Recomendações
Entrada é o preço de fechamento do ativo na data de publicação. Atual é o último fechamento registrado.
-
Entrada $661,75 16 nov 2026Atual $608,45 20 nov 2026Resultado −$53,30vs. índice −1,3% SPY −6,8% no mesmo período
Citar esta recomendação Ver trecho original (1452s) *
Contexto da transcrição original
…tlook principally due to energy. However, we expect these headwinds to recede by the end of 2027. More inflation being persistent is a clear risk uh that the Federal Reserve will have to deal with. So, my take I know it sounds kooky dukes. The Q's are now literally at 666. I know. I know. But it's simple. Stay out of margin. Stay out of debt. Buy the dip, baby.
The Q's are now literally at 666. I know. I know. But it's simple. Stay out of margin. Stay out of debt. Buy the dip, baby.
Transcrição Completa
Oh, this story is just getting wild. Donald Trump, who's shilling for Jensen, it turns out, has a stake worth $62 million in Nvidia. Or at least it was worth that at the end of August when they last reported it, which means we're down a little bit. The stock market is selling off and it's all because of Jensen Swang. I mean, Jensen Huang. Anyway, the stock market is selling off. Part of it is because people believe that Jensen got scoped by the New York Times. We'll talk about that towards the end of this video. And what's more important is that potentially the most bullish research firm on AI chips is all of a sudden cutting its forecasts for spend between now and 2030 by $1.6 trillion. In other words, an extra 1.6 trillion they had forecast gone. Zero, baby. This is insane. And it's all coming right after the Jensen disaster, which makes you wonder how many people knew that we were going to have insiders start rallying against Jensen. I mean, you've literally got this bullish AI research firm now cutting prices for Nvidia by up to 60%. Way beyond what other Wall Street firms are doing. Makes you wonder how many puts they bought on Nvidia. And then you've literally got Scott Bessant running out an anti- Nvidia campaign against Jensen Huang, suggesting that just 1% of the Nvidia chips sold to Singapore actually stayed in Singapore. Which then of course raises the question of why didn't the government actually intervene or do anything about it? probably because the government was secretly uh or pretty overtly going, "Yeah, just go ahead and keep selling. It's good for GDP." But anyway, in fairness, Singapore was up to 20% of Nvidia's sales in 2024. And now Nvidia is redefining where these sales are going to end up being registered based on the headquarters of companies, not necessarily where the chips are showing up. maybe to obfiscate some of where these chips are actually going. This is reiterating why people like Bessant on Fox News are like, "Nvidia needs a shakeup. The next CEO needs to come from outside the company." I can't help but think Scott Besson is fearful for his job. That if for some reason Donald Trump happens to support Jensen and the finger gets pointed at either Letic Besson that they're at the firing line. So Besson kind of has to act like, "No, no, man. We we we were trying to stop in video and video bad. It's kind of interesting, especially since about an hour and a half ago, Donald Trump came out and said, quote, I certainly wouldn't mind seeing Jensen back. And that Jensen has learned a huge lesson. This is coming at the same time as employees have discussed a demonstration at the Santa Clara campus. We'll see if there's a demonstration out at their headquarters. 175,000 people have signed the petition to bring Jensen back and stand with Jensen. they actually had the traffic or traffic to the website was so overwhelming that the website was starting to struggle. Uh in fairness, there are some pieces coming out that are suggesting that uh you know Jensen was also trying to limit the quote open checkbook that the days of the open checkbook of Nvidia financing were uh coming to an end anyway so that you know people should stop relying on Nvidia's circular financing. So, you know, some bare news on the AI cycle, but I think the biggest bare news isn't Jensen saying to the Neoclouds, "You're not charging enough." He said, telling them that Elon Musk was renting Nvidia chips to customers at better chips than they were. I don't even need you guys. I could just give all this capacity to Elon. Elon wants it. Which then also makes you wonder, is Elon underwriting the entire AI boom? Right? But let's put this aside for a moment and understand what just happened with this semianalysis drop because everybody on finit is freaking out about what happened with semianalysis. Semi analysis basically massively cutting their forecasts. This this is circulating everywhere right now. Uh this is what people are looking at. semi-analysis capex estimates. They're basically keeping 2026 unchanged to save face. They're reducing 2027 capex estimates from 770 billion to 627 billion, a 19% drop. 2028 1.05 trillion down to 758 billion, a 28% drop. Then 1.35 trillion for 2029 is getting dropped to just 796 billion. and 1.53 trillion for 2030 is getting dropped 40% to 923 which what's worth noting is that the second derivative here plummets. This growth in capex between 2028 and 29 is only 5% per this analysis. Now here's the problem. Okay, semi analysis knew the vast majority of this crap before they just all of a sudden decided to change their mind. They literally just 11 days ago reiterated their numbers about how bullish they were and how everybody else was too not bullish. So, let's be real about this. What actually changed in the last 11 days? Because that's what matters here. If maybe big things changed in the last 11 days, maybe this could be fair. I mean, Jensen did quit. We did have a dem sweep. Are those the things that changed? Well, let's analyze this a little bit. When it comes to what happened in physical hardware, nothing changed. Literally nothing changed. You have companies across the board still ordering, placing orders, and growing backlogs at either the same or greater pacing than what we had before 11 days ago. So either semi analysis is turning into a weenie baby and they're just fearful because the stock market is going down and they're trying to tell everybody, "Yeah, see our analysis actually justifies why the stock market is going down." Or they're just being dishonest about how good their models are. Because apparently if you've got to adjust your models from one and a half trillion or or you know you got to cut $1.6 trillion dollars out of your models in 11 days, your models must not be that good. Now yes, midterm elections happened. But midterm elections are expected to only delay data center buildouts by about 5.8 gawatt between 2027 to 2028. If you put 50, call it $60 billion on 5.8 28 gigawatt, you're talking about a delay of maybe $340 billion of spent from 2027 to 2028. Like delaying that over basically that's not eradicating $1.6 trillion all of a sudden. So I think what happened is semi analysis woke up from a really bad hangover or with a really bad hangover. I think they realized they were drunk with their numbers and you had this like ridiculous Tom Lee style version of a semi bull semi analysis bull in chips and then they woke up and realized, wow, we were drunk. We should revise because now the stock market's falling and we don't want to seem like we're drunk while the stock market's going down. In other words, I honestly think they're just adjusting their numbers down to make it seem like they're not drunk and that they're just following the market just like Wall Street analysts do all of the time. Now, how do we know that? Well, because when you actually look at the details of what they say, they literally walk back their fantasy perchip power numbers from 12,200 watts to 16,000 uh watts per socket. that all of a sudden can't be plugged in anywhere. They just they just cut this by like 50%. They're like, "You know what? We're going to cut how much power gets plugged in by 50%." Then they assume that memory prices are going to get cut by 18 to 30%. Which frankly, most people on Wall Street were already forecasting that memory prices were going to slow down way before this election. And then they decide they're going to cut Nvidia's future margins by assuming that OpenAI, Anthropic, and Amazon are just magically going to only use in-house cheaper AS6, which ASIC competition is real, but that Nvidia GPUs are basically going to get thrown in the trash. That's how I'm I'm being obviously a little bit extreme over here. But there's only one way you can actually in 11 days say Nvidia's prices are going to go down 44%. And ASIC makers like Broadcom's prices are only going to go down 6%. In my opinion, the only way you reconcile that is fear, mood, and credit. That's it. The VIX spiked to 28 on November 5th. Then we got downgrades after Jensen cheated on uh uh you know November 11th and 12th and 13. We got multiple uh downgrades. We saw CDS's rise and peak in October for uh you know financing for the NeoClouds like Cloudflare uh you know even meta CDS's went up and all of a sudden the stock market really started falling in conjunction with Donald Trump not getting a deal done with Iran. the Federal Reserve becoming more hawkish. Putting all of that together with a dem sweep makes the stock market go down. And I think semi analysis is fearful of looking like a fool. So, they're going in and cutting their ridiculous analysis and making it try to seem a little bit more sober to justify why the stock market's going down. But I really want you to think about for a moment. Think about all of the crap the stock market has had to digest. The Fed holds in October because they think Donald Trump's going to make a deal with Iran. Donald Trump doesn't make a deal with Iran. Fed has to go hawkish. That's one thing. No deal with Iran. Hawkish Fed. Inflation. All that's bad. Inflation numbers come in hot. The labor market comes in strong still. None of that's recessionary. It's just a sign that we've got an energy crisis. Okay. On top of that, you get the Dem sweep, which in fairness, that's going to delay probably some data center buildouts to future years. Not as much as semi- analysis is saying now, but that adds. So now you've got Iran, uh, no deal. You've got Fed hikes, you've got data center delays from de Dem sweep. Add on top of that the Jensen cheating scandal. Holy moly, dude. You've got a lot of pain that this market now has to absorb. But like when you actually look at what semi analysis is saying, their number changes are loony bins. They literally remove $1.6 trillion in spend. And how do they do that? Well, they assume that they've got to take big cuts, big acts to revenue for Nvidia in 27, 28, 29, 2030. They are lowering that entire stack by 32%. Morgan Stanley literally has 1.7x the semianalysis numbers for Nvidia. And in their analysis, Nvidia takes the vast majority of their hit or the hit. They are literally cutting Nvidia's data center revenue for 2030 by 66%. I swear these people have like puts or something on Nvidia. But anyway, 2/3 of this because of Nvidia is what they call lost silicon area. In other words, all of a sudden we're going to see massive pricing power declines because we're going to shrink silicon area. Now, little complicated. I'll make a very simple analogy here. But basically, I like to use a kitchen analogy. Think about the kitchen as the kitchen of your home. The more square footage you have, the more chefs you can have. Chefs are like transistors doing AI work. Okay? The bigger your pantry is, the bigger your refrigerator door, the more bandwidth you have to get stuff done. The doors are sort of a little bottleneck. And then how much power you have going to your fridges or your stoves like gas is coming in means how much you can actually cook, right? the AI industry as a whole right now, even silicon analysis says or uh yeah, semi-analysis semi- analysis says, "Oh, we're going to only shrink our forecast by about 15% for how much area we're going to get in kitchens." So, all everything's going to be the same except we're just reducing our kitchen sizes modestly because of the dem sweep or whatever by 15%. Yet, we're going to assume that power consumption is down 50%. And the way we're going to reconcile this is we're going to say Nvidia is going to take the biggest hit. To me, this is not evidence of anything that's happened in the last 11 days other than semi analysis waking up and realizing, wow, Nvidia stocks going down. How are we going to justify that Nvidia stock's going down? Why didn't our models tell people upfront that Nvidia stock was going to go down? Oh, uh, why don't we just go blame Nvidia now? And so that's what you're getting. I mean, you take a look at the chart that we get up here. Nvidia gives up 40% of its markup. Okay, chips all of a sudden are going to see a massive power cut. And that's how we get rid of $1.6 trillion of spending cut between 2027 uh to 2030. And that all of a sudden Nvidia's chip revenue is going to go down 66%. That's the bare story today. That's why the markets are red today. This is also relatively why even other companies are getting hit. Semi- analysis comes out and says Broadcom's going to have to cut their prices by 6% and AMD is going to have to cut their prices 10 to 15%. So across the board, they're trying to become a little bit more reasonable with their numbers, but they're coming from this crazy high state. Anyway, semi analysis argues that up to 30% of their cut is because of memory price declines. Now, Bernstein argues the opposite. They actually think that high bandwidth memory contract prices are roughly going to double in 2027 and that memory prices are still going to be a a major driver of semiric growth. But semi analysis is now taking the flip argument and they're saying no high bandwidth memory costs are going to come down. That's going to squeeze margins that premium that you're able to charge for chips like Nvidia which are basically packaging the high bandwidth memory and reselling it at big premiums because of those memory cost increases and that A6 are going to take over. Fine. All that sounds really complicated. The basic argument here is that a big research firm that people see as the big bull on Wall Street is now coming out saying, "Hey guys, we're going to be less drunk." Uh, and part of the reason we're going to be less drunk is because we think that, you know, Anthropics growth is slowing more than we thought before, which is really interesting because Anthropic's growth rate was already declining uh before we got to this. So, that's that's where I get so confused as to why these folks all of a sudden are changing everything now. See, 11 days ago, we already knew that Anthropic's growth rate wasn't going to be like it was last year. The big growth happened in Q1, Q2 of this year. So, the year-over-year numbers, of course, are going to lap and look lower. We knew that before 11 days ago. We also knew that financing got tighter before 11 days ago. Credit default swaps peaked in October. We already knew those two things. The only thing that changed is we see a delay of 5.8 8 gaw of data center, you know, buildouts basically. And that's a delay, not a cancellation. We don't see cancellations or delays from Meta, Microsoft, Google. They're all spending like crazy. If anything, they're spending and borrowing more. So, two of the three justifications they give for changing their numbers, we already knew when they were reiterating their bullish numbers. The election happened, but that doesn't have that big of an impact. So really what you have is you have a company here that's like, "Hey guys, the stock market's going down, so we're just going to follow the market down with our analysis." That's what I think is happening here. And and I mean, you know, don't sue me, bro. I'm not trying to dump on them. This is my opinion. I could be wrong about this, but I honestly think this is an analysis firm that got caught being too bullish, and now they're trying to justify why Nvidia is down 34% from its peak. And this is how they're doing it. by using mostly old information that the market already knew and then trying to magically say, "Oh yeah, Nvidia is going to take the brunt of this whatever." Now, I don't care. I have very, very little exposure to Nvidia. I sold the vast majority of my exposure to Nvidia last year because I had to pay a big depreciation recapture bill and I had these massive gains on Nvidia and so I sold millions of dollars in Nvidia. Makes I have such little exposure to this stock. It really doesn't matter. I'm not I'm not trying to shield for Nvidia, but it does bring up what's going on with this whole idea about oh well, you know, the AI labs like now Microsoft data is showing that Claude is stalling against Copilot. Well, this is really good advertising for Microsoft, especially as anthropic is going into their IPO week. See, Claude for Microsoft is less than onetenth the number of weekly users of the paid version of Copilot is what Microsoft is trying to say. They're basically trying to say fewer people are using Claude and more people are using Copilot. For some reason, I have a really hard time believing this. Uh semi analysis also says that, you know, oh, now more policy is going to get passed like the Remote Access Security Act and you know, now the odds that that passes rises. that's gotten stalled twice already before. Every no vote was Republicans on that. So yeah, is it possible that it passes now because Democrats are going to control the Senate? Maybe. But Trump will just veto it. And I feel like they don't even consider that. So that's, you know, trying to restrict restrict chips going to, you know, foreign countries basically. So I I really take issue with this. Now, regarding the SCOP and the anthropic IPO, this is an interesting one. There is now analysis that US intelligence basically Marco Rubio's memos from the CIA pushed out Jensen Huang. But the whole reason it all got publicized was because basically the argument is the Jewish people who lead the New York Times decided to coordinate this to beat Jensen out of Nvidia. Now you might ask, what do Jewish folks have to do with this? Why are we all of a sudden talking about Jews? Well, here's the thesis, and it's mostly a conspiracy, I think, almost completely. But anyway, so Jack Clark of Anthropic was reportedly the one who leaked the Jensen spy story to the New York Times. But during Jensen's October tour, he publicly voiced support for open weight, basically anti-anthropic and open AI. This, people believe, is a threat to Israel's control over the entire AI industry since they have people in place at the head of anthropic, open AI, and Palanteer. That's this is very conspiratorial. Uh I just want to address this. I don't think that this was an Israeli scop to oust Jensen because he's pro openweight. Uh but I want to address it because this is what some people think is really going on. So now what happens next? Tuesday, November 17th, Nvidia reports. That's tomorrow without Jensen. It's going to be big. Okay. I kind of expect they're still going to beat and raise and they're still going to say the backlog is growing like crazy. We'll see. We'll cover it tomorrow. It's going to be interesting. November 25th is the 10 days from Bessant's call for answers. Thursday, Anthropic IPOs. If anthropic tanks, the whole market's going to keep tanking. Okay, we've got some forecast that we've been talking about in the course member livereams. Hope you join us over at reinvest one.com. December 3rd is the date tentatively that Tim Scott wants Jensen Hang to testify in Congress. We do not know if that is going to end up being public or behind closed doors. And then December 8th to 9th, we'll get the next Fed rate decision. So, how do I reconcile all this? Personally, when I put all of this together, personally, I think that semi analysis adjusting their numbers down is being done to save face and argue that they knew the sell-off was going to happen. Total bull crap. I don't trust these folks. I look at their stuff and then go drunk. I have said it before. Say it again. I don't look at their numbers. I trust what I see it. earnings comp uh earnings from companies, earnings calls and forecasts from analysts that are actually talking to the very companies on the ground. The semi-analysis numbers, I don't know if they're getting paid by Leopold or whatever. My opinions, I don't I don't like them because I feel like they've been too blindly bullish and now all of a sudden they're following the market down. I think that's dishonest and disingenu like frankly disingenuous. I think that lacks integrity and I think it's honestly dumping on Nvidia when they're down. I'm not here to say, oh, you know, you know, buy the dip on Nvidia. I really don't care if Nvidia goes to zero. It makes no difference to me. Do I think Nvidia is going to go to zero, though? No. I actually think that we are setting up for a big bull run. I know, I know people say, "Oh, Kevin, that's delusional." I know, I know. Let me read you this, okay? Because I think it's very interesting. So, I hate, by the way, agreeing with institutions. I am really anti-institution, but I want to read you this. Okay, so this here says, Morgan Stanley, the AI capex cycle we view as intact, a resilient global economy, and solid earnings growth extending into 2027 provide a constructive foundation with major risk catalysts behind us. risk assets have cheapened over the past six months and we enter 2027 with a broadly positive stance on risk. I agree. I hate agreeing with them, but I agree. Here's another analysis. This one I think was JP Morgan. We do expect global growth to soften modestly in 2027 before returning to its 2026 pace in 2028 as energy headwinds recede. The US remains the principal engine of global growth. There are clear challenges to global economic outlook principally due to energy. However, we expect these headwinds to recede by the end of 2027. More inflation being persistent is a clear risk uh that the Federal Reserve will have to deal with. So, my take I know it sounds kooky dukes. The Q's are now literally at 666. I know. I know. But it's simple. Stay out of margin. Stay out of debt. Buy the dip, baby.
Comentários 0
Entre para participar da discussão.
EntrarAinda não há comentários. Seja o primeiro a compartilhar sua opinião!