Markets Weekly November 15, 2025
发布时间 2025-11-15 17:58:50 来源
Hello my friends, today is November 15th and this is markets weekly. So this past week was another exciting week in markets. We had finally the end of the government shutdown, the longest in the US history. And we also learned that some of the data that releases that we missed, we may never see. So we are probably never going to find out what happened to the JAWS market in October nor will we have an October CPI.
你好，朋友们，今天是11月15日，这是市场周报。在过去的一周里，市场再次经历了激动人心的时刻。我们终于迎来了美国历史上最长的政府关闭的结束。此外，我们了解到一些我们错过的数据发布可能永远不会公开。所以，我们可能永远不知道10月份的劳动市场情况，也不会得到10月份的消费者价格指数（CPI）数据。
Looking at the equity markets, it was also about a week. We rallied earlier in the week, but towards the end of the week, we kind of had a lot of volatility, big declines on Thursday. On Friday, we once again tested the 50 day moving average on the S&P 500, bounced off of it, and remember, we are in a seasonally positive part of the year and that Christmas center rally is a real thing. Now it seems like the S&P 500 bottom markets were upset by two things. First off, of course, was that it looks like a Fed December cut is becoming less likely. And secondly, there are cracks emerging in the AI narrative, which has been a big driver in the equity market.
在股票市场方面，这一周也是不平凡的一周。我们在周初有所上涨，但到了周末时，市场出现较大波动，周四大幅下跌。周五，我们再次测试了标普500指数的50日移动平均线并反弹回升。请记住，我们正处于一年中季节性较为有利的阶段，“圣诞老人反弹”行情是确实存在的。现在看来，标普500指数的市场底部受到了两件事情的影响。首先，市场对美联储在12月降息的预期变得不那么乐观。其次，人工智能这一股市大推力的叙述中开始出现裂缝。
So today, let's talk about two things. First off, let's talk about why the market is pricing in only a 50% chance of a December rate cut and what the cracks are in the AI narrative, starting with the Fed. Now let's rewind a little bit. Now after the September FOMC meeting, the markets were pricing in a 100% chance of a December rate cut. Why would it because the Fed basically guided towards that in September? Remember, in September, we got a dot plot. And if you look at the dots on the dot plot where each dot shows where a Fed official thinks Ritz would be at the end of the year, you'll see that a majority were guiding towards a December rate cut.
今天，让我们来谈谈两件事情。首先，我们来讨论一下为什么市场预计12月降息的概率只有50%，以及人工智能叙事中有哪些漏洞，从美联储开始说起。现在让我们稍微回顾一下。在9月的FOMC会议后，市场预计12月降息的概率是100%。为什么呢？因为美联储在9月的指导基本上指向了这个方向。记得在9月时，我们看到了一个点阵图。如果你查看点阵图上的点，每个点代表一位美联储官员认为年底利率会在哪里，你会发现大部分官员的意见指向12月会降息。
But not all, it was a very big majority. You had a sizable plurality that thought that maybe we just need one or two cuts this year. But in any case, the markets saw the dot plot and was basically thinking that, yeah, we're going to get a December rate cut. Moving on to the most recent October meeting, Pawe had very, very memorable and strong words. Here's what he said. In the committee's discussions at this meeting, there were strongly differing views about how to proceed in December. A further reduction in the policy rate at the December meeting is not a foregone conclusion far from it.
并不是所有人都这样想，但绝大多数人是这样的观点。还有相当一部分人认为可能今年只需要一到两次降息。但无论如何，市场看到点阵图后，基本上认为我们将在12月进行一次降息。接下来谈谈最近的10月会议，鲍威尔说了一些非常难忘且有力的话。他是这样说的：“在本次会议的讨论中，关于如何在12月采取行动，委员会内部意见严重分歧。在12月会议上进一步降低政策利率并不是板上钉钉的事情，远非如此。”
So Pawe was basically strongly telegraphed into the market that he did not have enough support, at least at that time, to guide towards a December rate cut. There seems to be a lot of vocal opposition toward a December rate cut. And so he basically pushed back against it. And so December rate cut odds gradually faded. Now this past week, we got a lot more Fed speakers, specifically a lot more Fed presence speaking who were sounding a very hawkish note.
帕维基本上已经向市场明确表示，他在当时缺乏足够的支持来推动12月降息。似乎有很多反对声音对12月降息持反对意见，所以他基本上对此进行了抵制。因此，12月降息的可能性逐渐减弱。在过去一周，我们听到了更多美联储官员的讲话，他们的言论显得非常强硬。
Now what they were saying is basically that, you know, I don't think we should cut rates. We should basically wait and see from their perspective, they were worried about inflation and they felt that the labor market was okay. Now if you look at the data, again, this is going to be out of data a little bit, but it's the most recent official data we get from the government, you can see that the Fed is in kind of a difficult situation.
他们基本上在说的是，从他们的角度来看，他们不认为我们应该降息。他们觉得应该先观望一下，因为他们担心通货膨胀，并且认为劳动力市场还不错。现在如果你查看数据，虽然这些数据有点过时，但这是我们从政府得到的最新官方数据，你可以看到美联储处在一个有点困难的局面。
On the one hand, inflation has been stuck around say 200.7% for some time. It really hasn't been going anywhere. On the other hand, you see the unemployment rate gradually creep up. So they are at a time where things are moving in opposite directions and they're not really sure where inflation will be because some of them are afraid that tariffs will actually move inflation higher. And they're not really sure how the labor market would perform because, you know, we have this changes in population growth.
一方面，通货膨胀率一直稳定在比如说200.7%左右，已经持续了一段时间，没有明显变动。另一方面，失业率却在逐渐上升。所以，目前情况是两者朝着相反的方向发展，人们不确定通胀会往哪个方向走，因为有些人担心关税可能会导致通货膨胀上升。而对于劳动力市场的表现也没有明确的预期，因为我们面临着人口增长变化的问题。
Whereas the weak job numbers we saw earlier in the year could be in part due to just lower population growth. And the third dimension that they're really not afraid of, not sure of is how restrictive is monetary policy is say 4% through point 8% Fed funds rate. Is that restrictive today? Some people will say that, you know, the economy has changed between now and the pre pandemic world. And so where we are now, not super restrictive and you have other people would say that, you know, the structural forces that kept the neutral rate low are still here.
今年早些时候我们看到的就业数据较差，这部分可能是由于人口增长放缓。而他们真正不确定、不太担心的第三个方面是货币政策的限制性程度。比如说，4%到4.8%的联邦基金利率在今天是否算是限制性政策？有些人会说，经济已经从疫情前的世界发生了变化，因此我们目前的利率并不算非常严格。但也有其他人认为，导致中性利率保持较低的结构性因素依然存在。
And so we are still pretty restrictive. So we have all sorts of uncertainty. And on top of that, different risk tolerances, some people are just have a lot more sensitive to inflation. Being above target since it has been above target for a number of years. So all this confusion is ending up to the market just saying that, hey, 50-50% of a Fed December rate cuts. And a lot, it seems at least part of the enthusiasm in equity markets has been due to the fact that, hey, markets are kind of in a mania, Fed's actually going to be cutting into it.
我们仍然相当保守，因为存在各种不确定性。此外，不同的人对于风险的承受能力不同，有些人对通货膨胀特别敏感，因为已经有好几年通货膨胀率高于目标。因此，所有这些不确定性导致市场上流传着美联储在12月有50%的可能降息的说法。而股市的热情部分似乎归因于这种想法，即市场有点疯狂，而美联储可能真的会在这种情况下降息。
And for sure, expanding their balance sheet. So let's party on. Now this is actually historically quite strange. Usually when we have these big mania, let's say during the dot com or during the 1920s, we have the central bank actually raising rates. Not so much to proud of pop the bubble, although that may play a role in it. But usually when you have strong stock markets, you also have a strong economy, it just doesn't seem to be the case at the moment. Now for my perspective, I think all this hawkish talk is really, really strange because, you know, if you were kind of dovish in September between then and now, we really haven't had a lot of official data. So you haven't had good data to change your mind.
当然，他们正在扩张资产负债表。所以让我们继续庆祝吧。不过，从历史上看，这种情况其实相当奇怪。通常，当我们遇到像互联网泡沫或1920年代这样的大泡沫时，中央银行会提高利率，尽管这样做不完全是为了刺破泡沫，尽管这可能会有一定作用。通常情况下，强劲的股市伴随着强劲的经济，但目前似乎并不是这样。从我的角度来看，所有这些鹰派言论真的很奇怪，因为如果你在9月份还是比较鸽派的，从那时到现在，我们并没有太多官方数据。因此，没有好的数据来促使你改变主意。
And I think just as importantly, the government shutdown itself is a headwind, right? You had all these government workers who were doing anything, industries that were reliant upon their patronage, not doing anything, and all these flight disruptions that are impacting airlines and so forth. Now, a lot of the economic slowdowns from the government shutdown will reverse as government workers get back pay and spend that money, but some of that will be lost and never come back, right? If you were a government worker, let's say buying a sandwich at a local restaurant by your office every day, you're not going to go back and buy two sandwiches instead. That restaurant owner is just never going to get that last revenue back.
我认为同样重要的是，政府停摆本身就是一个阻力，对吧？很多政府雇员无法正常工作，依赖于他们服务的行业也停止运转，还有所有这些影响航空公司等行业的航班中断问题。虽然政府停摆导致的许多经济放缓会在政府雇员收到补发工资并花费这笔钱后反弹，但有些损失是无法弥补的。比如说，你作为一名政府雇员，可能每天都在办公室附近的餐馆买一个三明治，但停摆结束后，你不可能回头去买两个三明治。那个餐馆老板失去的这部分收入就永远找不回来了。
So if you had these headwinds, it would seem that you would, you know, at least continue your dovish bias. There's certainly no reason to switch to hawkishness. And yet we do have some Fed presence who seem to have switched. Now, at the end of the day, though, the most vocal hawks are Fed presence and many of them don't even vote at the moment. The more important people, the Fed governors, there's still a very strong contingency there that are leading dovish, Governor Moran. I would assume that the other Trump appointees are also leaning towards a cut in December. I suspect that John Williams, President of New York Fed, is as well. And I think that Pablo probably is leaning towards a cut as well.
所以，如果存在这些阻力因素，你应该至少继续持有鸽派倾向。显然没有必要转向鹰派立场。然而，确实有一些美联储主席似乎已经改变了立场。不过，总的来说，最激进的鹰派主要是一些美联储分支主席，他们中的许多人目前甚至没有投票权。更重要的是，美联储的理事们，其中仍有很强的鸽派倾向，比如莫兰理事。我想，特朗普任命的其他理事也倾向于在12月降息。我猜纽约联储主席约翰·威廉姆斯也是这样。我还认为巴勃罗可能也倾向于支持降息。
So at the end of the day, I think it's still likely that we would get a December cut, especially if the data that comes in is weaker than expected. One thing I'll also highlight is that we have this brand new research out of the San Francisco Fed that Nick Timoros of the Wall Street Journal highlighted. And that's showing that tariffs are actually disinflationary and raise unemployment rates looking through 150 years of history. And the reason, of course, is really simple. Tariffs are basically a tax. When you increase taxes, that slows down the economy, which, of course, you have less demand. That means unemployment rate goes up and inflation goes down, which usually a combination would lead to more monetary policy easing.
总的来说，我认为我们仍然有可能在12月降息，特别是如果即将发布的数据比预期的更疲软。需要注意的是，《华尔街日报》的Nick Timoros强调了旧金山联储的一项全新研究。这项研究表明，从150年的历史来看，关税实际上具有通缩作用，并提高失业率。原因当然很简单，关税本质上是一种税收。当税收增加的时候，经济增长会放缓，因为需求减少。这意味着失业率会升高，通胀率会下降，通常这种情况会导致更多的货币政策宽松。
So all these talk about tariffs pushing up inflation so that the Fed should say more hawkish, that's just not supported in data. And so far, I think more and more people on the FOMC are becoming to realize that the hawks, not really sure what they're looking at, but a lot of time in the data for them to change their mind. All right, the second thing I want to talk about is cracks in the AI narrative. So again, as we've talked about before, AI is the thing that's driving markets higher, companies are spending hundreds and hundreds of billions on it. Not really clear how they're going to make money off of it, but that doesn't stop the equity market.
所以，所有关于关税推动通货膨胀上升的讨论，认为美联储应该更强硬，这在数据上并没有支持。而且到目前为止，我认为越来越多的联邦公开市场委员会成员开始意识到，强硬派其实并不确定自己在看些什么，数据还有很多时间让他们改变主意。好了，我想谈的第二件事是人工智能叙事中的裂缝。再次强调，正如我们之前所谈到的，人工智能是推动市场上涨的因素，公司在这方面投入了成百上千亿美元。虽然不太清楚他们如何从中赚钱，但这并没有阻止股市。
Now the two cracks emerging in the equity market this year are increases in the CDS. So credit default swaps for certain AI companies suggesting some concern from bond investors about the AI narrative. And secondly, you have, I guess, more concern from the earnings perspective due to depreciation chargers. And this was raised by famous short seller Michael Blurry, who was of course famous for calling the mortgage bust in the 2008 financial crisis. There's a movie made about it as well. Christian Bell plays Michael Barry there. It's a good movie.
今年股市出现的两个问题是信用违约互换（CDS）的增加和贬值费用导致的盈利方面的担忧。首先，一些人工智能公司的CDS增加，表明债券投资者对人工智能前景存在一些担忧。其次，由于贬值费用，市场在盈利能力方面的担忧加剧。著名的做空者迈克尔·布雷（Michael Burry）也指出了这一点，他因在2008年金融危机中预测房贷崩盘而闻名，并且还有一部电影讲述了这个故事，克里斯蒂安·贝尔在电影中扮演迈克尔·布雷。这是一部好电影。
Now first let's talk about the cracks in the AI from the bond market. So over the past few weeks, we see that the CDS spreads for Oracle and for CoreWeave, which are two companies that are engaged in the AI build up boom, basically widened significantly. And what that what a CDS is, it's basically insurance premium for default on their bonds. And so when CDS spreads widen, that means investors are demanding, that means insurance premiums are going higher, that means investors are getting more concerned about these companies defaulting on their debt. Now Oracle we've talked about before, they had a just a tremendous mind-blowing rally from their most recent earnings report.
首先，让我们谈谈债券市场中与AI相关的问题。在过去几周里，我们注意到Oracle和CoreWeave这两家参与AI建设热潮的公司，其信用违约掉期（CDS）利差显著扩大。CDS是一种债务违约保险，也就是说，当CDS利差扩大时，意味着保险费率上涨，这表明投资者对这些公司可能违约的担忧加剧。之前我们谈到过Oracle，他们在最近的财报公布后股价经历了一次令人瞠目的大涨。
At that earnings report, they guided towards tremendous amounts of AI related build out. They were going to build all these data centers and sell this compute that they had contracted out with AI companies. And so massive AI build out, massive AI contracts for compute, and they basically went to the moon. And these are all promises in the future. And if you were to find, to be a bit skeptical, I think you'd be understandable. But in any way, in order to provide all that compute, obviously they have to spend a lot of money to build out data centers, part of that money was financed in the debt market. So they went and borrowed some money and started, are starting to build data centers.
在那次财报中，他们表示将大力拓展与人工智能相关的建设。他们计划建立许多数据中心，并将向人工智能公司出售已签约的计算能力。这意味着巨大的人工智能建设和大量的计算合同，基本上可以说是前途无量。然而，这些都是对未来的承诺。如果你持怀疑态度，我认为这是可以理解的。为了提供所需的计算能力，他们显然需要投入大量资金来建设数据中心，其中一部分资金通过债务市场融资。因此，他们借了一些钱，已经着手开始建设数据中心。
Now if you are a bond market investor, financing that debt, again, you're going to be a bit more skeptical because in order for you to be repaid, that data center spending actually has to generate revenue that covers its financing costs. So bond market investors, they're really cared about being repaid. They don't have that to the moon upside that equity investors would have. And so it seems like some of the bond market investors are becoming concerned that, hey, maybe these AI data centers might not generate enough cash flow to repay me. And so I'm getting a bit concerned again, that's a crack.
如果你是债券市场的投资者，你为这些债务提供资金，你可能会更加谨慎。因为为了确保你能被偿还，这些数据中心的支出必须产生足够的收入来覆盖融资成本。所以，债券市场的投资者非常关心能否被偿还。他们不像股票投资者那样期待股价大幅上涨。因此，有些债券市场的投资者开始担心这些AI数据中心可能无法产生足够的现金流来偿还他们。这种担忧就好比市场中出现了一道裂缝。
And the belief that AI is just going to be this huge thing that's going to change the world. A similar company core weave, which also of course is in the business of building data centers, also saw their credit spots widen. So that is some crack in the AI narrative, but I have to remind everyone that the bulk of the AI spending is financed from the hyperscalers like meta, Google, Amazon and so forth, Facebook. And these guys are paying for most of their AI spending from their cash flows. And they generate enormous amounts of cash. They are the actually some of the world's most successful companies, generate huge amounts of cash.
人工智能将会成为改变世界的重要力量的想法依然受到人们的关注。类似的公司CoreWeave，也是一家从事数据中心建设的公司，他们的信用差价也扩大了。这显示出人工智能叙事中的一些裂痕。但我要提醒大家，大部分的人工智能支出是由大型科技公司如Meta、谷歌、亚马逊和Facebook等进行资助的。这些公司主要以其现金流支付人工智能方面的开销。他们产生了巨大的现金流，是世界上最成功的一些公司，创造了庞大的现金收入。
They're taking that cash and building AI stuff. Some of it, they have some debt as well, but of course, they generate so much cash in such exceptional businesses, not something to worry about. So the bigger picture is that you have some cracks on the periphery, but really the big part of the AI story, the max seven, the hyperscalers, these guys, they're still fine. You can make an argument that if you spend cash on data centers, you don't have cash for stock buybacks, but it seems like they are saving money by cutting headcount so far anyway. And of course, we get to hear from Nvidia this week, which will really help us understand the AI narrative better.
他们正在把资金投入到人工智能项目上。一部分资金用来偿还债务，不过不用太担心，他们的业务非常出色，能够创造大量现金。从大局来看，虽然外围有一些小问题，但在人工智能的核心领域，像Max Seven和超级规模化公司，这些大企业运作良好。有人可能会说，如果把资金用于建立数据中心，就没有钱回购股票了，但似乎他们通过裁员来节省开支。我们将在本周听到来自Nvidia的消息，这将帮助我们更好地理解人工智能的发展。
The second crack that's emerging has to do with whether or not these tech companies are popularly accounting for their depreciation expense, and this was raised by Michael Blurry. So what is depreciation expense? So if you buy a physical asset, let's say that's actually you buy a rental home, it costs $500,000 to buy it, and every year you get rental income, let's say $50,000 in rental income. So when you for tax purposes, that $50,000 in rental income isn't pure income because you have costs to gain it. You have actually you bought a house and that housing cost has to be spent over the lifespan, the useful life of the house.
第二个问题是关于这些科技公司是否如实列出它们的折旧费用，这一点由迈克尔·布里提出。那么什么是折旧费用呢？举个例子，如果你购买了一项实物资产，比如说一套租赁房，花费了50万美元，每年可以获得5万美元的租金收入。在进行税务申报时，这5万美元的租金收入不算作纯收入，因为你为此有费用。你实际上购买了一套房子，而这个房子的购买成本需要在房子的使用寿命内进行分摊。
Now the useful life of the house, it's going to be from accounting purposes, let's say that it's 20 years, 25 years, in any case, you divide the deprecial part of that house over a depreciation lifespan, and that is your annual depreciation cost. That's costs that you incurred in order to generate that $50,000 in rental income. And so when you have higher depreciation, you have lower income because you're allocating more of that real asset cost to the income, to against the income. Now at the moment, a lot of these guys are buying a tremendous amount of GPUs and are giving them a depreciation schedule of about five years.
现在从会计的角度来看，房子的使用年限可以假设为20年或25年。无论具体多少年，你需要将房子的折旧部分分摊到整个折旧年限中，这样就能计算出每年折旧费用。这是一种用来获取房租收入（比如5万美元）而产生的成本。因此，当折旧费用提高时，收入就会相对减少，因为你将更多的实物资产成本分配到了收入上。与此同时，目前很多人正在购买大量的图形处理单元（GPU），并且为其制定了大约五年的折旧计划。
So if you spend $100 on your GPU, let's say that every year you're recording $25, $20 of cost against that. So let's say you made $100 in revenue from your AI data center, your own cost is your GPU, you record $20 of GPU cost depreciation against that. That's basically saying that you had to buy a GPU to get that revenue. That's a cost. The GPU has a useful lifespan about five years. And so you're just saying that, hey, let's do a $20 cost a year. The longer the lifespan of the GPU, the smaller depreciation you get each year. And so the bigger your earnings look.
如果你在显卡（GPU）上花了100美元，假设每年你记录25美元或20美元的成本。那么，如果你从人工智能数据中心赚了100美元，你需要记录20美元的显卡折旧成本。这基本上是在说，为了获得这些收入，你必须购买一块显卡，这是一个成本。显卡的使用寿命大约是五年，所以你每年分摊20美元的成本。显卡的寿命越长，每年分摊的折旧就越小，这样你的盈利看起来就越高。
Now what Michael Blurry is suggesting is that the GPU lifespans that are being proposed or being used by these big tech companies are too long. And because they're too long, the depreciation charge is too low. So they're not properly accounting for the expenses they are incurring in getting their revenue and artificially boosting their profits. Now that's a large argument. So when I look at computers, when I look at GPUs, usually technology progresses pretty quickly. Five year lifespan, probably not super reasonable. Maybe it's three years, maybe it's four years, it's hard to say. Nvidia is going to come up with a new GPUs, say every year. And so maybe your old GPU becomes absolutely faster.
现在，Michael Blurry 的观点是，这些大型科技公司所提出或使用的GPU使用寿命太长了。由于寿命定得太长，折旧费用就偏低。因此，它们在获得收入时没有正确地计入相关支出，从而人为地抬高了利润。这是个较大的争论点。通常我看计算机和GPU时，技术进步都相当快。五年寿命可能不太合理，可能是三年，也可能是四年，很难说。比如，Nvidia可能每年都会推出新的GPU，这样一来，你的旧GPU可能很快就会变得相对较慢。
Now, other people would say that even old GPUs are still used out of 100%, because they can do, you know, similar tasks. And so a five year lifespan is appropriate. Again, this is all hard to say. But if you think that depreciation is too long, then you would assume that they are overstating their profits. And then the stocks, you know, if the profits are overstated, that should be negative for their stocks. But of course, we all know better, right? Stocks can have no earnings and still go to the roof. But if you are a fundamental investor like Michael Blurry, again, you would point towards things like that. And that is totally legitimate, totally valid.
现在，有人会说，即使是老旧的GPU也能被充分利用，因为它们可以执行类似的任务。所以认为它们有五年的寿命是恰当的。不过，这个问题本身就很难判断。但是，如果你认为资产折旧年限太长，那么你可能会认为公司高估了利润。这种情况下，利润被高估对股票价格应该是负面的影响。但我们都知道，股票有时候即使没有盈利，也能大幅上涨。不过，如果你像Michael Burry这样的基本面投资者，你仍然会关注这些问题。这种观点完全合理和有效。
But earnings don't really seem to matter all that much at the moment. And so these concerns about the overstating of the GPUs and of the credit concerns in the AI narrative seem to have shaken confidence a little bit as well. So that seems to be what's happening. Now, next week we have a lot to look forward to. We have data that was supposed to be released. We're going to get the September NFP data. And we're going to get Nvidia earnings.
目前，利润似乎并不是特别重要。因此，关于GPU夸大其词和AI叙事中信用问题的担忧，似乎让人们的信心有所动摇。这就是目前的情况。不过，下周会有很多值得期待的事情。我们将看到一些原定发布的数据，包括9月的非农就业数据，以及英伟达的财报。
Now, before we get too bearish on the market, we do seem to be losing a little bit of momentum. You know, you could have a very clear path where say that Nvidia is totally okay, as it usually is, reaffirming the AI narrative. And you could have data come in weaker than expected going forward. That puts a December rate cut back on the table. And then, you know, you could easily see the seasonal effects kick in and see us at 7,000 by the end of the year. That is, in my personal view, probably a pretty likely path.
现在，在我们对市场过于悲观之前，我们似乎有点失去动力。你知道的，情况可能很明朗，比如说Nvidia一切正常，正如往常一样，这也支持了人工智能的发展。同时，未来的数据可能比预期的要弱。这种情况下，12月的降息预期可能会被重新提上议程。然后，你可能会看到季节性效应发挥作用，今年年底前市场可能达到7000点。在我个人看来，这可能是一个比较可能出现的情况。
And of course, we could also go down. So in any case, next week will be very important. Let's watch Nvidia and let's see what the data brings. All right, talk to you guys next week.
当然，我们也可能会下跌。所以，无论如何，下周都会非常重要。让我们关注英伟达，看看数据会带来什么。好了，下周再和大家聊。