CNBC & Fox Today On NVIDIA Stock, Micron Stock, SK Hynix Stock, CoreWeave - NVDA Update

watching the chip sector with the sock semiconductor index riding an eight session losing streak. It's the longest since May of 2006. It's bouncing back right now. You can see there on your screen uh up two and a half%. But for more, let's bring in Tom Lee, head of research at Funstrap Global Advisors, also a CNBC contributor, and Ben Emmens, founder at Fed Watch Advisors. It's great to have you both here. Tom, I'm going to kick this off with you. Uh how would you describe this market right now, especially when we have an S&P that if you look over the last couple of months has been pretty rangebound? Um, good morning. You know, the S&P has actually held up very well. It's very close to all-time highs, but as you know, underneath is an incredible amount of turmoil because it is trying to digress this wall of worries. Uh, AI spending, how it's going to be funded, you know, even the growth rate of these frontier models. On top of that, it's energy, which is very important, but it's very un uncertain because not only are prices high, but you've got two wars, and both wars in the Ukraine and in Iran seem to be escalating. And I think it's at a time when there's a lot of criticism about valuations. So, I think investor sentiments negative. I know this sounds all very scary, but to me, this is is because fundamentals are so strong, this is a coiled spring. I I think stocks are going to really ultimately get through this, but we're at a time of great concern. >> Um to go back to the coiled spring, Tom, where do you see those opportunities right now, especially where when in recent trading sessions, um the focus has really been on this rotation under the hood in the S&P right now? I mean, and and actually, if you look at the stocks, if you look at semiconductor stocks versus the equal weighted S&P, there's been a bit of an inverse correlation. So, does that actually have to come together and and meld a little bit more to really move higher and see that spring bounce back here? >> Um, well, I I think to me where the opportunity lies in things that investors have been unwinding. You know, uh, >> I've saw some reports from the brokerages that AI positioning by hedge funds has been almost entire the overweight has almost been entirely unwound. I think that number one explains a lot of the selling pressure, but I also think that's where the opportunity lies because the the growth you you know the visible growth the source of drivers of S&P earnings it's really coming from the AI and all the bottlenecks and that's still where valuations are quite reasonable. So to me I I think as much as people are sort of fearful of this I I'd be sort of jumping into the AI semiconductor mag 7 software these are AI downstream names as well as crypto which which got caught caught up in the same selloff reports last week said that anthropic revealed slowing revenue growth when it briefed investors but our next guest says that that slowdown doesn't necessarily signal trouble for the broader AI industry. Joining us right now is Gil Laura. He is DA Davidson's head of technology research. And G, let's just lay this out for anybody who wasn't paying attention last week. If you look at the revenue numbers for Anthropic, in the first quarter of this year, they had $4.7 billion. That grew to 11.5 billion in the second quarter. That grew to 16.8 billion in the third quarter that we just saw these numbers. Um, that sounds like really phenomenal growth to keep seeing those billions. But what spooked people last week is that the rate of growth is slowing and and pretty drastically. Um, you say that this is not spell the end of days for the AI trade. Why don't you lay out your thesis? >> Yeah, I go back to basics at a time of great volatility like this. First principles, am I using the AI models more than I've ever used them before? Am I paying more than I've paid for them before? Are people in my professional personal circles using the models more? Are they paying more for them? Are we likely to pay even more next year? If that's all right, that means AI investments are going to continue. And that's good for your Microsoft, your Amazon, and Google because they're going to be the ones serving it. And they're going to be buying chips primarily from Nvidia, Broadcom, Micron. What's good, the good news is that all the companies I just listed are trading close to a market multiple or below. That's the AI trade. Which model ends up winning is actually a little consequence. If it's anthropic or open AI or Google or anybody else, open weight gets a bigger piece. That's of little consequence. And what's even little less consequence is if coreweaver and scale even make it as companies, that's of no consequence. Actually, the big picture is we're using AI more. >> We're using AI more, but it matters greatly if you happen to be an investor in Coreweave or if you're an investor in potential investor in anthropic. It matters greatly to people who are trying to figure out where they can make money on this AI trade. >> That's right. And that's why we're focused on the companies that are collecting those biggest profit pools and will continue to do so in a wide range of environments. That's why I mentioned Microsoft, Amazon, Google, Nvidia, Broadcom, Micron, AMD. Those are the companies that will capture the profit pools really one way or another. Anthropic will end up being a very speculative investment. They may do well, they may not. To your point about deceleration, the key to understand is they were at peak power earlier this year. They had almost the entire enterprise market. That's when their growth was really skyrocketing. What happened since then is OpenAI caught up and in enterprises realized they were spending so much that they had to start using open weight models as well and that's why anthropic decelerated. But overall, if you were to pull out, you realize that we're pulling as many token the the token growth never stopped accelerating. We're still on a very steep slope for that. And that's what really matters. Th those um Coreweave, Nebus, uh Nscale are also highly highly skepul speculative. They're very leveraged to the AI trade, but that doesn't mean they're going to win. That means that either they'll win big or they'll go to zero. That's not the real important part of the AI trade. It's the ones that are still trading at reasonable multiples that are due to capture the biggest profit pools. >> The ROI I mean you have to have the other thing is that I think that you and I would disagree on what is the worth of an H100 that's the basic Nvidia let alone uh Ver Rubin what is the worth out six years because remember uh Cory depreciates out six years. So on the seventh year it's free to trade. Is that free compute? I mean, and then do we wrap up the compute? You look, Larry Frink last week said here that the next mortgage back is going to be compute. Yes. >> Yet no one that I know has actually gone the step forward and bundle compute or make compute. So two-way market, >> not yet. But I mean this this effort to securitize chips and GPUs is going to be very important >> because conceivably it would lower the overall cost of capital. >> Yeah. because it would if you create a securitized market that is collateral of which is GPUs and other you you conceivably lower the overall cost of capital and that's what I'm waiting for but I'm not saying it's going to take a while no but it's going to take a while it's these these markets don't >> don't happen overnight what you're talking about a pretty significant effort >> yes >> uh that involves uh a lot of data >> and a commoditization in in in an effect because if you're creating this you know I don't even know the answer to some of the questions. For example, if you got a GPU sitting in a server rack somewhere here, is it the same as the one that sits here? You just drew a distinction between Corweave, what they do, and maybe what some of the other uh Neoclads do. Is it the same or not? So, because >> But don't you think that Morgan Stanley would have mentioned it on its call? That Goldman would have mentioned it on his call as maybe they weren't asked about it. >> We asked we asked Larry Frink about it last week. He gave an answer. He He talked about how they're putting all the data together right now saying like, "Let's make a two-way mark." You know, these guys come out, they say, "Listen, the market's wide open. We're all set." Where's that? Why weren't they ready? For heaven's sake, they're the earnings. >> I think the opening was with Becky when they all sat around the table and talked about the half a trillion dollar, you know opportunity. >> Sit around the table with you. It's not sit around the table and I want I want a two-way market, >> right? But that was really not very long ago. You're being very impatient. Uh, impatient. Yes, >> I'd say that it should have been done yesterday. >> You're being very impatient. Look at Nvidia. It this stock, this company just happens to be Ted Thatcher's stock pick of the day. Ain't that right, Ted? >> It sure is, Stuart. >> Well, that thing's really sold off big time last week, bouncing back 2% this week. I hope there were some dip buyers out there last week. You got amazing little uh gift headed into election season. And you know what I like very much about Nvidia is just how you can see the difference between a stock's price and its actual value in the marketplace actually show a little bit of a discrepancy. If you look this year they're about a 23 times earnings. But if you look at their forward earnings depending on exactly how it's calculated you're starting to see numbers in the mid to low teens even at a PE ratio like that. I think Nvidia is a nice buy. Obviously, they are the loadbearing wall, the tent pole in this AI trade, and I can I believe very much in their leadership and Jensen uh will continue to execute very well. >> I should say that when we get these big tech earnings, they better not disappoint. >> I completely agree with you. I think the theme of this year is if you miss by an inch, you miss by a wide wide mile. >> I think you're right. He also had some funny I think a funny tweet about compute and you know basically I say this all the time because all the reporting shows that there's been nobody who said there's a demunition in demand. >> Exactly. >> In demand continues. >> Right. >> All of those who are building this compute continue to build it as quickly as they can. >> Right. And the the blockages zoning is a problem and uh power is a problem. And let's not forget, it's not like Jensen Juan can get all all of the chips that he needs out of Taiwan semi just like Intel's having a problem. Ju just we have you could say that ARM is having a supply problem. So yes, well they can't get what they need. >> Yeah. >> So I mean Nvidia was up four uh about a half hour ago. >> Yes. >> And this is what I was worrying about. This is that retest that I think you have to have four down to you have to bring out these sellers. you have to clean them out. Uh we have to find out whether there were some hedge funds that had overextended themselves. Perhaps that was part of the problem. We don't know. >> We don't know. Last week there were any number of as you typically see sort of some rumors around wasn't able notational awareness which really does still doesn't have a public book of equities. I did report on them buying options some time back but some other names some other funds that had similar trades on. Were they suffering last week? Of course they were. Were they getting taken out? No. I I heard nothing about that sort. At least we had rates that were above where we are right now. Yes, we did. >> And it wasn't a gating issue. >> We have to get used to it perhaps because I remember 67 was not >> a lot of the spend there may have been as a result of equity that was being raised by companies that had no business raising anything because their business models were completely flawed. as opposed to now where we have the biggest companies in the world and once the most profitable companies in the world that are willing to go to free cash flow negative and issue debt in order even though their balance sheets are still >> No, I did a think piece this weekend for the club are willing to finance this thing. >> You know, if you want the analog of the bears is that Lucen Rich McInn >> that Jensen is Rich McGin and I'm just I rebel at that. Jensen Jensen Wong is a visionary who is seeding a lot of different companies because otherwise they wouldn't have the capital. He's knows that fiber is the way. >> He does like 50 billion in free cash flow in a quarter. >> Well, and he's not throwing the money away. He's not putting in an incinerator. He's giving it to companies. >> If he wanted, he can just turn off his $150 billion buyback. What's he going to buy back $150 billion over the next few months? I've been calling for half throw some money this way. >> I've been calling for half trillion. >> I mean, it is the National Bank of Nvidia. Yes, it is. But the National Bank Nvidia was up four. Now there's only one because people were too excited. >> The question about Nvidia, Jim, is how long can they maintain those incredible margins? What are people saying going to really come for them? >> Well, where are Look, I I'm a big believer in Lisa Sue. I think she does a remarkable job. There was a rumor this weekend that they she's having the GPU prices are going down. I think that rumor is completely false. But it doesn't matter because there are so many people who are non-believers right now. Uh, and you couple them with the the incredible amount of money going against you. Here's a little 99.7% of the dollars spent in this election are uh campaigns critical of data centers. 99%. You is it time to stand up in the data center? Maybe take the other side a little bit. Do we think that what what we're witnessing in the market right now is is based on the fundamental story in any way because everybody who's coming out talking about it, Jensen Wong, everybody else is like there is no sign whatsoever in waning demand >> for any part of this story, >> including your interview last week with Brad Gersonner. Um, so there are a lot of people out there who are inside uh the so-called beltway of Silicon Valley, if I can use that expression, who are saying that this isn't slowing down. The questions are there about where peak spending will be and we're going to get some catalysts coming up in the next few weeks particularly when the hyperscalers report Scott >> and that is this week now. >> Okay. And you know what? But that's the point. The next two weeks you're going to get some catalysts that are likely to say the following. Number one, spending isn't slowing down. Yes, of course at some point it is going to slow down, but they're likely to say that it's not on the near-term horizon. Number two, what they're also likely to say is that the growth rate is still constrained by the supply of commute. That's what they're likely to say. The last thing that they're going to say, we got some insight from Andrew Jasse last quarter about how quickly they get payback on these data centers. Now, I'm talking not about the Neoclouds. I'm talking about the bigger the biggest scaled players, and I'm looking forward to what those comments are now. Both about the demand that's out there and how quickly they pay back just on the data center built. I'll also point out some other catalysts that we think are coming. Now, we're predicting the future could go ary, but we think Anthropic is going to go public sometime before Thanksgiving. Um, when that happens, and I honestly don't care if it's one and a half trillion or two trillion. The point is it comes in, they get money that they continue their build out with. Oh, by the way, OpenAI is doing its private funding round right now. These are catalysts that may put to bed this question of whether we're at peak spending. I told you guys. >> Here's the example. Jimmy, you bullish on Micron? >> Yes, I am. >> Steve, you bullish on Micron? >> No. >> Okay. >> I don't know. I'm sold,000 bucks. >> I'm bullish on Micron. I'm bullish on the semi- equipment cycle. Jimmy, when did Micron have a temporary peak? June. >> Yeah. So, it's And has the demand slowed? Has memory pricing slowed? No. But if you wait until everybody rings the bell that it's slowing, then you're going to sell at a lot lower price than you want. You can't wait. Why do I have to sell? That's the That's the question. It's the It's the great example of Nvidia. It goes through a period where you're frustrated. It goes sideways for 6 months and then it revives itself. That's what I'm telling you is going to happen with your microns, with your semi- equipment news. >> You watching Lamb? >> Yes. This week they report middle of the week. >> It it it reports this week. It's going to be a great quarter. The stock has not traded well. It's down 37% off of its June high, but it's up 62% year to date. >> That's fine. But Joe, you're a student in the markets. I'm student. Jimmy is Scott. Everybody. >> Isn't this sound so eerily familiar that Micron six times earns? It's so damn cheap. And their order book is so big, right? And that they can't meet all demand. Have you ever heard that before? Hold on. I've referred it about 10 times you my career. >> There's one big difference. Your point is well made. But here's the key. Big difference at the end of >> this one. This one matters. They have no debt. Micron, they have about 100 billion in cash. They're generating 130 billion in free cash flow. And what are they doing with it? They're recapitalizing the company by buying back shares at $6. >> How much debt do you think they had in prior cycles? >> A lot. Jimmy. >> A lot. >> Jimmy, it doesn't matter when the order books dry up. >> Jimmy, let me ask a question. Do you think the universities are going to dry up >> in ter dry up in terms of what they believe? I'll phrase it a different way. Their take or pay contracts for 40% of the revenue 27 if the market if the if the spending peaks those takeay contracts mean zero like in every cycle. >> Listen, I'm not going to dispute that with >> but that's what the bull case says. Half their revenues are guaranteed. a simple question that we're disagreeing on which we have rights to do which is where is the peak where is the peak and again if we look >> I'm saying I'm saying it's prudent to think a year and I would agree with you that it peaks a year before spending peaks I agree with you what I'm saying is this cycle and it is a cycle okay is longer than that >> what happens if the demand story isn't close to peaking at all >> isn't that what the conversation is about that's >> Isn't that why don't Isn't that why Jan I is out today suggesting that what we're seeing now is no different than the you know the the so-called SAS apocalypse but now it's just become now we're worried about the this >> the answer and demand isn't waning at all >> the answer to your question which is the question lies somewhere between a year from now and four years from now most of the projections are this spending goes through the next four years I find that a little bit hard to believe but equally and just based on what I hear from managers of these companies I grant you Steve biased of They have their perspective. But what I'm hearing from them is it's not one year either. It's longer than that, >> you know. But but keep in mind, I'm not selling all my AI stocks, right? I've just reduced my exposure to the group. >> So you've reduced you you've reduced Meta. Have you? >> No. Meta actually, you know, I took the opportunity to increase the size. >> But what's the difference between that and the others? >> Because I think Meta is a product company ultimately. So they'll benefit. They're on the other side of the capex cycle, right? So of course they're going to spend. They have a unique ability to spend given their balance sheet. They're also going to benefit from what they're spending in terms of their products they offer to the public. >> So if you're reducing your exposure to AI stocks, which ones are we talking about? >> We talked about Micron. We talked about Verdive. You know, Micron was a really big position. It's gone. It's been gone for a few months now. >> You own Nvidia. >> Again, small position. Relatively small position. But you haven't you haven't se you haven't reduced that, >> have you? >> Uh, yeah. At one point I got out and then I got back in lower prices. So, yes, >> and it's smaller now than it was. >> Taiwan semi is still probably my largest position. >> All right, I hope you're all doing well today and staying calm in this market. Monday was a mixed day in a market with some green action in many of the large tech hardware stocks that sold off heavily last week. Over the weekend, Soul Economic Daily published a report saying that Nvidia has requested that Samsung SKH Highix and Micron manufacture the base dies for its next generation NVHBM using TSMC's 3 nanometer process. Nvidia also wants sixth generation DR RAM technology incorporated into the memory. Samsung reportedly agreed to Nvidia's requirements even though it normally relies on its own foundry operations for HBM production. Samsung is reportedly evaluating its own foundry as a possible alternative if TSMC's capacity proves insufficient. Put simply, Nvidia is exercising greater control over the architecture and manufacturing requirements of memory used in Nvidia systems. As a reminder, Nvidia's NVHBM is being positioned primarily for XBUS through Envy Link Fusion. Envy Link Fusion allows companies to integrate their own custom accelerators with Nvidia systems. That way, if another company develops an accelerator that's better than Nvidia for specific workloads, they can simply integrate those XBUS with Nvidia systems and remain within the NVIDIA ecosystem. Over the weekend, it was reported that Microsoft has pulled out of ancale data center deal in the UK due to power connection delays. The site was expected to begin operating next year with roughly 23,000 Nvidia GPUs. A Microsoft spokesperson said that this decision relates to a single site. So, this is a situation of power delays rather than a situation of diminishing demand. The good news for Nvidia, as Jensen mentioned on a recent episode of the BG2 Pod, is that they have such a large customer base globally that they can simply redirect compute to the sites that are built and powered. And speaking of in the scale, Bloomberg is reporting that the company is in talks with Anthropic for a compute deal worth about $45 billion. Importantly, this is in addition to the $45 billion deal that was announced between the two companies back in August. And the newly reported deal involves the same West Virginia site. So, we're essentially talking about a doubling of the deal that was announced back in August. In other news, Nvidia and German officials signed a joint declaration of intent focused on developing advanced AI models for physical AI and expanding AI adoption throughout German industry. Jensen also announced Nvidia's investments in four German startups. This is all happening ahead of GTC Berlin this week. Now, let's cover some memory news. We have an important update to the Micron Taiwan labor union situation. As a reminder, there are two different labor unions involved in the situation. The first union met with Micron. Conversations were constructive and a second mediation was scheduled for October 22nd. That's this Thursday. The other labor union met with Micron. Mediation was unsuccessful and that union voted in favor of strike authorization. And now we have an update. The union that approved the strike authorization will hold a solidarity strike in support of the other labor union on October 22nd while mediation is ongoing. Importantly, the union says that this strike will last for just one day. The bigger piece of news this week will likely be what comes out of the mediation that takes place on Thursday, October 22nd. A one-day strike isn't really a problem for Micron. If there were a prolonged strike stretching over the course of many days or weeks, that would be a different situation. But a 1-day strike isn't a major problem. So now, let's see what happens later this week following Thursday's mediation. Hopefully, we'll get positive news. As a reminder, Micron revealed in an SEC filing earlier this month that the board approved an increase to the share repurchase authorization of $33 billion. That's almost all of Micron's reported quarterly adjusted free cash flow in the most recent quarter. Micron said they intend to return 100% of excess cash to shareholders, primarily in the form of buybacks, and they appear to be delivering on that promise. Micron should be allowed to conduct typical share repurchases starting in December after restrictions related to the chips act subside. Micron's share repurchase authorization should increase further in future quarters as the company generates additional free cash flow. It is very plausible that Micron could buy back 10% or more of the company's current market cap over the course of a full year. That's very good. Also, on Monday, Micron announced that it will host a technology webinar this Wednesday, October 21st, starting at 8:00 a.m. Mountain time. That's 10:00 a.m. Eastern, 7:00 a.m. Pacific. The webinar is to be led by Micron's president and chief technology and products officer as he discusses innovations across DRAM, HBM, and NAND that extend Micron's technology leadership. The press release says, quote, the webinar will detail how Micron's technology foundation, multigenerational roadmap, closed customer, and long-term research position the company to extend its technology leadership over the next decade and beyond. There's a lot happening this week. Looking ahead, Jensen Hong is scheduled to deliver a GTC keynote at GTC Berlin on Wednesday, October 21st, starting at 11:00 a.m. Central European Summertime. For those of us in the US, that's October 21st, starting at 5:00 a.m. Eastern, 2 a.m. Pacific. And then we have Hypers Scale earnings with Meta earnings this Wednesday after market close. Alphabet and Microsoft earnings are both scheduled for October 28th. And then we have Amazon earnings on October 29th. What these companies say about capex, capex guidance, demand monetization, and their overall AI strategies will have important impacts on how the stocks of Nvidia, Micron, SK, Heinix, and the entire AI ecosystem trade moving forward. Let's briefly go through the four major hypers scale companies. Last earnings season, Meta said they expect 2026 capex to be in the range of 130 to $145 billion, and they did not give a guide for 2027 capex. Since that earnings report, Meta launched Muse. The app rose to number one on the app store. User reviews are overwhelmingly positive and early adoption rates are very strong. Meta is also showing impressive progress with multiple model releases from Meta Super Intelligence Labs as they continue on their scaling ladder. Because of these developments, many market participants are starting to believe that Meta will generate a return on their massive investments in AI infrastructure. That was not the case last quarter. I would not be surprised if Meta provides a 2027 capex guide that is notably higher than 2026. If that happens, it will be interesting to see how the market reacts. Last quarter, Alphabet said they expect 2026 capex to be in the range of 195 to $25 billion. And they said, quote, "We continue to expect our capex to increase significantly in 2027." Google Cloud has been showing very impressive growth with revenue up 82% last quarter. Alphabet has a strong incentive to increase capex and the company has many valuable use cases for comput. I expect Alphabet's capex guide to be strong given their previous comments, GCP's impressive growth rate, and the fact that Alphabet has so many different use cases for the compute they bring online. Now, let's talk about Microsoft. So, last quarter, Microsoft said it expects Calendar 2026 capex to be roughly $175 billion, which was down from the previously announced $190 billion. The downward adjustment to the capex guide was made because moving forward, more of Microsoft's future data center leases will shift from finance leases to operating leases. Operating leases are included in opex, not capex, which is why Microsoft reduced their capex guide, even though they still expect to spend money. Microsoft also said on the last earnings call that they expected fiscal Q1 capex. So capex in the quarter that they report results for on October 28th to be over $50 billion, which is notably higher than the previous quarter's $41 billion. Microsoft also told us last quarter that they expect fiscal 2027 capex to increase year-over-year given the strong demand signals across the portfolio. And the big news last quarter was that Microsoft said it expects to remain free cash flow positive in fiscal 2027. Now, I do expect Microsoft's capex guidance to be strong given the strength we saw from Azure last quarter, leadership's commentary on the earnings call, and the fact that Microsoft is really making a push with their own AI products and services. You may remember just last month, Microsoft announced a major overhaul of Microsoft 365 copilot with home code and autopilot. And they also announced usagebased billing for more computensive AI features. So, Microsoft needs more capacity to fulfill its massive cloud backlog worth hundreds of billions of dollars, and they also need a lot of capacity for their own models, products, and services. I expect Microsoft's capex commentary to be positive overall. And then there's Amazon. Last earning season, Amazon CEO Andy Jasse laid out a great case for how Amazon is monetizing new capacity as soon as it comes online. They have clear line of sight to strong financial returns on their investments, and there is far more demand than what they can supply. AWS is the largest cloud, by the way. Last quarter, AWS hosted accelerating revenue growth and expanding operating margin, which is very impressive given AWS large size. Amazon said last quarter that they expect to spend $220 billion on capex in 2026. And CEO Andy Jasse said, quote, "Even at that amount, we will still not have enough capacity to meet all the demand we have in 2026. And I believe this dynamic will also be true in 2027, too. In fact, the demand we already have for 2028 is striking." And more recently, Ben Horowitz joined the All-In podcast in an episode published on October 16th in which he said he just recently spoke to Amazon CEO Andy Jasse. Listen to what Ben said about that. >> And this this this shortage is so bad. You know, I was speaking with Andy Jasse who runs Amazon yesterday and I said, "Yeah, like we're out of everything, you know, till 2028." He's like, "Oh, it's beyond 2028 bit. Like I don't think we get stuff back like end of 2028 maybe." Uh so everything else is sold out. So he he has the advantage of that he's got supply in a great shortage which is extremely different than the internet bubble. The internet bubble we didn't have enough demand because we had to build the internet. People had to get broadband you know you needed uh all this stuff. People had to buy you know originally they had to get their browser app fries all this kind of thing. There weren't enough consumers on the internet to absorb all the bandwidth that had gotten built out. and they had a massive amount of debt, not unlike uh we're building today, although not as big. Um but but the lack of demand killed it. We don't really have a lack of demand in AI. So I think the dynamics will be different. >> All things considered, I expect Amazon to have very positive things to say about capex on the Amazon earnings call. Really quick, I want to address something important. You may have noticed a recent report going around from a particular firm saying that by 2030 most Neoclouds will cease independent operations. I noticed an analyst behind that report speaking on CNBC Monday in which he claimed that we have far more supply today than we have demand. And he pointed to recent price cuts by the frontier model companies as evidence that we have more supply than we have demand. Respectfully, that is inaccurate. As I've said many times in the past, at this current stage, profitability is not the Frontier model company's priority. Right now, they're prioritizing building a large user base, and capturing as much market share as possible. Building a large user base, a large customer base, capturing market share, that's the hard part. Once you get to massive scale, optimizing for profitability is relatively easy. Meta is a great example of this with the recent success of Muse. They already had the large user base, and so it was only a matter of time before they found additional ways to monetize that user base with AI. Now, going back to the Frontier Labs, they are prioritizing growth and claiming market share over profitability at this current stage, which is what many of their long-term investors want them to do right now. We need to be careful not to conflate reductions in token prices with a lack of AI demand. Additionally, with each new generation architecture from Nvidia and software optimizations, token prices will continue to fall. Nvidia and other hardware companies stand to benefit greatly from this. As token prices fall, it catalyzes greater usage. More developers are able to build more and more workloads become economically viable. Whether you're generating closed source frontier or open- source/openweight tokens, you need the necessary infrastructure to generate the tokens. That's where Nvidia comes in. This is why even as token prices have gone lower. Usage has increased notably and GPUL prices have remained robust. You may remember open router experiment not that long ago which clearly showed that usage increases drastically when token prices are reduced. And notice this, as usage increases, you need even more compute for inference in order to serve that demand. And as I've said before, a key fundamental difference between this AI revolution and the shale revolution is the number of use cases. If we're just talking about oil by itself, then sure, there are many use cases. But when compared to AI, it's not even close. With AI, there are already thousands of use cases and potentially millions of endpoint applications. As token costs move lower, new use cases will be unlocked, resulting in greater usage. And some of those use cases are very valuable. What we saw earlier this year with a decoding is a great example of this phenomenon. Anthropic and Open AI have both benefited greatly from the explosion in aic coding. And that's just one example of a valuable use case at scale. There will be others. And one final point on the frontier labs. When token costs come down thanks to new generation architectures and software optimizations. It's up to the frontier labs what they do with that reduction in cost. They can either pass the benefit along to their users. They can pocket the benefit and improve profitability or they can reinvest in a business. Right now they are prioritizing growth and capturing market share over profitability. Do not conflate reductions in token prices with a lack of AI demand. That's not how this works. 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 compute constrained 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.com 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 do-com 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 dotcom 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 compute constrained 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-t 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 spinning will reach 3 to4 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 Finn Vid. 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