CNBC & Fox Today On NVIDIA Stock, Micron Stock, Tech Sell-Off - NVDA Update

But how about the skeptic? The skeptic says that almost every day invid opens up and finishes down. >> Today it's down. >> Okay. >> Well, the skeptic says if it opens up every day and then finishes down, which is bleering >> that if it starts down that it'll end up. >> This is the first time I've seen Well, I'm saying the skeptic. In other words, there are lots of people here to say Kramer says. I'm not saying that. I'm making the observation that the stock opens up all the time and everybody's believer. What happens if the stock opens down? In other words, not not because they own compute. They're like Mr. Compute. >> They are they're also Mr. Bank. >> Yeah, they're Mr. Bank >> for when we say Mr. Bank, meaning that they are certainly helping to fund a lot of the buildout as well >> with various guarantees that they have provided for certain offerings and and things of that nature. But I'm just saying, look, this stock has been one of the worst stocks since >> this is this is the cheapest multiple on Nvidia about a decade. We're at 16 and a half now, >> right? And remember, we had that period of an unnatural moves, right? The June period and we uh the Koreans were buying and everyone was excited. that blowoff is uh it's still not >> but if there is a slowdown in data center construction that begins let's call it next year >> and extends will that result in fewer GPUs being sold >> u I think undoubtedly you have to say that it's going to be stretched out >> in the same way that if you were in the pipeline business and you woke up and Biden won I think you'd say you know what guys we don't need to make as much pipe >> well there's the there's also the dumber the double whammy Jim if there's a data center slowdown and the war ends. Defense competes for the same chips, the same power, the same industrial capacity as uh data centers. >> Like I said that like we're all writing off Nvidia, but it is the cheapest that it's been. We got to focus on that. >> Very cheap. Yes, Carl just told us 16 times. That is quite a low multiple even the balance sheet of that company is still great. Last I looked, Jensen Wong is still Jensen. >> Amazing margins, incredible free cash flow. Jensen Wong is still Jensen Wong. >> It's still Jensen Wong. >> There's a price to own it, but people would rather own Apple. Interestingly enough, I was I was on this call earlier this week. >> Thought it was pretty negative. It was the Apple call and uh they I was coming with this DM negative DR rap, you know, cuz like I I looked at the Micron and stuff, >> but they came out with the P, you know, like listen, D going to go up from now till king of K. >> It's still up 28 and a half% year to day. I say own it, don't trade it. >> It's above the street high target at this point. >> That is something you have actually consistently said. >> You have said consistently >> $5 Apple don't trade it. I will give you that >> what you've been saying on Micron. I have no idea. >> Okay. But I I I appreciate what you just said about me and Apple >> and I like Turtis. I'm just giving them a hard time because in truth, yeah, it was a remarkable quarter and I thought they were a little defensive and I wouldn't have been particularly when I can go to T-Mobile beginning today and get a duo for $800. >> Well, and if you're right about memory, then that margin story is not as >> the margin story is going to be fabulous. So, I am right on memory. >> Wait, you're right on memory. Why? Cuz data center construction is going to slow and then therefore we're not going to need as much memory. >> It's peaking the stocks leading >> peaking. it is >> in the outear shifts because there's a lot of look at the demand destruction in PCs. I'll sit I'll give you some DRAM. I got DM. You want some DRM? I got to get some DAM over here. I need you my D RAM filled with DAM. I got a special deal with the DRAM. I want to come back to the concept of of a conceivable short-term bottom in the stocks that have just been eviscerated. Okay. >> Like I really I you know I respect Nvidia greatly and I think at a certain point Nvidia represents some value. uh at a certain point I think micron represents some value but people decided you know what they can never reflect what what the people have spoken they can never reflect it enough they can all go down but I do think that these companies are not necessarily dead I think that they were they got too high in June during the blowoff they got too high they've been going down consistently now David the one I'm going to ask you about which just seems at this point to be unbelievable is Marll No. >> Yeah. >> What what what what's >> Well, it's funny. We talked about it in 10 yesterday. Obviously, you I mean, that just keeps going down. That again, this goes back to the idea of anybody selling into the data center conceivably is not going to be selling as much as quickly, right, Jim? Exactly. >> So, whether it's Marbell andor their merger partner or their the company they're buying or not, they are both relying on continued spending on data centers and selling into them. Look at what happened to Dell yesterday as well. It got absolutely crushed. Um >> this has been you know this has now taken on a life of its own in terms of it's not deal related. There's a look at Dell again today. Look at the sell off in that stock. That is right by the way right at the 200. Look at look at Wells Fargo. Okay. Look at JP M. You watch the banks. I mean what the Democrats like pro bankers lost 37% of its market value in a month. What? >> Yeah. >> Yeah. You don't need to say more than I mean servers plus PC. Oh, >> servers is really I mean that was one of the worst IPOs. That's like a you know that's kind of the benchmark of this Europe in terms of wow people can we have the IPO? See that's >> David see that chart. >> Yes. >> That that's suboptimal. >> It is suboptimal. David, I just want to take us to the current moment for a second uh in which certainly investors are trying to understand what that enormous change in Congress is going to mean for the future of data center buildouts. I'm curious as to your thoughts given conceivably it could impact your business over time. I mean, you're just talking about how important data center is going to be with you. I know it's early, but how are you seeing things unfold, if in fact you're seeing them unfold any differently than you were on Monday? >> Uh yeah, thank you David. Uh good to see you as always. uh when when when we think about what that landscape looks like and I'm going to talk about it with two hats here. One is my my ARM CEO hat and then also obviously I have a a role inside SoftBank and inside Soft Bank we're doing a lot of work inside data centers. Uh we've got a lot of uh projects underway in states like uh Ohio with the Portsmith project and other areas. Uh the shift in terms of some of the politics going on uh could have some short-term impact in terms of how data centers get built out uh across the US. I think in some states where it gets harder, other states may pick it up to make it easier. But I do believe strongly that if the US hesitates in any way, shape or form in terms in terms of data center buildout, the rest of the world is is ready and and and willing. Uh there is just huge demand for this capacity and I think globally uh the US really wants to be able to compete against China. China's not slowing down at all. So I think the West uh is definitely going to continue to invest Whether that happens at the rate inside the US that it was going to happen a week or a week or so ago, we'll we'll find out. But I'm not really worried about that in the uh in the long term. It's been suggested so far this morning that this maybe is the result of some desks giving especially these state house races a second look. Do you think there's more to it? Yeah, I mean look this is coming as was highlighted off you know a few months of we've had you know a lot of volatility big draw downs in some of the areas where you know many participants were were were focused and so you know kind of sitting there jittery I think you know I think before this 40% of long short funds were were down year to date you know there there's a lot of pressure and a lot of force selling and that's what we're seeing today is kind of one of the first kind of totally correlated you know draw downs across you know different areas you know yesterday we saw the mega caps up and some of the AI infrastructure names down today it's everything everything down including software and there's just a bit of a buyer strike um today with with all sort of participants selling so that's why we're seeing some of these outsides moved unfortunately I was going to say is there any area of tech that you can see where positioning is improving today >> so you know there's there's some areas motor solutions some some good results and was up we're seeing a little bit rotation into um some of the you know the the internet names you know overnight we saw some rotation in um you know from from Taiwan and Korea into into Hong Kong and Tencent was up um sort of strongly. You know, Sony reported some good results and and again, they were up strongly overnight. So, there are some some some brief pockets of of rest bite from this, but obviously, you know, pretty pretty painful across the board today. >> Is it justified? I mean, is there is there a real rethink on on how much is going to be able to get built in this country and what that investment might mean for the overall economy? Yeah, I think some degree of blue wave was was priced in, but you know, just the degree of it and then obviously that the Texas governorship um you know to say swinging blue and given that's you know about a third of of future kind of data center build out in in the US you know that's just you know led to some additional jitters. Now you know if you you kind of analyze what you know Hosea said you know it's not kind of you know don't build any data centers ever. It's just let's build them right. Let's make sure we've got the the right legal framework in place to do that. You've seen, you know, some of the builders react to that. So, you've had some Alman come out and, you know, let's not have these NDAs. Let's just be very clear and engage with communities as to what we're doing here, what we're getting, what the impact's going to be to your community, to electricity uh prices and things like that. So, you know, ultimately that that's a good thing. Um, but, you know, obviously there's going to be a knee-jerk reaction and that could lead to some delays and that's going to be some concerns. We're not going to solve for that in terms of knowing the answers to that anytime soon. Unfortunately, >> it does kind of feel like a moment in which if you had theoretically a press conference uh with the Texas governor who would essentially say, "Look, we're not here to demonize data centers, but we're going to have robust conversations about it." I mean, it I guess it feels like you might be in a moment where that would result in a bit of a face ripper rally. >> Yeah, absolutely. I think that that would be helpful. Um, so once I think everyone's digested their wins and had the parties and get to the serious stuff of you know what they're actually going to be doing uh when they when they get into into the state house um then you know that would be helpful. I I think obviously we've got the you the anthropic IPO process going on in in the background and again you know we'll see if they can come forward and talk about you know a reaceleration in their business. Some of their plans obviously they've made huge commitments to build out and you again can update us as to you know they're going to continue with those plans. They think those are still doable. So, you know, there are things that can I think can come out and kind of make everyone a little less fearful than they are today. Um, but we we knew a lot of these issues were were going to be kind of plaguing us along the way. It was never going to be linear. Um, I I think kind of our advantage as long only managers, as active managers, is we can, you know, do the work and try and assess that ourselves and we're not slaves to volatility. Obviously, many market participants are either because they've they're getting margin called and they've got leverage or they're, you know, they're sitting in a pod and and they have a a draw down, they're going to get fired. So um you know we can hopefully take advantage of that volatility where it's appropriate. >> I think I think no matter what you're definitely going to get pushed out the data center build I mean the in this particular case of te Texas what she has said is she wants a moratorum until there's legislation passed around permitting data centers. We know that can't I mean she doesn't come in till January. We know the legislative crisis takes time in Texas and they need twothirds but they want to do anything before September. So all of this just I think Richard feeds the feeds the worry and the concern that was already there from Wall Street which is how long do the revenues get pushed out right if the one thing that was happening fast was the data centers were getting built fast but if that now slows down too then what do we see on the other side >> yeah and look you know there's there's a little bit of near-term rest bite because I think you know the the current freeze from Abbott you know ends in December and there's a tiny window over Christmas maybe there'll be a few applications put in. It's going to be a busy Christmas for a few folk in Texas before, you know, she she she comes in and then has that, you know, sort of immediate freeze. Now, obviously, that's going to be impacting, you know, pipeline, you know, beyond 28. Um, so, you know, there's still plenty in the pipeline through next year and and beyond that that that's still there. And also, you know, there was over 400 gigawatts in the pipeline in Texas, which is, you know, way it was a lot of speculative building there that needed to be cleaned up anyway. And when we think about investing, you know, you got to think about that risk curve. You know, we we've seen some of the the impact of the NeoClouds, but yeah, the hyperscalers are probably the the most likely to actually be able to deliver on their road maps of building out those data centers. As you go down the spectrum and you go to Oracle, you go to the Neoclouds, you go to some of more speculative stuff. Yeah. Then it's going to be much more likely that that's not going to get done. That's going to be much more risky and that's before you adding, you know, some of the financing concerns as well. So, you know, again, from an investor point of view, you can think about that risk spectrum and where you want to sit on that. And obviously there were just a lot of market participants who are sitting at the riskiest end of that that are seeing the you know the biggest draw downs and the most volatility at the moment. >> Yeah. >> My next guest does some of the best work out there when it comes to the AI boom. He's in in studio right now to help address your concerns. Let's bring in now BEP Research founder Ben Palladian. Ben, uh first let's start with uh with with the financing issue, right? Because people, you know, listen, we just talked with Julian Emanuel. Uh lot of money has been raised here recently but people are questioning can it keep happening the circular financing will there be cracks in that what are you telling people >> I don't uh thanks for having on but I don't believe that there's circular financing we have one of the biggest private equity firms in the world Apollo Global who's basically made deals lending on large assets becoming the biggest buyer of Nvidia GPUs and Google TPUs to Lee Satanropic and to other firms and they just did a $40 billion deal with SpaceX to basically buy all those GPUs for Elon's biggest cluster. So, you can generate revenue and tokens from these GPUs. They're revenue generating assets, but you need to have a whole organization under it. I'm not saying everybody on this list will end up a winner. >> Well, let's talk about the main the name maybe the biggest name for the public that's the scariest right now, and that's Cororeweave, right? >> Right. Corweave is the one that's kind of leading this nose dive. the the neoclouds are are in question. Uh people are concerned about it on many different levels. Is there enough cash? Will there be enough cash for them to make it through? They probably have to raise more debt and cash as well. I mean, that's part of the business. It's very capital intensive, but they have a customer base like Microsoft and Meta and other customers that are using them for compute. So the the question is is if you go look at Microsoft's recent earnings, they're they're taking workloads, putting them on on Azure and speeding up business processes and generating revenue and benefit for their customers. >> That leads me to my next question. You the great answer because return on investment. Can these hyperscalers make money? Now we just came through this earning season. Their cloud revenues through the roof. I mean I I felt like they answered the question the last two earnings periods, but Wall Street doesn't agree. This is a great uh piece you put out here. more output, fewer payroll jobs, and and and and ultimately though, uh you know, with will they be able to say, "Hey, get back to those free cash flow levels because I've seen some charts that suggest in 29, 30, 31, they're going to have free cash flow that we've never imagined." >> Okay? I mean, as investors for the last 10 to 15 years, we've been conditioned with the MAG 7 being free cash flow monsters doing buybacks and dividends and because software had no cost. But today because of AI and the capex they need to invest and build out these data centers and that's where it comes down to just like how Amazon built AWS 20 years ago and the benefits are there now. This same buildout the future of compute is high performance compute and it's AI. So they have to make the investment today to make those dollars tomorrow. >> Right. And the bottom I love the word you're using is investment. We say in the media spend and I think that's negative right. We think about spending as something that you leave you don't get anything in return. These are investments. I keep telling people they're investing to avoid creative destruction. We don't they don't want to be talked about like Sears, you know, 10 years from now. Remember that company Google? Uh speaking of which, Anthropic is is pivotal right now. Uh because of its relationship, uh the commitments, I think 900 billion if you add everything up. That's a lot. Will they be able to do it? Because you've got some big names here. You say, of course, in your piece, they're still ordering Nvidia, but people are worried that, you know, that that maybe uh either that the open- source models will catch up with them and make them irrelevant or they just won't be able to live up to all the hype. I think a company has a great product, Cloud Code, and Opus 5.5, and they've basically from a business decision have spread the risk of their silicon amongst different suppliers. Any good business would would do that, right? You don't want to be tied to one supplier. But obviously the best uh compute for training these larger and bigger models at the lowest cost per watt is the Nvidia product and those Nvidia data factories. That's that's where they can't get away from. And because they have only like a small amount of $1 million customers, the runway to grow their revenue is is big. The market is still big. People still focus on coding. But if you really think about it, anything in life can be turned into code. any business, accounting, finance legal >> and you're confident that our frontier models will continue to have lead the way so that they dominate on the enterprise side of this and you know and live up to all the expectations. >> Yeah, you're going to use a mixture of frontier and then open source with your own train models in house. It's a it's a mixture of both to blend that cost. >> So every day we hear more and more listen this is the docom era all over again. Uh and of course different parallels have been used. Now uh you know recently we had this uh deal Marll and and Lum and it seemed like a pretty interesting deal on paper. That day uh the stock got hammered. Let's bring that other piece back up one second because I want to talk about this story. So back in the day JD Unif which is one of the top names in the whole do era bought SDL for $41 billion July of 2000. It was an ominous sign in hindsight. Are you concerned that this Marllum deal is a similar ominous sign right now? >> Not really ominous, but if you really look at the trajectory of the data center and where data needs to head and how fast it needs to go, you need Marvel's chips and Lumenum's laser optics to basically bring the data closer to the chips and move it faster. So, it's actually a good combination to make this system work faster for 2028 and beyond. By the way, I want everyone to go and maybe go to your website or on on social media. This is what was put out just on the combination of Marll and Lmentum. So, if you're long either stock, if you're long optics, if you're long the tech trade, you may want to be able to study this. You have to understand, folks, what's going on. You the charts, the movements, that that doesn't tell you the story here. This is one of the charts. Man, I'm glad you came. This is a phenomenal work. It's a great worksheet for anybody if you want to understand what's happening. Before I let you go, I want to talk about some of your core positions, right? So, I know in Nvidia is a name that you've been in, you say for 10 years you've been in Nvidia. So, you've written a lot of ups and downs. Why did you always have the conviction to stay in that and then you bought it on dips? >> Because when you take time and actually study the company, it's not just a hardware stack. It's the software, the CUDA libraries, Dynamo. Basically, they give you all the tools to run the AI factory. So, I I implore any investor to spend time and watch Jensen's GTC keynotes. He tells you everything and that road map is all there. You just have to pay attention and listen and have a long-term mindset of where the future's going to be. >> Free cash flow at both of those companies are accelerating. And in Nvidia, which is the third, it's exploding. So, it's Nvidia's now trading at a 13 and a half multiple. So, I think Aaron Gibbs is going to jump in. >> Hey, real quick. Uh, Nvidia, you know, pulling out a two-year chart. Looks phenomenal. But here recently, it has come down, right? We were near 230. We're near 180 right now. For you, this is a gift, right? >> Yeah, it is. I mean, again, we were buying this stock around 100 uh adding to it. Uh during Deep Seek, these dislocations are actually good for long-term investors. So, you want to pace yourself. If you want to buy it, add a little, add a little bit more. We're going to continue to hold it and add to it. It's actually now a member of our value portfolio. >> All right, folks. The key word there is pace and stay in there. The midterm result has AI bull Dan Ies rethinking not only the data center story itself but how long hyperscalers will continue spending on the AI buildout. He joins us now as you see at Post 9. Good to have you. Good to be here. >> Uh we really needed to speak with you because you put out a couple of notes over the last 24 hours I think with a more cautious stance on certain parts of the AI trade. 10% fewer data centers. That's your baseline view. You also suggest that big tech capex the growth rate now peaks in 2027 versus your previous view that wouldn't happen until 28 or 29. So are you putting words into action and then cutting back some of your expectations for the stocks that are in your universe? >> Yeah, that's essentially I mean you know me and you talked going into the midterms that there was definitely in terms of a potential blue wave what that could do. I think what we've seen especially in Texas, you know, now that that 10% of data centers, you know, and that and that could be low. It could be 15 20% of data centers that ultimately don't happen. But our view now is that capex for big tech it peaks out in 2027 next year. And from a growth perspective, >> much earlier than you expected. >> Yeah, we thought it it wouldn't that probably won't happen 2028 maybe or 2029. The reason that's ultimately happening is is that the capacity is going to be constrained. But as we've said, the reason I view it for the hyperscalers bullishly because the market's been screaming worrying about overbuild, worrying about free cash, worried about these ATMs, all the capital raises to some extent that that now really gets thrown out the window. The hyperscalers are in the best position in terms of how we change our view. I think ultimately it's software over semis because my view that the use cases are going to continue to build out in terms of the capacity that's there and then neo clouds it's going to be a game of thrones there in terms of ones that will ultimately emerge. Nebby is probably the one that you'd have to front and center. >> So you don't think there's any earnings fallout for any of the hyperscalers based on you pairing back what they what you think they're going to you know in terms of pairing back their spending. >> Yeah. to put numbers with it and we've talked to people across tech down in Texas engineers ones that are building data centers. So let's just say right now from a gigawatt perspective call about 40 gigawatts in terms that we have you know in the US by a year end yet about 40 to 45 that was going to come online. So let's call 10 to 15 of that is probably now you know either not going to happen or it's going to get slow rolled. My view is that demand to supply today demand still 40 to 50% over the supply. So capacity those that have it will continue to be ultimately those are the those are going to be the winners that's the hyperscalers the ones that really are much more threatened. It's open AI. It's anthropic. They need to bring on that capacity. That's a huge part of it. You see with SpaceX, Korea, what's happened in Texas as well. >> But if if you have basically been suggesting that, you know, this capex runway is long and we're only in the third inning of the buildout, right? You've said that, right? You're talking about the infrastructure buildout. >> Now, you're telling me that maybe we're in a later inning than than you thought. Correct. I would view it that we're third inning, but now >> because my view is that when you look at it, it's an elongated cycle. So in other words, now if I thought it was third and at the seventh inning, that's where we've talked about maybe you now start to, you know, you start to get much more caution on tech. My view now is that the amount of capbacks that's going to be spent over the next 3, four years, it's not changing. It's yeartoyear how it's going to change. >> Yeah, but what if they call the game after five or six innings? You know what I'm saying? Like if things are are slowing as dramatically as you think they will and if they actually slow even more than that, you're making the assumption that it just elongates the cycle. What if it shortens it? >> And that's a risk. But to that point, if demand was starting to wade, that would be different. When I look what's happened on the Palunteer Snowflake, when I look what's happened from enterprises on a use case perspective, demand basically is 4 to one. If you if you started if right now you actually go to a software company and try to actually get a project, they're basically telling you late 2027, maybe early 2028. My point is that yes, some of that could fall apart in terms of deals that don't happen. But relative to what these stocks are baking in, when I look at Alphabet, when I look on chips is a good example, like with AMD, when I look on the software side with Palunteer, on cyber security, I Scott, I just view this as like we went through moments like this in CO. We went through moments like this in terms of liberation day and tariffs. I think that's what's happening here. I think this is the market way, I think, mischaracterizing that this AI revolution buildout is not going to happen. it's just going to be more elongated and capback's going to peak out earlier than we thought. >> So, okay, I can I can get my arms around, I think, you know, why you would remain bullish on the uh the so-called incumbents, the the hyperscalers who already have all this compute. They have more levers to pull. Maybe by, you know, doing all this, their free cash flow goes back positive sooner than it otherwise would. So, I'll give you all that. What I don't understand is that if you pull back your expectation of capex itself, if you lower demand for data centers by at minimum 10%. How can your view remain the same on Nvidia's numbers? >> How? >> Because my view there is that the issue with Nvidia and we just saw when we were in in Asia 13 to1 demand and supply. So let's just say that demand and supply now goes seven or 8:1. Still, when I look at Nvidia numbers, I think they're underestimated by 15 to 20%. So, when I look at a stock trading where it is, my view is that numbers are way underestimated. I get that they're going to ultimately have issues if you don't have the data center build that we think. But when you look at the runway to 2028, 2029, I don't see at a stock where it's trading that I could be any less bullish given where I think the street numbers are and what's actually being baked in to the capacity. >> Where do you stand on, you know, commitment is a is a dangerous word in this market. People talk about the amount of money that has been committed by whether it's the Frontier Labs or others in this universe to spend on compute and everything else in this infrastructure buildout and we take it as the gospel. What if it isn't? And what if some of that so-called commitment is rolled back or doesn't live up to what we think it will? Is is that a significant game changer? Are you overly optimistic that commitment will turn into reality? >> I view it somewhere in between. So let's just say right now 40 gawatt capacity commitment let's call it is about 4x that more what's in the pipeline from a backline perspective. So let's just say 30% of that comes off relative to where the stocks are trading here. you buy every single hypers scale or you buy every single big tech name and specifically some of the software names Nvidia included and AMD on the chip side neoclouds that's ones where they're clearly on on the the wrong side of this to some I'd say Nebas and you can say well core we've get you know be ultimately one that could emerge but it's going to be massive consolidation musical chairs and then the power players from GE Vernova to everyone else because you're going to have less data centers >> yeah but you put risks around open AI and anthropic also No doubt you said that >> if OpenAI and Anthropic today were public companies, those would be the ones that I'd be coming here being like I relative it depends on valuation. You have to be more negative on those names because they're the ones that have to build out the capacity. Hypers scale is already on it and that's where it goes to where rubber meets the road with anthropic IPO. >> Aren't you going to cover you you're you're going to end up covering Anthropic, don't you think? >> Sure. I mean now it's that that would be definitely one that we would cover but the point there is like okay 1.5 trillion now is it 1.3 is it 1.2 to where's the number given what we're seeing here because ultimately they have to be the ones to build out capacity Nadella Yuka Jasse they already have it >> let's bring in Tom Lee of Funstrat who's joined me here at at Post 9 I don't think you were expecting this based on what you told me yesterday >> no not expecting it at all but uh you know as you're pointing out this is a lot like tariff liberation day coupled with deepseek and coupled with situational awareness. You know, we got a policy shock because that blue tsunami uh some rhetoric that scared people and I think today's drop, especially the overnight session in Korea, I think has caused a lot of liquidation. I I think there are stories of some large hedge funds actually being liquidated now. I mean, the the difference between this and say a a liberation day is just as the president held up the, you know, now infamous board with all of these large tariff numbers, he could just as easily take that away, which he essentially did, which is why the market rallied so far back. What's the mechanism for reversal here? It's not as simple. Well, yeah. I mean, for for foremost, you know, there's just a lot of uh fears building about what could happen with AI progress, but I think we have to just keep some things in perspective. One is that the demand side for AI isn't going to slow because there's still a lot of discovery and companies are benefiting. And the supply side, there's not really new AI competition or uh something that could wipe out the industry. What's really being discussed here is a slowdown in the US, but that means a lot of opportunities are going to emerge for either states that embrace AI or other nations that have lower power. So I I think in some ways this is sort of a a speed bump question, but does that warrant the massive massive repricing of these stocks? I I think that's overly done. Well, I mean, unless the speed bump becomes a pothole and you have a more sizable sort of clamming up of the the the money. >> Yeah. >> And that people just are forced by, you know, market forces, market turmoil, other things in the environment that just cause a either a delay or an absolute, you know, rethinking of some of the spending. Are are you considering any of those? Uh I am but I I think we should just have some perspective on the consequences. You know 10 gawatt of uh AI construction is equal to almost 350,000 homes or like 5,000 hotels. So a state that wants to prevent this construction is going to have a gigantic hole in their budget. Number one. But number two, I think we have to remember that this is this is really completely changing the narrative around inflation risk and even the Fed hikes. Today's productivity number means I think the September hike looks like a mistake. The December probably doesn't happen, but then I think the Fed put comes into play if it becomes a employment risk story. So I I think that there's inherent stabilizers and again I think it's just too early for people to say AI construction won't happen. You saw the the Texas governor elects they're going to actually allow the existing projects to to be completed. So I think there's a lot of doom saying that's maybe overshot. What if growth you know total GDP slows? If let's say 75% of it was directly tied to the AI infrastructure build out for 2026. What happens if that slows? Isn't there a a ripple effect through financial markets? >> Yeah, I I think we have to be give companies some credit. They were very ingenious about dealing with the tariff shock, the tariff liberation day because earnings held up. I think companies are going to find ways to deal with this policy shock of the blue wave. And I think we even heard stories that for instance, you may be retrofitting empty industrial plants and office space to put in data centers. So I think there's going to be ways for power and gigawatts to be constructed. Um, so I look at this as an opportunity for American businesses to be use some ingenuity. >> Would you be a buyer of many of these names today? >> Uh, just want to point out tariff liberation day, deepseek liber, situational awareness day. The bottom the the longest it took to bottom was 3 days and then it was a V-shaped symmetrical recovery with some six-month gains of over 20%. So, >> I know. Are those really analogous though? Like I said, you know, Liberation Day was a scare about numbers that weren't real numbers at the end of the day. That's why the market had a a V-shaped recovery. The the Deep Seek thing was a scare. It didn't materially change the fundamentals of any story whatsoever. This has the smell of a fundamental shift in a significant part of the AI story. >> Yeah. Um, just want to point out a few things. One, the market that there's still a very tight supply market for memory and semiconductors. That's still going to be very tight and the multiples have come in. So, we we've now priced in skepticism of earnings and then a belief that the market is actually going to get loose. I I think compute's going to get quite expensive if we don't actually build data center capacity. I think the overall economics get better. So I I think in some ways Scott we repriced the whole AI trade on a belief of things that actually haven't happened yet. So in some ways I think it is equal to deepseat plus tariff liberation day plus of course because there's real liquidations take you know situation awareness. I mean it's three shocks in one. >> Okay. >> All right. I hope you're all doing well today and staying calm in this market. Thursday was a rough day in the market and it was an especially rough day for AI infrastructure stocks. This is on continued concerns regarding data center push back and potential moratoriums following this week's midterm elections. As I mentioned in last night's video, the main election result that spooked market participants was the governor race in Texas. This is because a large amount of data center capacity is being built in Texas and the incoming governor has advocated for a moratorium on data centers until the legislature can meet to put protections in place. As I mentioned in Wednesday night's video, while I expect there to initially be opposition to data centers, I do expect that opposition will decrease at some point in the coming months because of the economic benefits, both jobs and tax revenues that data centers and data center construction provide to Texas. And so, while I can understand the short-term reaction, I do think it is somewhat overdone. Keep in mind that there are other states and countries that are welcoming of data center construction, Nvidia has a large global customer base, allowing them to redirect capacity as needed. Jensen Hong spoke about this at length back in October and I don't think the hyperscalers are going to pull back substantially on their capex plans given that they're seeing very strong cloud growth with backlogs worth hundreds of billions of dollars and they're also investing heavily to protect the modes of their core businesses that said some notable voices reigned in their bullishness on the AI buildout Thursday which exacerbated the selling. I've also seen indications that there may have been some foreselling on Thursday which would have exacerbated the downward price action if that is correct. I noticed Gavin Baker pointed out on X that the governor-elect of Texas previously indicated in an interview that she will pass laws pertaining to newly permitted data centers going forward for delivery in 2028 and 2029, but all currently permitted data centers in Texas will be allowed to finish construction. Gavin makes the point that this covers all of 2027 and likely well into 2028. And so again, I understand the knee-jerk reaction in the market, but I do think some of the selling is overdone. Now, let me cover a few news stories about Nvidia and Micron, and then I'll cover Iron earnings after that. Reuters is reporting that about 600,000 Nvidia H200s are stored in Hong Kong, awaiting shipment to mainland China as Beijing withholds approval for large-scale imports despite there being demand from technology firms in China. Sources indicate that some chips have been allowed to enter China in recent months with some Chinese internet companies obtaining about 10,000 H200s each in the past few months. In other news, The Information published a story claiming that Nvidia has approached Apollo about converting existing telecom facilities into AI data centers. The project could span about 20 US states. The buildings belong to Apollo backed Brightpe which reportedly operates more than 130 facilities. Nvidia reportedly identified about 1 gawatt of available power capacity across the facilities. I have no idea if this story is true or not. I'm just bringing it to your attention so that you're aware of it. Bloomberg published a story saying that Micron paused work on a $4.5 billion memory plant in Taiwan due to an extended labor dispute. That added fuel to the fire, the AI infrastructure sell-off, as some market participants perceive the reported pause as being related to the midterm election results and a signal of weakening memory demand due to a slower AI buildout. That doesn't appear to be what's actually going on with this story. I'm just telling you what some market participants knee-jerk reaction to this story was. Now, let's cover Iron earnings. I'm just going to rapid fire important points both from the earnings report and earnings call. Iron reported quarterly revenue of $228 million versus roughly $221.7 million expected. Quarterly AI cloud revenue was $200 million. I was expecting somewhere in the range of $160 to $185 million. So that's better than what I was expecting. Bitcoin mining revenue was only $28 million for the quarter as iron decommissions mining hardware. 28 million is lower than what I was expecting. Overall, that's a positive from my perspective because iron exceeded my expectation for topline revenue even though Bitcoin mining revenue was lower than what I was expecting for the quarter. Bitcoin mining revenue declining is overall positive because we want iron to transition from mining to AI cloud quickly. So that's a positive. AI cloud ran at an 86% gross margin for the quarter. Quarterly net loss was $263 million driven by $148 million of depreciation and amortization, including charges on newly commissioned data center and GPU assets, $16 million of non-cash, stockbased compensation, and $22 million of net finance costs. Iron's more than $4 billion ARR target is fully contracted, which is a change from last quarter, and it is fully funded. Iron expanded its contract with the unnamed Frontier Model Company after the quarter end. That's likely anthropic, but we don't know for certain yet. Iron also said contracted pricing continues to rise, up 11% on three-year deals and up 12% on 5year deals. And this next point is big. Iron signed a short-term contract for $40 million per IT megawatt of revenue. That's very strong. Importantly, that's for it megawatt of revenue, but we don't know which site that contract pertains to, and so we don't know the pee. But the important takeaway is that contract pricing is moving higher. As a brief side note, if data center capacity is truly going to become more scarce following the midterm elections, that's positive for the prices that companies like Iron can charge customers to rank capacity. Iron says that the more than $4 billion ARR target for year end 2026 is now targeted for January 2027 following a customer requested change and supplier timing and it may even extend to February 2027. On the earnings call, leadership confirmed the rumors. Microsoft made a request to reconfigure some of the network infrastructure after Iron delivered Horizon 1. Iron said yes to the requested change and that is the reason for Iron pushing its year-end 4 billion ARR target back to January 2027 possibly to February depending on supplier timing. So the rumors are true. Importantly, Iron built Horizon 1 to spec and Microsoft accepted delivery, meaning Iron delivered the data center to spec. So it's not like Iron made a mistake on Horizon 1. Microsoft requested the change after they accepted delivery. Iron also mentioned signing a contract after quarter end at approximately $22 million per IT megawatt per year on a 5-year contract including customer prepayment covering 90% of associated GPU capex. Iron also mentioned that they have $20.6 billion of total assets including $6 billion of data center assets with 20-year useful lives and their assets and revenue are underpinned by customer contracts with 15 customers. Twothirds of revenue backlog is with investment grade counterparties. Iron also said they have $5.8 8 billion of cash, including $2.3 billion of restricted cash reserved primarily to finance GPU deployment. On the earnings call, we learned that Iron expects $250 million of AI cloud revenue to be recognized on the company's next earnings report. That number would have been higher if there were no delay related to customer requested changes. By June next year, Iron expects to have $25 billion of GPU and data center assets. Iron CEO was asked about modular data centers and said they're talking with Nvidia about that. As a reminder, Jensen Hong recently announced NVIDIA DSX Pods at GTC Berlin, which is essentially a ready to go AI factory assembled in the supply chain and housed in a transportable container which serves as the shell of the data center. It removes the need to construct a large building to house the AI factory. As I've said before, I think Nvidia's DSX pods will likely be a major success and an important accelerant to growth. That said, it will take time for the industry to get fully on board. Iron recently recontracted H100 capacity at double its original pricing. Iron spent about $5 billion on capex during the quarter and their expectation for capex through the remainder of the fiscal year is about $15 billion which is down from their prior forecast. Iron was asked why expected capex was lowered and leadership said it's not because of a lack of demand. It's purely because of the current state of capital markets. Iron wants to strike the balance between expanding to meet strong customer demand while at the same time not raising money at all costs. They want to strike the right balance. Market participants like that. That said, leadership made clear that the demand signals are there. They recently signed a short-term contract at $40 million per megawatt and iron recently recontracted H100 capacity at double the original pricing. Leadership indicated that Nvidia liquid cooled Vera Rubins are expected to start arriving late next year and the supply of Vera Rubins next year is incredibly scarce. Iron is excited about the economics in light of that scarcity. Also, Sweetwater 1 and 2 are both under construction and included as base load in the batching process, meaning that the 2 gawatts across those two sites in Texas are secured. As a reminder, the incoming Texas governor has indicated that sites that are already under construction will be allowed to continue. Sweetwater 1 and two are both under construction. Overall, I would say this earnings call was much better than the last one. Iron appears to be exercising some restraint on raising capital and capex in light of current market conditions. And at the same time, they're seeing stronger contract pricing and demand from customers. As I said, quarterly AI cloud revenue was higher than I was expecting and topline revenue was higher than I expected, even though Bitcoin mining was less than I was expecting. That's all positive. Previously, I shared an expected price range of $100 to $150 per share for the first half of calendar 2027, and I said I would adjust that range over time as needed. The reason the range was so wide was because of Iron's large $6 billion ATM program. $6 billion is a lot for Iron. And then after the last earnings report, I narrowed the expected price range to $90 to $120 per share. And I told you that I expected to narrow that range even further as Iron issues more shares. I also said repeatedly that the expected price range was not guaranteed. It was simply a general range of what I thought was reasonable. I also said repeatedly that there were multiple assumptions baked into the expected price range. Those assumptions were one that there's not a market downturn, two that iron reaches their ARR target on time, three that the world remains compute constrained, and four that there is no significant dilution beyond the $6 billion ATM. I said repeatedly that if any of those assumptions do not hold true, then the expected price range should not be expected within the relevant time frame. Now, two of those assumptions have not held true. Of course, we've seen the recent market sell-off in tech stocks primarily concentrated in AI infrastructure stocks and especially concentrated in the Neocloud stocks. So, that's the first assumption that didn't hold. They've been a notable market downer. The second assumption that has not held true is that iron will not meet its ARR target by year end. Again, that's due to Microsoft requesting a change after Iron delivered Horizon 1 according to spec. So, it's technically not Iron's fault. Still, the assumption did not hold true. Therefore, the expected price range should not be expected within the relevant time frame. If I account for the dilution that occurred during the quarter and the fact that the $4 billion ARR target was pushed out to early calendar 2027, that gets me to a price range closer to 70 to $95 per share in the first half of calendar 2027. But there are a couple problems with that range. First, Iron indicated that there's already been more dilution after quarter end. If I account for that additional dilution, that gets me to a price range of 67 to $91 per share in the first half of calendar 2027. We also have to consider that iron will likely issue more shares under the ATM in the days ahead and that would bring the range even lower. Additionally, we have to acknowledge that capital market conditions have changed notably in recent weeks and there's been a notable selloff in AI infrastructure stocks. So, put simply, two of the assumptions baked into the original expected price range have not held true. One being a market down and the other being Iron not reaching its year-end ARR target on time. And so, the expected price range should not be expected within the relevant time frame. The range of 67 to $91 per share in the first half of calendar 2027 is what I think is reasonable, but that range does not account for additional dilution that will happen in the future. I expect to narrow the range further in the future as iron issues more shares under its ATM program. Something else that I said previously was that if Iron delivers on its ARR target by year end, then the earnings report released in May 2027 could potentially be the first earnings report that really grabs the market's attention in a positive way. With the ARR target being pushed back to early 2027, I now think the earnings report released in August 2027 is likely to be the first one that really grabs market participants attention in a positive way. Now, I continue to be bullish on Iron and I expect that I will likely be bullish at least through the first half of calendar 2028, depending on what happens. The ARR target being pushed back is simply a timing issue, so I'm not worried about that. And hopefully in the future, Iron will be able to lean more heavily on asset back financing and larger customer prepayments as well as ramping revenues. So, yes, I'm still bullish overall, even though I reduced my expected price range for H1207. As I alluded earlier, I do think the selling in some of the AI infrastructure names is overdone and misses important context regarding the AI buildout. That said, I don't know where the bottom is for stocks, and I don't know how long this will last. If you need help with your situation, consider reaching out to a certified financial professional who can help you with your unique circumstance. As I often say, remember to stay calm in this market. Remember to maintain a long-term perspective and to not make any hasty or irrational decisions. Looking ahead, we have Nvidia earnings on November 17th. Now, in case you're new to the channel, I want to make sure that you have at least a basic understanding of the underlying long-term thesis. So, let's cover that. Now, I don't know what's going to happen in the short term, but from a long-term perspective, I am very confident that Nvidia will be worth much more in future years than it is today. When Jensen was on the Lex Freedman podcast not that long ago, he was very seriously raising the possibility of Nvidia becoming a $3 trillion revenue company in the near future. If that happens in the coming years, then it is very plausible that Nvidia could one day be worth tens of trillions of dollars in market cap. That might sound crazy, but that's what Jensen is implying when he raises the possibility of Nvidia becoming a $3 trillion revenue company. I guess the question at that point is what multiple the street will be willing to give Nvidia. I don't know the answer to that question, but I truly do think that Nvidia will be worth much more in future years than it is today based purely on the fundamental growth of the business. Based on everything I'm seeing, the world is still computed, and I expect that to continue at least through calendar 2028. In a computed environment, developers will use whatever viable compute they can get their hands on. Today, there are no GPUs that are sitting dark due to a lack of demand. Like, there was fiber sitting dark due to a lack of demand at the height of the dot bubble. Back then, companies were laying fiber in the hopes that use cases and demand would eventually show up. Today, we are seeing the complete opposite. As I've said many times, when market participants compare this AI revolution to the dotcom bubble, they ignore the fact that the internet is already here this time. This means that mass adoption of the technology and new use case development at scale are immediately possible. We don't have to wait years for it to show up. It's already here. The world is compute constrained, which means there is not enough supply to satisfy demand. New capacity is utilized as soon as it comes online. The hyperscalers are monetizing capacity as soon as it comes online. Each of the hyperscalers spoke about being supply constrained on their most recent earnings calls. Additionally, many of the clouds are building out into contracted demand. They're not blindly building in the hopes that demand will eventually show up. No, they're building out because they have signed contracts and in some cases significant prepayments from their paying customers. This AI revolution is fundamentally different from the do-com bubble. And 2026 will be a pivotal year for the AI industry thanks to the rapid adoption of agentic AI and the proliferation of agentic systems in the world's leading enterprises. The leading AI labs revenues are surging right now. Agentic coding and the implementation of agentic systems in large enterprises are new use cases that are increasing inference demand significantly. That subsequently is increasing compute demand. The rapid adoption of agentic AI is why we're seeing an inflection in inference demand. It's why we're seeing the leading AI labs revenues surge. I wish both anthropic and open AI were public so the public could see the ramp in their revenues. I think the leading labs surging revenues may be the initial proof point that grabs market participants attention and causes them to realize that there will be a clear ROI on AI infrastructure. I think the leading labs surging revenues will also help assure investors of the longevity of Nvidia's growth since these labs revenues are directly tied to compute. If they had more compute, they would have greater revenues. It really is that simple. Demand is not the problem. The problem is a lack of supply to meet the demand. As I've said previously, I expect the world to be computed at least through calendar 2028, possibly longer. And so regardless of what happens in the short term, it's important for long-term investors to remain focused on the fundamentals, maintain a long-term perspective, and remember that we are only in the early stages of aic systems being adopted at scale. This will increase compute demand significantly, and after that, the next surge in compute demand will likely be fueled by physical AI. We're no longer talking about digital agents performing digital tasks. With physical AI, we're talking about physical AI agents performing physical tasks in the real world. NVIDIA CFO has called physical AI, quote, a multi- trillion dollar opportunity. and the next leg of growth for Nvidia. This industry will fundamentally transform society and Nvidia has positioned themselves to benefit massively. NVIDIA sells the hardware for the data centers where the models are trained. They offer omniverse where the models are taught and tested. And Nvidia also sells the hardware that allows ondevice real-time inference through NVIDIA AGX allowing robots to have intelligent interactions with the real world even when they are not connected to a data center. Notice that Nvidia is taking a holistic platform approach to physical AI and they're embedding themselves as the underlying foundation supporting all of it. Over 3 million developers are already building on the Nvidia robotic stack and this is not getting enough attention. As for production ramps, Blackwell Ultra has ramped and remains in high demand. Vera Rubin is rolling out to customers. Nvidia Gro 3 LPX is in full production. Later on, we're expecting the launch of Reuben Ultra in 2027 and Fineman after that in 2028. We have a clear data center product roadmap stretching into 2028. And Jensen believes that AI infrastructure spending will reach three to 4 trillion annually by the end of the decade. That means Jensen is expecting growing AI demand and an expanding total addressable market underpinning all of this. I don't think we are anywhere near any type of bubble bursting type of event. With all of this in mind, I seriously think that Nvidia still has plenty of runway ahead of it. And I think this company will be worth substantially more in future years than it is today. At least that's my view of the situation. Quick note before I wrap up. All of the compilations on this channel are edited by Finn Vid with original structure and commentary. Occasionally the same edits appear elsewhere on YouTube. If you're looking for the original version, it's always here on this channel. Thanks for watching, Finnvid. I appreciate your support. Remember to stay calm in this market. Remember to maintain a long-term perspective and do not make any hasty or irrational decisions. With all of that being said, I hope you all have a great rest of the day. And I'm curious to hear your thoughts about Nvidia in the comments below. Please leave a like on this video so more people will see it. And while you're down there, please consider subscribing. It's free and you can always change your mind. Thanks for watching and hopefully I'll see you in the next