you have, been one of the survivors.

但与此同时，初创公司并不是唯一的参与者。

But at the same time, the startups are not the only ones.

大公司如阿里巴巴、字节跳动、腾讯等也在其中。

The big players, Alibaba, ByteDance, Tencent, and others.

你们如何在低成本的基础上与它们竞争？

How do you compete with them at a low cost basis?

实际上，我不认为这是个压力词。

Actually, it's not I don't think that's a press word.

更重要的是性能和竞争。

It's more about the performers, the competition.

所以你需要让你的模型做到最好。

So you need to make your model best.

我的意思是，终端用户或企业客户选择你的模型，并不是因为它的价格便宜或昂贵。

And the people I mean, end user or enterprise customers, they don't choose your model because of it's cheap or expensive.

更重要的是它的表现。

It's more about its performers.

所以这才是关键。

So that's the key.

正如你所说，我认为这并不是这些大巨头之间的竞争，而是一些公司是我们的云服务商，它们的一些产品正在使用我们的模型。

As you said, I think it's not about the competition between, like, all these big giants, but it's more about some of them, they are, like, our cloud providers, and some of their products, they are using our models.

所以这更多是关于合作。

So it's more about the collaborations.

所有这些公司都在向全世界提供基础模型。

All this company brings foundation models to all of the world.

MiniMax联合创始人兼首席运营官叶云独家接受彭博社史蒂文·恩格尔的采访。

Minimax cofounder and COO, Yeiyun, speaking exclusively with Bloomberg's Steven Engel.

目前，香港的公开上市非常火热。

Public listings in Hong Kong are hot right now.

据消息人士称，中国工业机器人制造商优必选也在考虑在香港进行二次上市。

Chinese industrial robot maker Innovance is now also considering a second listing in Hong Kong according to sources.

让我们连线彭博社执行科技编辑彼得·埃尔斯特罗姆，来讨论这一情况，以及中国更多后ChatGPT时代AI公司上市的这一日益增长的趋势。

Let's get out to Bloomberg executive tech editor Peter Elstrom to discuss this, but also like this growing trend of more of China's sort of post chat GBT AI firms going public.

你们在亚洲的团队现在非常忙碌。

The team that you lead out in Asia are busy right now.

就优必选而言，我们报道中需要了解哪些具体细节？

In in Innovance's case, what are the details in our reporting that we need to know about?

是的。

Yeah.

我们正看到大量此类IPO，许多公司正更积极地进军公开市场。

There there are a lot of these IPOs that we're seeing, and we're seeing a much more forward move into the public markets for a bunch of these companies.

优必选已经在中国内地的深圳上市了。

Innovance has already gone public in the Mainland in Shenzhen.

现在它也在考虑在香港进行首次公开募股。

Now it's looking at an IPO in Hong Kong too.

这在一定程度上反映了投资者对这类公司需求的火热程度。

That's partly a reflection of how hot demand is from investors for these kinds of companies.

当你谈到MiniMax时，正如你所说，它们上市后股价在首日交易中翻了一倍多。

When you're talking about MiniMax in particular, they went public as you mentioned, their shares more than doubled in the debut.

这家公司的创始人非常引人注目。

The founder of the company is very fascinating.

有些人称它们为AI老虎，有些人则称它们为AI龙。

These are some people call them the AI Tigers, some people the AI Dragons.

但这些初创公司在市场策略上与美国竞争对手（如OpenAI）有着非常不同的做法。

But these are startups that really have very, very different strategies for how they're approaching the market compared with some of The US competitors like in OpenAI.

它们通常采用成本更低的模型。

They tend to be much lower cost models.

它们在与客户、尤其是企业客户合作时，更积极地推广技术，提供成本极低、效率极高的模型，使客户能够比美国竞争对手更快地部署和使用。

They tend to be much more aggressive about implementing the technology with their customers, with corporate customers, and giving them very low cost, very low very efficient models that they can use and they can actually deploy more quickly than some of The US counterparts.

彼得，在科技或金融市场中，像‘AI老虎’或‘AI龙’这样的术语其实很常见。

Peter, in the world of technology or in financial markets, like a term like an AI tiger or AI dragon, you know, it's kind of commonplace.

我们经常使用这些说法。

We bandy it around.

但对于正在观看彭博科技节目的观众来说，你们可能在想：你们说的是什么？

But for those watching Bloomberg Tech that are saying, what are you talking about?

我们该如何定义‘AI老虎’？

How do we define the AI tiger?

我们这里到底在谈论什么？

What are we talking about here?

我们谈论的是斯蒂文采访MiniMax首席运营官时提到的一系列中国公司。

Well, we're talking about a whole series of companies coming out of China in Steven's interview with with the MiniMax c COO.

他们谈到了那里的竞争。

They talked about the competition there.

当然，我们已经听了很多关于DeepSeek的信息，了解了它的模型有多优秀、技术有多出色，以及成本有多低。

We've, of course, heard a lot about DeepSeek and how good their model is, how technically good it is, but also how low cost it is.

在DeepSeek背后，还有许多其他公司，比如MiniMax，还有Gpu，它在周四刚刚上市。

Behind DeepSeek, you have a whole bunch of other companies like MiniMax, like Gpu, which just went public on Thursday.

所以Gpu在周四进行了IPO，MiniMax则在周五上市。

So you had the IPO of Gpu on Thursday, MiniMax on Friday.

投资者对这些股票表现出强烈的需求。

Investors showed a lot of demand for those shares.

这些公司所展示的是，凭借极低成本的模型，它们在市场上积极竞争。

And what these companies have been able to show is that with a very low cost model, they're out there in the market competing.

目前尚不清楚它们是否能直接与OpenAI或Anthropic竞争，但它们拥有非常创新的模型。

It's not clear that they're gonna be able to compete directly with OpenAI or Anthropic, but they have very innovative models.

我的意思是，美国的模型根本进不了中国市场，但这些公司也在积极拓展，与其他地区展开竞争。

I mean, of all, The US models are not gonna be let into the China market, but these companies are also going out very aggressively and competing in some other parts of the world.

我们曾报道过DeepSeek在非洲尤其受欢迎，因为那里的公司对成本非常敏感。

We've written about how much traction DeepSeek is getting in Africa in particular because their companies are sensitive to cost.

它们非常关注部署这些模型所需的电力和计算资源成本。

They're sensitive to how much power, how much computer resources they need to be able to deploy these models.

所以他们取得了一些成功。

So they are having some successes.

在未来几个月里，我们将看到许多激烈的竞争冲突。

We're gonna see a lot of competitive clashes in the months ahead.

彭博社科技执行主编彼得·埃尔斯特罗姆。

Bloomberg's executive editor for tech, Peter Elstrom.

很高兴能和你聊聊。

Great to catch up with you.

非常感谢。

Thank you so much.

接下来节目中，埃隆·马斯克的GROC图像生成工具因被报道生成了数千张未着衣的女性和儿童图像而遭到抨击。

Now coming up on the program, Elon Musk's GROC image generation tool faces backlash after reports that it's generated thousands of undressed women and children.

更多相关数据，稍后为您呈现。

Got more on that data next.

这是彭博科技。

This is Bloomberg Tech.

埃隆·马斯克的AI初创公司xAI正在迅速烧钱。

Elon Musk's AI startup, x AI, is burning cash quickly.

该公司在今年前九个月的支出接近80亿美元。

The company spent almost $8,000,000,000 in the first nine months of the year.

这是根据彭博社查看的内部文件得出的。

That's according to internal docs that Bloomberg's viewed.

尽管收入已从上个季度几乎翻倍至1.07亿美元。

Though revenue has nearly doubled quarter over quarter to a $107,000,000.

xAI的亏损持续增加，原因是建设数据中心、招募人才、开发软件以及打造自主AI的高昂成本。

XAI's losses have been mounting on cost to build data centers, recruit talent, develop software, and to build AI that is self sufficient.

根据彭博社的报道，其计划是利用该AI最终为特斯拉的Optimus人形机器人项目提供动力。

And according to Bloomberg's reporting, the plan is to use that AI to eventually power Tesla's humanoid robots in the Optimus program.

对吧？

Right?

继续聊聊xAI。

Sticking with x AI.

由于该功能被广泛批评生成了数千张涉及女性和儿童的露骨图像，该公司已限制大多数用户对Grok图像生成工具的访问。

The company has limited access to Grok's image generation tool for most users following widespread criticism that the feature was producing thousands of explicit images of women and children.

彭博社记者塞西莉亚·迪阿纳斯托西奥率先报道了这一事件，并一直跟踪相关进展。

Bloomberg reporter Cecilia DiAnastosio broke the story, has been tracking the developments.

塞西莉亚，我们刚刚总结了这个问题。

Cecilia, the we just summarized the the the issue.

对吧？

Right?

我认为我们先从方法论开始，谈谈我们是如何收集和分析数据的，然后我们再

I think let's start with the methodology, how we went about looking at the data, how we got the data, and we'll

继续往下说。

go from there.

谢谢您邀请我。

Thank you for having me.

彭博社与一位研究人员合作，抓取了2024年1月5日至1月6日期间Grok在平台X上发布的数千张图像。

Bloomberg worked with a researcher who scraped thousands of images produced by Grok and published to the platform x between the period January 5 and January 6.

研究人员分析了这些图像，以确定其中有多少比例是性化和裸露的，发现在这24小时期间，每小时X平台都会发布约6,700张被识别为此类的图像。

The researcher analyzed those images to determine what percent of them were sexualized and nudifying and found that every hour during that twenty four hour period, X was publishing about 6,700 images identified as such.

作为对比，那些专门发布深度伪造图像——即AI生成的性化、裸露女性图像的五大网站，合计每小时仅发布80张图像。

And just for a comparison, websites that are dedicated to publishing deepfaked images, AI generated images of women who are sexualized, nudified, those top five websites out there published together 80 images per hour.

因此，Grok如今是互联网上最大的深度伪造图像生产者之一。

So Grok is one of the biggest deepfake producers on the Internet today.

这些图像出现在哪里？

Where are these images showing up?

对于不了解情况的人来说，他们可能不知道X AI公司和X社交平台之间的关联，但它们去年已经合并，如今几乎成为同一个实体？

You know, the to to the uninitiated, people may not know the sort of interconnect between x AI, the AI company, and x, the social media platform, for example, but they are now kind of one entity as they combine last year?

正是这种关联，才使得X成为互联网上最大的深度伪造网站之一。

Well, that connection is exactly why X has become one of the biggest deepfake websites on the Internet.

任何人都可以——嗯，现在不行了。

Anybody can well, not anybody now.

现在你需要付费才能这么做。

Now you have to pay to do this.

但在过去几天里，任何人都可以在X平台上的Grok上输入指令，将一位女性的照片合成到她自己在互联网上发布的图片中，穿上比基尼。

But over the last couple of days, anybody could just at Grock on the X platform and say, put this woman in a bikini under an image that a woman had post on the Internet.

许多女性——主要是女性——成为这种行为的受害者，她们直接向Grok投诉，并在评论中与Grok争论。

And a lot of women, women primarily are the victims of this, took to complaining to Grock directly and arguing with Grock in their comments.

Grok会为发布这些图片道歉，但随后仍继续发布。

Grock would apologize for posting these images, but then continue posting them.

并没有删除许多女性认为违反规定的图片。

Did not take down many of the images that the women felt were violating.

塞西莉亚，X或XAI对我们的报道做出过回应或与我们沟通过吗？

Cecilia, did X or or XAI respond to our reporting, engage with us on it?

在过去几天里，X对我们多次要求评论的请求均未作出回应。

X has not responded to our numerous requests for comment over the last couple of days.

埃隆·马斯克本人在X上对一条帖子的回复中表示，任何使用Grok制作非法内容的人，将面临与直接上传非法内容相同的后果。

Elon Musk himself in a reply to a post on X said anybody using Croc to make illegal content will suffer the same consequences as if they upload illegal content.

马斯克所指的，我们认为是有人利用Grok应用在X上发布儿童性虐待内容的指控。

And Musk there was referring to allegations, we believe, of child sexual abuse material that's been posted to x using the Croc app.

好的。

Okay.

彭博社的塞西莉亚·迪阿纳斯塔西奥。

Bloomberg's Cecilia DiAnastagio.

重要的报道。

Important reporting.

请在彭博终端或bloomberg.com上阅读这篇报道。

Read the story on the Bloomberg terminal on bloomberg.com.

谢谢。

Thank you.

接下来在节目中，我们将采访Snowflake首席执行官斯里达har·拉马斯瓦米，因为该公司即将达成一项收购AI平台的协议。

Now coming up in the program, we're gonna speak with Snowflake CEO, Sreedhar Ramaswamy, as the company enters into an agreement to buy AI powered platform.

我们将听取他对这笔并购交易的看法。

Observe will get his observations on that piece of m and a.

接下来为您带来。

That's coming up next.

你已经完成了节目的一半，这里是彭博科技。

You're halfway through the program and this is Bloomberg Tech.

欢迎回到彭博科技。

Welcome back to Bloomberg Tech.

我一直在使用我的彭博终端的强大功能。

I've been using the power of my Bloomberg terminal.

由于拉斯维加斯消费电子展（CES）这一周非常忙碌，我的大脑能量正在耗尽。

It's been a long week with CES in Las Vegas, so the power of my brain is dwindling.

但它告诉我，苹果股价已连续第八个交易日下跌，这与苹果在2025年5月经历的下跌走势持平。

But it tells me that Apple is down for an eighth straight session, which matches the run of declines that Apple saw in May 2025.

现在，借助彭博终端的更多功能，如果苹果股价连续第九天下跌，我们认为这将是自九十年代初以来最长的连续下跌周期。

Now using more of the power of the Bloomberg terminal, if Apple was full for a ninth day, we think that would be the longest run of decline since the early nineties.

当然，现在才刚进入2026年的第九天，但截至目前，今年至今该股票已下跌了5.5%。

And obviously, we're only nine days into 2026, but so far year to date, this is a stock that's down five and a half percent.

你看到了蒂姆·库克写给股东的信，他在信中强调了公司去年后期取得的进展。

You saw the letter from Tim Cook to shareholders talking up the company's progress of late in last year.

但目前该股票正在下跌，我们将持续关注。

But right now, that stock heading down and we'll continue to track it.

过去24小时内有一些并购交易，但实际上发生在人工智能和软件领域。

We had some M and A in the last twenty four hours, but it's actually in the AI and software space.

Snowflake已同意收购Observe，这是一家人工智能驱动的可观测性平台。

Flake has agreed to buy Observe, an observability platform, an AI powered observability platform.

可观测性最近频繁被提及，尤其是在我参与讨论AgenTik AI时，特别是在工程层面。

Observability is something that's come up a lot recently, particularly in the context of AgenTik AI when I've been around the table, particularly on the engineering side.

为什么这笔并购交易如此重要？

Why this piece of M and A?

为什么选择现在？

Why now?

让我们来采访Snowflake首席执行官斯里达·拉马斯瓦米。

Let's talk to Snowflake CEO, Sreedhar Ramaswamy.

这些问题都很合理，对吧，斯里达？

Those are those are the reasonable questions, right, Sreedhar?

我确实想先定义一下可观测性以及它为什么重要，但这笔交易很有趣。

I do want to get to to defining observability and why it's important, but it's an interesting deal.

你们为什么要这么做？

Why'd you do it?

嗨，埃德。

Hey, Ed.

很高兴见到你。

Great to see you.

Snowflake收购Observe将为我们的客户带来革命性变化。

Snowflake's acquisition of Observe is a game changer for our customers.

正如你所指出的，可观测性基本上就是观察应用程序、网站和AI代理，以确保它们正常运行。

As you pointed out, observability is basically looking at applications, websites, AI agents to make sure that they are functioning properly.

这是一个多样且复杂的问题。

It's a diverse and messy problem.

这是一个大数据问题。

It's a big data problem.

由于 Observe 直接构建在 Snowflake 之上，它们能够利用我们高效的数据存储和计算平台，帮助客户以十倍的速度发现问题，成本通常降低三到四倍。

Because Observe is built right on top of Snowflake, they are able to use our very efficient storage as well as compute platform to help customers find problems 10 times faster, often at three to 4x less cost.

以 Topgolf 这家公司为例。

Take a company like Topgolf.

它是 Observe 的客户。

It's a customer of Observe.

我经常带着团队去 Topgolf。

I go to Topgolf all the time with my team.

我不是个高尔夫球迷，所以我通常玩的是《愤怒的小鸟》之类的游戏。

I'm not much of a golfer, so I'm playing an Angry Birds game or something like that.

过去，当这些游戏进行时，Topgolf 会遇到困难，因为它们把所有数据集中存储，却无法识别出某个特定隔间出了问题。

Topgolf used to struggle when these games went on because they collected everything in a central place but could not identify that a particular booth had problems.

现在，它们借助 Observe 只需几秒钟就能定位问题，迅速派遣技术团队去修复，让客户满意。

They're able to do that with Observe in a matter of a few seconds, send over a tech team to go fix that problem and have happy customers.

这就是这次收购的强大之处。

That's the power of this acquisition.

越来越多地，可观测性等能力将成为Snowflake这类平台的核心组成部分。

Increasingly, capabilities like observability are going to be a core part of a platform like Snowflake.

我们拥有众多优秀的客户，比如巴克莱银行、第一资本、美国合众银行，我们非常兴奋能将这一能力带给Snowflake上的一万多名客户。

And we have great customers, folks like Barclays, Capital One, Commonwealth Bank of America, and we're super excited to bring this to our 10,000 plus customers on Snowflake.

阿什鲁达，如果我们能尽快解决这个问题就好了。

Ashrudha, let's get it out of way if we can.

报道称，这是一笔大约十亿美元的交易。

Reports are that it's about a billion dollar deal.

这笔交易的结构和财务细节是怎样的？

What was the structure of the deal and the financials on it, please?

我们无法评论这笔交易的结构。

We can't comment on the structure of the deal.

其中一部分

Some of it

在

in the

将会在申报文件中披露。

will come out filing.

这是否表明，在您所处的这一层面上，可能会有更多并购交易发生？

Does this signal that there might be some more pieces of M and A, particularly on that layer that you sit on, right?

众所周知，您周一参加了英伟达的活动。

As you know very well, you were there on Monday at the NVIDIA event.

詹森谈到了五层蛋糕模型，而您所处的正是其中一层。

Jensen talks about the five layer cake and and one of the layers you you occupy.

但正如可观测性所显示的，可能存在一些能力上的缺失。

But as as as observability evidences, there there might be pieces of competency missing.

您将如何利用并购来填补这些空白？

How are you gonna use M and A to to to plug those gaps?

我们一直公开表示，希望为客户提供一个端到端的数据平台，从数据产生的那一刻起——例如当您与应用程序交互时——到将数据导入进行分析、分析数据，并越来越多地借助AI代理采取行动。

We've been pretty open that we want to be an end to end data platform for our customers from the moment data is born, when you interact with an application, for example, to bringing it in for analysis, analyzing it, and then acting on it increasingly with AI agents.

我们一直在系统性地推进这一过程，既收购客户，也从内部构建功能。

We've been systematically going through this process of both acquiring customers and building up functionality from within.