CNBC On NVIDIA Stock, Micron Stock, SK Hynix Stock, Memory News - NVDA Update

And the latest comments from Broadcom CEO Hawk Tan adding more fuel to the fire. Speaking at a conference yesterday, Tan said, "A lot of investors have no clue what they're investing in and set high expectations only to find out they're in a casino gambling." Joining us now is Gil Laura from DA Davidson. Try to wrap this all together. Uh Gil, you you've had this overlay of skepticism on the semis for a while. It's not because earnings estimates have stopped going up. It's not because anybody thinks the overall pie is is not growing into next year, but unwilling to to maybe give credit upfront for for how long this boom is going to last. >> Yeah, that that's exactly right. Or at least it's giving the credit in an uneven manner because let's be clear, all the results that have been reported over the last 8 to 10 weeks have been not only positive, but companies across the value chain have said that results have improved, that the outlook has improved, and the visibility has improved. That's everybody from the chip uh companies that make the the chip equipment, the companies that make the chip, the hyperscalers, the model companies, Anthropics going to show up tomorrow and and give us probably a very positive outlook as they prepare for their IPO. So the news, the reporting from the companies has been very consistent. that co that comment from Hawktown was in the context of the the open AI mishap that happened last week where everybody uh had to dust off the definitions of gross revenue and net revenue and that's all really that happened. He was commenting on the fact that if anybody sold because they misunderstood the difference between gross revenue and net revenue. They're behaving as if they're in the casino. So the commentary from Broadcom, including his during the quarter and and over the last couple days has been consistently positive. results have gotten better, outlook gotten better, visibility's gotten better. >> Okay. Um, but let's talk about why the the chips are down at this point, Gil, and specifically focusing on the memory space. I mean, one headline that really jolted this particular subset of of chips is CXMT coming out with a denser wafer by the end of the year. I mean, this notion that China will be able to scale very quickly. um it had 9.5% of global uh RAM memory u RAM revenues uh in the last quarter versus 7.2 I think the quarter before. So it gained two points in in revenue share and I'm just wondering you know is that something that investors should factor in that sort of China dynamic where they can figure it out they can scale pretty quickly and they can sell it for cheaper. >> That's right. we definitely have to factor for it and that is what riled investors uh in the Korean memory and in Micron today. Having said that, when we say that we think that memory is going to continue to grow over the next couple years, China has been in the math and continues to be in the math. These companies aren't new. They're newly public, but they're not new. They've been catching up on the technology. Having said that, they only participate in the low end of the market. They do not participate in the high bandwidth memory part of the market. And by the way, they can't sell to American companies, right? Let's not forget Micron's biggest customers, Microsoft uh Micron's biggest customers are Microsoft Amazon Google Apple Nvidia. None of those companies can buy Chinese chips, right? So, if there's an impact, there's some impact on Highix and and Samsung. It's not on Micron. Having said that, the Chinese are catching up. The other thing to keep in mind though is that the discrepancy between demand and supply exists in China as well. meaning these Chinese memory companies can't supply enough chips to meet the demand in the domestic Chinese market and they won't be for a while. So again, it's been in the math when Micron says that the tightness is going to get worse next year and even worse the year after that. They are fully aware of what CXMT is developing, their product roadmap, what they have, what they don't have, and how many clean rooms they're going to open. Gil I have to imagine I mean it's impossible to kind of isolate this but the psychology of investors is definitely also aware of the push back on the pace of data center development and just how costly memory is in the entire kind of expense stack out there and can they possibly maintain those economics you know beyond uh the the several quarters we have ahead of us. So, I'm looking for ways to explain Micron under six times forward earnings, which is super cheap, but also where Micron has tended to trade at cycle highs in the past, >> right? And and that's the that's the big discrepancy in the market right now is the fact that you have let's just use memory and CPUs is as the two markets. The two markets look almost identical in terms of the competitive nature in terms of the exposure to the eye cycle and yet you have AMD and and Intel trading at at 40 times and above and micron is six times that is a really big dislocation. The reason is that people even if they just remember 2023 think of memory as a commodity that's that has its own uh cycle. That is no longer the case. And as investors wrap their head around the fact that memory has a completely different role in the data center than it did in 2023. And that means that it's getting incorporated into its large customers. Long-term planning. They're signing long-term deals. That changes the competitive nature. So Micron no longer sells a commodity. A and B, the cycle that impacts Micron is the cycle that impacts everybody else. We still think that as the market wraps its head around that, there'll be a convergence in multiples between AMD and Intel on one side and Micron on the other side. >> So, it's just a matter of time before the market realizes all of the the bull case themes that you had laid out, Gil, because I mean the socks is still down 18% from June highs and you had laid out the case before that the the companies have been delivering. They've been delivering consistently. The message has been consistent. The stocks have not responded. What do you tell investors? Yeah, I mean I think part of it, you mentioned psychology earlier. Part of it is psychology. So, uh, first the investors that bought Micro at 100 and sold it at a thousand think they're done and they're very happy with it, what it contributed to their performance. Then investors that missed it from 100 to a,000 think that they missed it and they don't want to touch it. That's psychology. as as the company continues to deliver as it becomes clear that again it's one cycle it's a non-commoditized product people will wrap their head around it the reason I'm confident that I've spoken with dozens of investors that are in the process of doing so let's not forget Micron was a small nichy stock as as late as a year ago now it's a trillion dollar stock and and it's not just semianalists that are looking at it generalists have to look at it it's only 3% of the NASDAQ 100 by market cap, but it's 10% of the NASDAQ 100 earnings over the next 12 months. So, if I'm right and they end up trading at a market multiple a year from now, that's seven points of performance from one position. Everybody's doing the work and I think they're going to come to similar conclusions. >> Uh let's now get to the uh dramatic or dramatic moves in memory stocks today. The group under continued pressure from China's CXMT and concerns that cheaper, denser memory chips will threaten existing players. A now deleted expost from Elon Musk adding to the concerns. He warns that China typically can scale production faster than anyone else once it figures out a product. Though he noted it might be harder for CXMT to compete in high bandwidth memory specifically, which is where Micron focuses. The major memory players all up triple digits this year, but they are well off their record highs hit back in June. So, what does this all mean for the direction of the AI trade? Let's bring in Gene Munster, managing partner of Deep Water Asset Management. Gan, great to see you. Hello. >> We've had the concerns about Chinese competition before, particularly when CXMT went public. Um, I'm wondering from your standpoint, is this just sort of the context in which the concerns are being raised once again? Because the context here is that there are all sorts of jitters surrounding semiconductors in general, which are down 18% from their June highs, and memory specifically. So Melissa, I mean the big picture here, we can actually go back to January of 25 when Deep Seek came out. That was kind of this concept of a magic bullet that was going to fundamentally change the need for the most cutting edge frontier models. We did not see that happen. What we're hearing here is this, as you mentioned, kind of the concept that memory will need about a third less amount of memory because as as you said, it's more dense falls into that same category. And I'd put this conversation in the category of the magic bullet. And the reason why we're talking about it is because the underlying AI trade is nervous. And so when you sprinkle on these data points, last week the nervous point was around open AI and their revenue recognition. This week it's around what's going on on memory. And so then you have Elon step in here and try to calm people and basically reassure that it's not that easy. And as we go to the high bandwidth memory and more sophisticated like GPUs and the cooling pieces, it's going to be much more difficult. So what do we make of all this is first I I think investors should recognize the difference between weather and climate. And the weather piece here is this nervousness. There is no magic bullet. There's no magic bullet around memory. There's no magic bullet around frontier models uh for being that top 1 or 2% intelligence. It needs to come from closed proprietary models. there's no magic bullet around power and so I think uh just knowing uh that piece I think should be somewhat reassuring and there's a second piece around the climate and this is what's more important is that ultimately the question comes down to let's take out of the the nervousness and what the story of the week is that's going to get investors nervous the big picture and the big picture is we're still early and how early we haven't even started on personalized AI these bots that have come out in the last few weeks they take 10 to 15 times the amount of tokens is a typical chatbot. That's going to be a massive increase in inference and compute that's going to be needed to power that and physical AI hasn't started. And so each of those whether it's the personalized AI or physical AI is as big as the whole digital AI that we've seen over the past four years. So uh just make the distinction between weather and climate. Weather's pretty choppy right now. The climate still looks really good for the AI trade. >> So Jean, I'm gonna start where you just left off. So I think about it as in five years from now there's going to be dramatically more agents. Agents are run on tokens. Tokens are uh inference. Inference is memory. So the best memory to buy in in my mind is micron. Exactly. And it it's sometimes the most simple answer the obvious one is the best answer. And in this case u I think micron is in a great position for that. Again, I believe uh and internally at Deep Water, we've had similar conversations that Elon's tweeting about today is just that it's just much harder to do all this. So, I would agree with you, Steve. Well, look, the most confusing part of the market right now, though, isn't the isn't the banks, although they are trading so awfully. It's the once red-hot semiconductor space. >> The group is so visceral that you figure that they're going to cause pain even before trading starts. >> The house of pain. And that's exactly what happens is these stocks are so behind the eightball and they get crushed from the moment that the Cosby the Korean index gets annihilated which is in the middle of the night which is both the time when I get up and when I go to sleep. So I see the damage happening. It's being exported right here. For the past six days the semis have experienced one of the most harrowing gains on record. And that's rolling right into the averages. The SMH the VANX semiconductor index has now uh fallen 7.22% in these last six days. That's the longest losing streak since May of last year. The group has always been streaky with 10 longest losing streaks in the last 25 years, going from minus 33% back in 2021 to minus 8% in May of 2006. In short, when the semis roll over, it tends to last for much longer than you might hope. So, don't presume we're about to have a serious bounce after the worst streak of the year. That said, the last three losing streaks did produce some spectacular gains. uh at least when they were through. The estimates was up 17% since the last week ended this July, up 73% since the previous one ended last December and up 148% since that one before that, which ended in May of 2025. So there is reason to be hopeful. The problem here is that the SMH ETF has wrapped up every important semiconductor in the industry. So they all trade together led by SKH Highex, the memory giant that trades both here and in Korea. What can you do? SKHEX is so big that it's become the key to the semiconductor solar system. Not Nvidia, SKHEX. The fact it's been down, down, and then down tells you all you need to know about the cell being exported from Korea to here. Our chipmakers are joined at the hip with SKHEX, not Nvidia. So we come in each day with a huge handicap because we know the initial direction of this important group and it's mighty hard to reverse that direction because it's controlled by SK highex. The problem is when you have this kind of joint at the hip relationship, we tend to forget why the heck we ever would own these stocks to begin with. But let's think about that. Let's think about what they really do. Today I went to this incredible Apple showroom where I saw interactive devices tied to my phone bound by Siri that uses knowledge about me that I give it but I give it to no one else. It's a personal assistant that knows my predelections and it also knows my messages and it has what Apple calls calls broad knowledge. In practice, it makes me want to rip out the intrusive Alexa who's in my kitchen who thinks that Vagner is Wagner and she tried to sell me another kind of soap because I just bought a soap and welcome this new version of Siri into my home that doesn't want me to do buy anything. I like that there's a beautiful HomePad mini that can recognize up to six users in a household by their voice. Exciting showroom. Get in sync with the doorbell camera to show who's there. Downward-facing camera showing you what they've brought for you. I love this stuff. I want everything. I think about Apple and I know that the stock's a buy ahead of the iPhone Duo, the foldable two screen version. But what I don't think about is how much memory and CPU power goes into these kinds of devices. Let's talk memory for a second. After a lengthy period of massive outperformance, this group has been the most beaten up and most discouraging. It's driving people crazy with Micron at the heart of the disappointment. We read yesterday the Chinese might have a better memory trip and that could crush the entire group or that Microsoft's getting in the chip. Whatever. But let's use this let's use this Chinese example that they're getting in it's going to hurt Micron. This is nonsense. See, no US-based company of any repute will accept Chinese memory without worrying about confident confidentiality. This is just stupid. Who writes these stories? Why they go to college? It's stupid. I MEAN, HONESTLY, WHY isn't there an editor saying let's kill this? I mean, is it that important to be wrong? Do you really have to fill her up like it's a gas station? More important, do people not know how many long-term contracts Micron has with its customers? How about 26, which uh cover revenue until 2030? Take or pay? You know what that means? It means you have to pay. Micron's currently negotiating for 2031. If the Chinese competition were so serious, why would anyone take these long-term leases? Are they idiots? Still, Micron stock is chained to SKH highness because they both make memory shifts. But SKH highest business is much larger on Microns. But I say these long-term contracts put them in a much better position. If Micron had no agreements, then indeed I would be worried. Regardless, the point is to remind you where these chips go and the desperate need for them. A market contrast to what the stocks are saying. The biggest disparity comes from AMD and Intel. These two companies make the CPUs that are basic building blocks for practically every machine under the sun, including Apples. Of course, now Apple doesn't necessarily use their chips, okay? Although I think ultimately they may use some some Intels, but we know you can't make devices without these chips. Use Intel's foundry, I should say. And Apple's going to make an awful lot of these devices. These are going to be big sellers. It may not matter though to the stocks, though. Maybe we have another down day for the semi-index. There have been 10 big losing streaks since 2000. It's reason to expect a sharp pullback after how far these stocks traveled. Ultimately, I just think that when you see the devices, when you know the exploding use cases, you can understand why Micron or Intor AMD has been moving up wasn't all situational awareness that Wayward Hedge fund that took so many semis up to unsustainable levels with borrow money. A lot of it's just raw demand and that's going to include Micron even if it's joined the hip to SK highinex. So, here's the bottom line. Before you decide to bail from the semis because you can't take the pain, >> the house of pain, >> go to see the showroom. Go see where these chips are really used. Go understand what really happens in this world. They appear in places you'd LEAST EXPECT THEM. DON'T WATCH THE semiconductor index. Watch the devices. Watch the business. It may turn out that you may like it and not just the action. >> The house of fame. >> All right. I hope you're all doing well today and staying calm in this market. There was a lot happening in the market on Tuesday. We closed red at the index level, but it was a mixed bag beneath the surface. It was a rough session for the Cosby overnight in South Korea. Investors were nervous. On Tuesday, ahead of an investor meeting at Anthropic that's scheduled for Wednesday, Cipher Digital announced a convertible note offering which put pressure on many of the NEOCloud stocks. And on top of that, we had everything going on with oil rates and Fed commentary. There were many things affecting the market on Tuesday. We got some Fed commentary around 2 p.m. Eastern that helped lift stocks after New York Fed President John Williams said that one more rate hike in 2026 may be appropriate, but further increases in 2027 are unlikely to be necessary. Those comments helped some stocks recover slightly Tuesday afternoon. I addressed the weekend story about CXMT in Monday night's video. I think the date of that video is actually listed as October 12th here on YouTube because I accidentally posted it slightly past midnight. Sorry about that. But you can watch that video if you want to hear my thoughts on the CXMT news. I'll just briefly mention that while CXMT made news regarding its claims about a 4F squared memory cell structure, there's still many unanswered questions regarding things like specifications, competitive yields, and high volume manufacturing among others. Samsung and SKH Highix have both already demonstrated 4F squared DR RAM and they could transition to a 4F square memory cell structure today if they really wanted to do so. But it doesn't make much sense for them to do that yet because of challenges surrounding yields. And so the media framing that CXMT has surpassed the big three memory makers is inaccurate and lacks important context. And as I've said many times, CXMT is not able to fully satisfy the memory demand in China, let alone the rest of the globe because the demand is so great. A 4F squared product announcement from CXMT is not the same thing as competitive yields, high volume production, or large quantities of near-term supply coming online. That's not what this past weekend's CXM news was about. I do think the recent sell-off in memory stocks is largely overdone and it has been exacerbated by inaccurate media framing. In other news, Jensen Hong and Nvidia are in Europe ahead of GTC Berlin. Nvidia announced planned investments in four French AI startups, while Jensen said Nvidia expects to announce approximately $2 billion in additional European investments over the coming week. Nvidia has committed nearly $10 billion across Europe's technology ecosystem. The European trip includes Paris, London, and Berlin ahead of GTC Berlin on October 20th through the 22nd. As a reminder, Jensen's GTC Berlin keynote is scheduled for Wednesday, October 21st, starting at 11:00 a.m. Central European Summer Time. For those of us in the US, that's October 21st, starting at 5:00 a.m. Eastern, 2 a.m. Pacific. I'll try to cover the important details in a video on the night of October 21st. NVIDIA also published a technical blog post on Tuesday saying that NVIDIA RunAI is now open source as part of Nvidia's DSX AI factory platform. This will give customers, developers, and ecosystem partners access to production grade AI workload and GPU orchestration capabilities. The release includes scheduling, GPU resource sharing, organizational resource management, and workload management. This is especially important because by making runai open source, Nvidia is enabling the broader ecosystem to bring differentiated solutions to market faster as customers needs are evolving. Nvidia is so much more than just a chip company. As for memory stocks, in addition to the weekend's CXM news, we also got some news about China's YMTC. I noticed Counterpoint Research published a piece saying that YMTC has filed for Shanghai IPO and the chipmaker rose to number three globally by NAN shipment volume trailing only Samsung and SKH while leading Kioia and Micron. YMTC is benefiting from the enormous increase in demand for enterprise SSDs for AI data centers. I haven't seen any update on the Micron Taiwan labor union situation. As I mentioned last week, Micron revealed in an SEC filing on Friday that it increased its share repurchase authorization by $33 billion, which is nearly all of the company's quarterly adjusted free cash flow reported on Micron's most recent earnings report. The repurchase authorization should increase further in future quarters as free cash flow increases. It's very plausible that Micron could repurchase roughly 10% or more the company's current market cap over the course of a full year. That is very good. 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, 2:00 a.m. Pacific. 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. Agent 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 revenue 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 aenic 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