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

So, do you recalibrate who winners and losers are in the AI race now? >> Well, I mean, we've we've said all along the biggest winners right now, we are still very much everybody wants to move on from the infrastructure trade. We are not moving on from the infrastructure trade. This is this is still about building as much infrastructure, data centers as possible to service the demand that's out there. The >> chips, >> yeah, chips, memory, uh connectivity, power, labor, everything. We we are in this place where we have and I' I've talked about this here before. We have this ear this this vertical wall of demand hitting a horizontal line of supply. And when you get into this sort of irresistible force and movable object kind of point, something has to break. And that's pricing. And so everything's getting more expensive. The chips, the labor, the power, the financing is all getting getting more expensive. And that's that's what's driving. >> What's not getting more expenses is the token costs, right? >> Those are coming down on the other end. So does that I mean is that a concern that there's the the buildout is still going up the cost while the token prices are going down. >> Well, we need token prices to come down like to you know this this is a technology technology is supposed to get cheaper over time and particularly if you're trying to drive enterprise adoption with this. You need token costs um to to to get cheaper. They should be right. We've got much more performant hardware. The current generation of chips are 35 to 50 depending on whose numbers you trust. X more performant than the prior generation on a per gigawatt basis. And so that increase in supply should drive overall token prices down. It's the consumption. It's how many more tokens we're all burning as companies um that's really going to going to drive that. So you need the per unit cost to come down, but what you take some level of comfort in and you see this in the Frontier Labs ARR numbers, you're getting a increase in volume from that. The whole Jevans paradox thing. you're getting an increase in volume from that that more than abstracts any of those price declines. >> If you play that out, the whole Jebans paradise thing, people use more when the price comes down, etc. Does the pie of profits continue to grow fast enough for everybody to get their share? And I I asked that because it happened in wireless data and fiber optic, you know, broadband and all the rest of it. And that infrastructure layer has not on a long-term basis been where a lot of the profit bonanza has settled. >> Yeah. and and it comes down to this being a very different technology and eventually we will get to that point in this cycle as well. We're just not there yet. There's minimum three probably more like 5 years more to go on the infrastructure side of this where the infrastructure builders are able to get real premiums for the infrastructure they have just because demand is so far um outside of supply. Eventually we'll get to some level of of moderation or equilibrium around that. You'll see margins. I mean to be clear, Micron's not supposed to be a 90% gross margin company. Like, you know, you are supposed to see kind of more normal margins here. We'll get to that, but it's it's a ways out. >> Yeah. And arguably, you know, Micron's down 250 bucks from the high. The market is kind of, you know, realizing traveling people. >> If you don't think that AI supremacy equals national security, listen to this story. It's wild. A new report reveals a suspected Chinese spy developed a close relationship to Nvidia CEO Jensen Wong. US officials concluded the suspect was funneling information and potentially diverting advanced chips to China. Now the information is reporting that Chinese labs Deep Seek and Moonshot AI relied heavily on OpenAI's powerful GPT6 Astra to train their latest models. Mike, I want to get your reaction to this. I mean, this is corporate astron its finest form, right? except it's the Chinese government. Uh but but also Deepseek, they also uh excuse me, Anthropic said that they found 16 million exchanges with Claude uh with these Chinese companies as well. Deepseek and the other uh we we knew that they were trying to steal our technology trying but but and then they got the chip story with Wong which is wild as well. >> I would say get used to them stealing our technology because they've been doing it and it's going to continue. But as far as them sending a spy, um, you know, sending if you're sending it for someone, um, I I'll bet on Jensen Wong. You know, I think he's, uh, been around the block a few times. I I doubt very highly that he was fooled by this person or was giving trade secrets to this person. I mean, there there's probably a lot more to this story that won't be revealed, but um, it it's how they operate. Um, you know, and they'll continue to operate that way. So the idea of having more manufacturing of chips, these advanced chips that right now we can't manufacture in the United States, the more we can do that here with within our own country, the better for all of us. >> Well, what Lou, one of the things that this information reported though is that she had a quote close relationship with Jensen Wong. Uh, and that also the fact that that she was taking all of this surveillance and all this information and taking it back to the CCP. Uh, my question is, do we need to be having trade talks with the Chinese? Does Treasury Secretary Bessant really need to be once again extending talks with with >> We have to. I mean, these are this is I think we don't have a choice, but we talked about these trade talks that just keep getting kicked down the road, right? We don't have a deal yet. But I think this is an important context and conversation to have in relation to the other one we had earlier in the show and the earnings reports related to AI. Everyone bemoning all the spending we're doing, thinking that this is just a tech trend. This is not just about uh the next big tech trend. It's national security and economic security. And this is a proof point here that you know if you're involved in AI trust no one that's near you and close to you right >> Intel earnings are out. Let's get to see Mod with those. Sema Mike a solid beat for the third quarter for Intel but fourth quarter guidance uh was lower than anticipated. Intel CFO David Zisner telling me that towards the end of August and into early September what Intel saw was a pretty big fall off in demand due to memory pricing. He says memory has just gotten so expensive. It's making these systems, these PCs so expensive that it's just really hard in a constrained budget environment for all these CIOS to be able to buy any more than you know that they can afford. Uh, Intel now sees its fourth quarter revenue between 15 to 16 billion versus the estimate of around 17 billion and earnings of 25 cents a share. The expectation guys was for 43 cents. The stock is responding uh down about 16%. Now on Foundry, Zindner telling me that Intel is in the final touches of what will be customer lockins and that on packaging there is clear progress where he says uh multiple of billions per customer type of business that he sees coming. Uh on capital raises, Intel did have that $23 billion equity sale uh that's been raised. Zinder says Intel does not see a need to raise additional capital. The only deal Intel might entertain is a strategic customer kind of like what they did with Nvidia and Soft Bank last year where Capital comes with demand for either product or wafers. I think the streets focusing on that Q4 guide, guys. Uh the expectation was for a beaten raise. >> Exactly. The commentary about memory prices, too. Very interesting. Amazon earnings are out. Let's get to those. Kate Rooney has those. Kate. >> Hey, Melissa. So, it was a a miss here on the top and bottom line. The AWS growth number though was stronger than expected on EPS. This was a seven cent miss 189 in the quarter and revenue also came up light 201.9 billion street was looking at for closer to 202 billion there AWS growth this is the cloud business for Amazon important number here this came in at 39.3 street expectations it's around 38% so that was better than expected it was also a miss on AWS revenue we've got a couple of the segments I'll run through here revenue of it looks like excuse me it was a beat on AWS revenue 46 billion versus around 45 on North America revenue though slightly light so this is the e-commerce side of the business 112.9 billion just shy of the consensus estimate online sales as well store sales 68.4 4 billion versus around 69 billion that the street was looking for. But you can see the stock reacting sharply to that top and bottom line miss at least. I I think we also guys hang on a second do have some guidance here. It looks like operating income guidance 28.1 billion. This is light street was looking for 26.2 billion operating income. It looks like at the midpoint is pretty much in line. We'll keep digging through here guys, but you can see the reaction in the stock. uh tells tells you the story here. Top and bottom line miss down almost 9% after hours, guys. >> All right, Kate. Thank you, Kate Rooney. In the regular session, we should note that uh Amazon shares were down by about 4%. It definitely absorbed some of that bad news um from Google. >> It did. >> Uh and today we have another decline. >> And I have to say this quarter with these big names, um a stock that's been down into the number has not meant a lower bar. It's actually meant the market getting anxious and sniffing something out because obviously Google have been an underperformer going into its report as well. >> Exactly. Let's go. >> Well, SanDisk reporting earnings just moments ago. That stock moving higher right now. With us now is Medi Husseini, senior equity research analyst covering technology hardware at Susuana Financial Group. He's got a buy rating on the stock with a price target of 27.45. Uh Mie, great to see you. Uh I guess uh you know given I guess the sell the news responses we've seen elsewhere uh in the AI food chain this is okay in terms of uh it holding its value after the uh after the print. Uh what are your main impressions of the of the quarter? >> Yes. First thanks for having me on the program and note this is not still on the news and if you think about the AI from a big picture and as we look into 2027 is actually going to get more exciting. We're m we're we're migrating from training to inferencing and in my opinion inferencing will require more storage and the kind of storage you need is a flashbased SSD and 27 actually could create an opportunity for us to differentiate SanDisk and in that context I believe in a annualized earning power of $250 pre-split. So there's still plenty of upside from here. I know Medie that not all storage is created equal, but in terms of Intel's comments about memory prices and the fall off in demand at the end of August going to September because of high prices and a reluctance uh to make purchases at that point. How do you interpret that onto the memory space? >> Yes, fair question. And uh for disclosure, I don't cover Intel, but I cover Foundaries. I cover TSMC and Samsung Foundry. and to to in my opinion there's still no uh viable uh comp competition to TSM and TSMC continues to be the primary leading edge foundry and perhaps this is where you can use a PC as a cover or or or or higher bomb cost preventing um um um adequate PC sell through but in reality um Intel was supposed to offer a viable foundry solution and I I just don't see any customer working with Intel for wafer loading one or two years from here. >> Now in terms of um I mean we talk about after the uh report that SanDisk up a little bit but it has really uh come down like 40% from its highs. What is it going to take for investors to get a little more convinced that there is durability in this cycle and SanDisk is going to have a spot at the center of it? Two >> two things two factors. First we have to be able to differentiate the semiconductor content and to that extent memory requirement for inferencing which my in my opinion will be different than training for instance HBM may not sustain the growth in 28 but SSDs will continue to grow at the same pace and number two uh that yes not all memories are the same and flash will be the key critical component part of inferencing and perhaps more so than hard disk drive. Let's dig into that huge drop with Bernstein, Stacy Rasgen. Stacy, great to have you with us. >> Good to be here. Thank you. >> Um, what do you make of the drop? Is it warranted? >> Well, probably. Yeah. I mean, look, people have been very excited about CPUs and they've been excited about Intel's more structural changes, looking for growth in in foundry business and packaging and all this stuff. That being said, you know, they they still have a lot of PC exposure and I I suspect that's where the weakness is. The the quarter actually was was fine. The quarter itself was was pretty good. Kind of in line on revenue, was above on margins. Um the guide is extremely weak though. They guideed 15.5 billion. The street was at at 17 billion. Um I personally have been worried about uh uh the the PC client market. We've been of the belief that CPUs have been overshipping for the last several quarters relative to PCs. And if you look at the the PC numbers themselves for Qi, they were absolutely horrendous for PCs probably in at least 20 years. Um, and I suspect that is what is driving the the week. I just there's probably a pretty sizable flush of the client business in into Q4. But I would also say like even I'm just trying to work through the what's implied in the numbers here. And I mean to get there, I've got to have client revenues probably down at least 20% sequentially in Q4. But even with that, data center would probably be up 10%. And I'm not actually convinced that would be enough even if client was was fine. Um and I I think in in the the press release they talked about trying to bring on supply to meet extraordinary demand for data center. They will probably talk about supply constraints. I think it's just taking longer presumably to ramp that stuff into into production. It just it just takes time. I also think people are probably looking for more upside maybe from pricing particularly in servers. They're guiding gross margins to 38% which is like you know 400 basis points below where the street was. Um, it's probably a combination of of process mix and and and lower revenue, but it doesn't seem like if they're getting any pricing, it's enough to help gross margins. So, all in all, given given these results, I'm actually not surprised at the move um uh uh aftermarket right now. >> And Stacy, I mean, worth probably mentioning the context and why perhaps there's been so much downside. The stock was down 35% from the highs because the valuation just got so dislocated really from from the business, right? Right. So, I mean, it went from 20 when the US government bought a stake, $20 a share, up to 140. And even with the decline, it's at 40 times forward, you know? So, it's it's sort of been an odd spot. >> Yeah. I I I mean, to be fair, it it's probably a different company now than it was when the government took the stake and and the balance sheets in a better spot. And the server market, even though maybe people wanted more, but the server market is saving saving them, right? And look, take take lucky, though. We're good like over the last like year or so, like it's it's fine. Um, so they're in a different spot. But, but, but that being said, you're right. I mean, the the valuations have have gone up. It's very expensive on earnings. People who are buying it now are not buying it for like the earnings over the next year. They're buying it for like a structural change over the next like 5 years as they ramp, you know, third party foundry business and everything. And I think people in their heads have much higher earnings power potentially like in a number in a few years than they have right now. But that's still, you know, up in the air. Like it's it's it's still speculative. um the actual hard earnings that we can sort of see right now on on that basis. Yeah. The stock is extremely expensive. It has been >> the care for high prices are high prices, Stacy. Right. That's how the say what the saying goes. So in terms of what their commentary is on memory prices, how do you sort of extrapolate that uh you know across your coverage universe and maybe even into some of the other ones that consume memory? >> Well, this I've been actually very nervous about PCs and smartphones and actually frankly even in our preview for Intel into the quarter, this was the biggest risk. We we were below in in Q4, not not this this far below this this but this is directionally consistent with with what we had thought was going to happen, but it's much worse on a magnitude basis. And the reason PCs are so weak, it's not because PC demand is necessarily it's because there's no memory and you can't build the the PCs. We'll probably see the same thing on on smartphones, frankly, or and and in broader consumer electronics. I've been very nervous about things like PCs and and smartphones. Um and and you look at the the shipments. I mean PCs in Q3 were just look at, you know, the the third party data sources, they were down 20% plus year-over-year, and they were down double digits sequentially. They're usually up, you know, potentially double digits sequentially. So, this was actually the worst sequential quarter for for PC shipments literally as far back almost as I have data, certainly over the last 20 years. Um, and I'm already I already am concerned. I I felt and we we've we've written this and talking of I've been worried that the PC CPUs have been overshipping PCs for the last several quarters anyway. So my guess is this is a combination of a weekend market because of these memory dynamics coupled with probably an inventory flush um that that's hitting it in in into Q4. But the the memory situation may not resolve for a while. And I mean even Intel in their release said they thought that the PC dynamics would last at least through the first half of next year. We'll we'll we'll have to see how it goes. Given this retrenchment you've seen across the group, um Stacy, I mean, what's uh what makes sense in terms of finding relative value here? Obviously, the things like Nvidia has been the defensive quality name and it's held up better. Uh but what do you think is is timely at the moment? >> Yeah, I mean, look, we we we've liked, you know, the Nvidas of the world. They are very cheap and frankly if on on all the AI bullishness, um uh presumably none of it works ultimately. If they don't, they're they're driving it. Um the other one that we we've liked frankly is semicap you know and frankly you know all of these shortages there's no memory I mean they they are adding capacity now to to make more of it but it does take time to add that capacity but I mean it's the tool vendors that are sort of at the forefront of of of that and we had we had some namings report or last couple weeks KLA reported last night for example um and we're starting to see this we've like semicap um as as another way to to to play this >> um we got some more details from Kate Rooney on Amazon Kate >> hey Mel so The call is kicking off. We are getting some comments from CEO Andy Jasse. He's talking about their custom semiconductor business. So, this has been an area he's focused on in terms of making them more efficient, also potentially driving sales outside of Amazon. They've largely built these and use them internally. But he did say with Trrenium demand so strong, customers continue encouraging us to make them available outside of AWS. We intend to do so in late 2027. So the end of next year they will start selling these NVIDIA competitors to outside customers. So that is news. Jasse has talked about this as a potential in the past. He's saying that explicitly. Also saying I mentioned AWS growth north of 40%. He does talk about these being trillion dollar revenue businesses in the next 5 to 10 years. And uh we'll bring you any other highlights Mel, but that's it for now. >> All right Kate, thank you. Um K Rooney, yes, he said confident that AWS will yield more revenue than previously expected. I would think that um capping capex at 220 here at least for for this year uh which was you know what it had forecast before so not raising it not lowering it just keeping exactly where it is but a being able to forecast more than 40% revenue growth for AWS in 2027 and 2028 >> is pretty positive. >> I think it's great. Isn't this what we're really valuing the company on? So so and and I know we wanted it at 42 43% and you know the whisper and all this stuff. It's it's so true because I mean every report I read on this, you know, I'm trying to get a sense of where people thought a AWS was going to be and that is where they thought it was going to be for this quarter. They thought it was going to be in the low 40s. Um but but if you're telling me that especially four or five quarters ago or maybe it was seven quarters because they've now been growing for six straight quarters that AWS um back into the mid40s is a great story especially with an order backlog that he also pointed out is north of 578 billion. um and has grown triple digits year-over-year. So, I I just it's just interesting to me where we choose to push back on Amazon. You know, this is a case where if in fact uh when the retail business was the bigger story uh and people felt the need to point out that AWS was really the more important part of the business. I don't know why we're so focused on retail cyclicality and the K-shaped consumer unless you think this is about um Agentic getting in the way and maybe this these sales are going elsewhere. But if you're worried about the consumer and Amazon, you should be happy today that AWS printed what it did. That's not the reason to sell Amazon. >> That's fair. But if you if you combine what Google said last night with what we heard today in terms of the miss on advertising, the miss on retail, then you maybe put the pieces together and say, you know, there could be continued pressure on this area of the business, which happens to be its core even though it's not the highest growth area of the business. >> But I don't think it's drive what's driving the stock really. It's interesting the, you know, the capex spend has obviously been a gigantic part of earnings for several years now. It's funny. We've gone from a place to, well, last night it was actually a lower slightly lower capex. That's no good. This is an inline capex. That's no good. God forbid higher. That's that's no good either. And um I mean Jasse being optimistic on the business. Um I don't that that gives me comfort. I mean it shouldn't, but it does. >> It's like Goldilocks and none of them support. of them. There's no say, you know, back in April, um, Amazon invested $25 billion in Anthropic, and with that was a 10-year commitment to buy a hundred billion dollars of tranium. And and so this is, I guess, one of the things that folks are kind of like, you know, like questioning at this point. It's a bit of a circular sort of thing. Now, this might prove to be a great investment uh for Amazon and it might be uh, you know, hundred billion dollars of sales in a a nent business outside of what they're doing. So, OpenAI has contracted, you know, for traniums, that sort of thing. we don't actually have a lot of data about how they're working outside the Amazon ecosystem, you know, outside of AWS and the like. So, again, a lot of this stuff is kind of on the come and we just don't know how it's going to shake out, but it also places a strong emphasis on how we keep seeing the circular and all the frenemmy stuff and that sort of thing, but it all points to a lot of competition for Nvidia. >> All right, let's uh get Jean Monster in here. I want to get his take on these managing partner, Deep Water Asset Management. Jean, always good to see you. Let's start off with Amazon. Um, do you think that the stock should be down on the back of these earnings? >> Yeah, it should be down. You know, they had the kind of the prime day shift. They guided in line, but the reality is, and you've covered it, this agentic piece, these personalized bots are going to become an increasingly part of the conversation. They're nent today probably starts impacting their business late in 2027, but at least the narrative has kind of shifted. Uh, so I'm sure that's going to be kind of a key in the call. Interestingly, Jesse didn't want to go there early. Uh, Sundar, if Google did want to talk about these personalized bots, he obviously wanted to talk about AWS, which I do want to emphasize is that if we set aside Amazon stock and look at uh what Amazon said, you were just a minute ago talking about the guidance for AWS, this is probably the most bullish data point in AI over the past three or six months of all the data points that we get. And uh agree Karen that Amazon does not do a good job of guidance. Typically they guide low on operating income and then they usually beat it. Revenue is kind of right on there. But there is some credibility. This is not like Elon type of guidance. And the fact that they went out 2 years I can't remember when a company has given guidance out on a segment for two years. And I think that uh speaks to a conference. Typically they only give one quarter if any if at all. They'll usually say acceleration deceleration. This is an over-the-top endorsement to how they think the cloud business is going to play out. That happens with the backlog that Tim's talking about. It happens with everything that Elon was talking about, what they're seeing with the amount of capacity that they're building out. It hap it plays into what Meta said about aggressively increasing spending next year. This is a very despite these stocks being down and the markets, they'll probably all be down tomorrow. I think that the significance of that guidance uh is going to be lost but it should not be. >> I mean you also pair this with what Jasse has said in recent days. I think it was just on Friday in a podcast he said uh in 2027 capacity is largely booked. Significant portion of 2028 is booked and that is how you get the clarity on the growth rate for 2027 and 2028. So in terms of this decline that we're seeing now, the AI portion of the business is going gang busters, but it is in fact the retail business that's driving the 6% decline. >> Yeah, there's I mean there's two I'm looking at the bigger picture, the read through with Amazon. And just one more on the guidance. They guided to number at least 40% which is higher than what they just printed. And so talk about like acceleration for multiple years. Something bigger is going on is my point. But as far as Amazon as a stock, again, this narrative now, this may end up being House of Cards about, you know, they may end up adopting to what's going on with the bots and embrace them in a certain way, but um again, we've talked about this a lot. Usage today is fractional at best. It's going to be widespread uh in the next couple years, and it changes how people start their commerce experience. If you talk to people who use these personalized bots almost universally they say they have changed how they start their uh buying experience and that impacts Google it impacts Amazon >> how about the tranium chip demand is this something again should we be worried about Nvidia and you know everyone's doing their own thing this isn't a new story but um it does seem like they're seeing all kinds of demand and and uh um they're also less of a buyer so what are your thoughts >> yeah I mean directionally it it has to fall in the negative category if we kind of put what we know a potential competitor that's well funded that's you know we're still a year away so that is negative by by any measure the positive side is what this AWS and the guide was relative to Nvidia because they got to build out infrastructure to support that kind of growth and so I think that uh yes it's negative but I think the the the bigger data point tonight is positive for Nvidia >> hey Jean it's Karen thanks for being on. Um I just want to clarify one thing which I I agree with him on the longer term guidance which is not something they do lightly. It's the short term the next quarter which they usually don't really care about. But um talking about their capex if you were they do have a great balance sheet but they do need to raise money. You were their CFO. How would you advise that they raise it? What instruments? >> Yeah I probably would do equity. I think that uh you know that debt piece is a little bit uncertain about what the the rate is going to be on that. Uh the stock's still in a relatively good position. Um I saw that's probably the way that I would go about it. But I mean this all falls in the category everybody needs to raise more and again if we believe they're competent in uh seeing what the demand is whether it's through equity or debt it still is is well spent because the demand curve is just so strong. I think that's the takeaway is like it's hard for us cuz we've been talking about this for 4 years. Like how do we kind of get our heads around it? I would just say um I would have lost a bet if you told me if Amazon was going to give a 2-year AWS growth rate that was higher than the growth rate they just reported. All right, I hope you're all doing well today and staying calm in this market. Thursday was a mixed day in the market following hyperscaler earnings Wednesday evening. We saw notable red action in many AI hardware stocks as well as some software names. We saw rotation into non-AI parts of the market and treasury yields move lower as some investors reduce their AI exposure in a flight to safety. I covered the highlights from Microsoft and Alphabet earnings in last night's video if you want to check that out. I'll cover some of the highlights from Amazon Intel and SanDisk earnings in a moment, but first let me address some of Thursday's news. The New York Times published a story concerning an alleged Chinese spy who reportedly had a relationship of some kind with Nvidia CEO Jensen Huan. The alleged Chinese operative is a former executive of Singapore-based Mega Speed International. The report claims that US officials identified this person as someone suspected of gathering intelligence for China. According to the report, the president was briefed on the matter but did not take action. The reported concerns included access to AI technology and possible attempts to evade export restrictions. Nvidia disputed the implications of wrongdoing with Jensen telling Bloomberg that the company's dealings with Mega Speed were routine and that Nvidia had investigated whether the company improperly diverted GPUs to China. Jensen said Nvidia had checked Mega Speed's GPU deployments. Importantly, the allegations do not establish wrongdoing by Nvidia or Jensen Huang. Obviously, I don't know if this story is true or not, and so I'm not going to spend much time on it. This story was published one day after Reuters published a story saying that authorities are inquiring about Nvidia's acquisition of Hugging Face, as well as other acquisitions, investments, and whether or not Nvidia pressured customers. That story caused Nvidia stock to trade lower Wednesday afternoon. The timing of these two stories is interesting. One story drops on the day Microsoft and Alphabet report earnings, and the second story drops one day later on the same day that Amazon reports earnings. Amazon, Microsoft, and Alphabet are the three major CSPs fueling a large portion of Nvidia's growth. So, the timing of these stories is interesting, but I'll let you draw your own conclusions about that. Both stories definitely had a negative impact on sentiment. In other news, you may remember earlier this week, I mentioned that Nvidia quietly lowered the HBM4 and NVLink bandwidth for its Ver Rubin platform. And now, WCCFT has obtained an official statement from Nvidia with an Nvidia rep saying, quote, "The specifications published during Vera Rubin's development were preliminary architectural targets. Final specifications on Nvidia's website and the Vera Rubin data sheet reflect what the system sustains at qualified volume production. As I explained in a video published this past Tuesday, lower operating speed doesn't necessarily change the number of bits being sold. And I would say this change is actually neutral to slightly positive from memory makers. I'm not going to get into the details of why in this video, but you can check out my commentary in the video I posted this past Tuesday if you want more details. We also got Samsung earnings Thursday morning in Korea. That was late Wednesday evening for those of us in the US. We already got Samsung's preliminary earnings guidance in early October, so I'm not going to go over all the details. I'll just mention a few important points really quick. 2027 HBM Supply is effectively sold out. Negotiations were completed at Samsung's targeted pricing levels, and major customers are still requesting additional volumes. HBM4 revenue more than tripled quarter-over-arter accounting for about 60% of total HBM revenue. Samsung has signed agreements with 10 major customers and is negotiating with another 10 customers. It aims to have contracts covering 60 to 70% of production capacity by 2030. DRAM ASPs increased by about 20% quarter-over-arter. NAND ASPs increased by a high teens percentage quarter-over-arter. Samsung expects supply constraints through 2028. And Samsung expects demand to exceed supply by an even greater amount in 2027 than in 2026. And now, let's cover some of the highlights from Thursday evening's earnings from Amazon Intel and SanDisk. I'll just rapidfire important details from each company with a focus on what's relevant to AI infrastructure companies. Let's start with Amazon earnings. Amazon reported a miss on both the top and bottom lines with quarterly revenue of $21.9 billion versus $22.2 billion expected and EPS of $189 versus $1.97 expected. Net income was down 67% quarter-over-arter at $20.69 billion. Amazon reported negative free cash flow of $42.4 $4 billion for the quarter, which was down 386% year-over-year. Considering everything that I just mentioned, you can understand why the stock traded lower in response to earnings. Now, similar to what we saw from Alphabet, cloud growth was a bright spot on Amazon's earnings report with AWS revenue of $46 billion, which represents an increase of 39.3% year-over-year. And AWS operating income increased 55% year-over-year to 17.8 billion. AWS operating margin was 38.7% versus 39.4% last quarter. As a reminder, AWS is the largest cloud and with year-over-year revenue growth of 39.3%. This was the fifth consecutive quarter of acceleration in AWS revenue growth. That is impressive. On the earnings call, we learned that AWS backlog grew triple digits year-over-year and is now $578 billion versus $496 billion last quarter. That's a quarter-over-arter increase of roughly 16.5%. Hyperscaler backlogs are very important to consider when asking the question of how long the AI buildout will last. The fact that each of the major CSPs have backlogs worth hundreds of billions of dollars indicates that there is still far more demand than what the hyperscalers can supply. AWS backlog growing $82 billion sequentially is a positive sign. Collectively, as of this earning season, the three major CSPs have a combined backlog exceeding $1.8 trillion. I know the Frontier model companies represent a large portion of those backlogs, but as I've said repeatedly, I expect the Frontier model companies will be able to fund their compute purchases for at least the next 2 to 3 years. And despite all the hysteria from the usual bears, Anthropic's recent S1 filing makes me even more confident in saying that after considering Anthropic's massive cash position, cash burn relative to the amount of capital they've raised and also the very large amount of capital that they're expected to raise in the near future as their revenues are continuing to ramp strong. We can debate what happens in the distant future, but based on what I can see today, I expect the Frontier model companies will be able to fund their compute purchases for at least the next two to three years. Also, as a brief side note, Anthropic spending commitments are spread out over many years with some of them even extending all the way out to the year 2045. It's not like Anthropic has to pay hundreds of billions of dollars today. Not even close. Now, back to Amazon earnings. AWS plans to make Tranium available to external customers in late 2027. As a reminder, there is plenty of room in this market for multiple chip makers to succeed. Additionally, Amazon's Tranium 4 adopts Nvidia's NVLink Fusion, enabling customers to use Tranium 4 in Nvidia systems. This is not zero sum. Amazon maintained its 2026 capex guide of $220 billion and CEO Andy Jasse said that Amazon expects quote to spend meaningfully more in 2027 with a clear line of sight to attractive financial returns. Jasse said quote the opportunity remains extraordinary. We believe that AWS can grow north of 40% in each of 2027 and in 2028 and be a trillion dollar revenue business in the next 5 to 10 years. We're in the very early days of this inflection. Our strategy is working and I expect it will for the foreseeable future. So, while Amazon's quarterly results overall disappointed market participants, AWS results were very strong and leadership's forward-looking commentary bodess well for AI infrastructure companies. Now, let's briefly talk about Intel earnings because this is directly relevant to memory makers. Intel reported Q3 revenue that was essentially in line with consensus and a decent beat on adjusted EPS. But what really caused Intel stock to trade lower in after hours was the company's Q4 outlook. Intel guided Q4 revenue in the range of 15 to$16 billion versus consensus expectations of roughly $17.3 billion at the midpoint of 15.5 billion that represents a miss of greater than 10% and Intel guided fourth quarter adjusted EPS at 25 versus 43 cents expected representing a miss of nearly 42%. A big reason for the miss appears to be the effect of higher memory prices on Intel's PC business. on the earnings call leadership said quote in recent weeks our customers began signaling a PC slowdown based on the memory prices and availability year-to-ate system endpoint prices have increased significantly placing additional pressure on already limited IT budgets the pace of the PC correction has accelerated over the last 2 months as our OEM partners are aggressively managing inventory to match the softer sell through we expect the current environment to persist until memory availability and prices begin to normalize leadership spoke about the memory shortage repeatedly throughout the earnings call and at one point They said quote we are not optimistic about next year in terms of memory. We think that is going to impact the total addressable market. We think TAM will be down 10% on a units basis next year and then we'll have to see in 2028 whether things improve but our expectation is that it does not improve in 2027. So this is bittersweet from memory makers like Micron SK Samsung and so forth. On the one hand Intel is clearly telling us that they expect the memory shortage to persist at least through 2027 which bodess well for the duration of memory makers pricing power. And on the other hand, Intel leadership is telling us to expect some demand destruction in consumer end markets, which is not ideal. As I've said before, while there are similarities between this current moment and memory cycles of the past, there are also key fundamental differences. A major fundamental difference between this current moment and memory cycles of the past is where the demand is coming from. In the past, memory demand was primarily fueled by demand from consumers. But now, the demand is primarily coming from very large hyperscalers, enterprises, frontier model companies, neoclouds, and sovereigns. That is a very different customer base that has the ability to pay much higher prices than consumers do. Additionally, the demand for memory is much greater today because of the effect that memory has on AI. Think about it. In the past, the amount of value added to consumer products via the addition of more memory and the subsequent premium that consumers were willing to pay for more memory was relatively negligible. But now with AI, the more memory you use, the better the models become. Faster response times, longer context windows, more valuable use cases in large enterprises, hyperscalers, and sovereigns. That's a very different scenario. In other words, this time, in contrast to previous memory cycles, the value that memory adds to the underlying technology is substantially greater today than it was in the past. Let's be real, most people don't care how much memory is on their iPhone. But frontier model companies training new frontier models, hyperscalers growing their cloud businesses, and sovereigns focused on defending their sovereignty. Those are customers and use cases that benefit greatly from more and faster memory. In this technology shift, more memory is a vital necessity rather than an optional luxury. There are similarities between this current moment and the cycles of the past, but there are also major fundamental differences. I do think this current moment is going to last longer than the typical memory cycles of the past despite market participants incessant skepticism. And now let's cover some highlights from SanDisk earnings. SanDisk reported revenue of 11 billion versus $10.7 billion expected. Revenue was up 23% quarter-over-arter and up more than 377% year-over-year. Adjusted EPS was $49.51 versus $47.15 expected. Adjusted EPS was up 26.1% quarter-over-arter and it was up more than 3,958% year-over-year. SanDisk reported non-GAAP gross margin of 85.1% during the quarter, which is half a percentage point greater than last quarter. SanDisk returned $5 billion to shareholders during the quarter through share repurchases. SanDisk announced an 8for1 stock split effective November 19th and SanDisk guided fiscal Q2 revenue in the range of 12.2 to 12.6 billion versus 12.29 billion expected. So at the midpoint of 12.4 billion that's a beat of roughly 1% and SanDisk guided Q2 adjusted EPS on a pre-split basis at 55 to $57 versus $5469 expected at the midpoint of $56 per share. That's a beat of 2.4%. Sandis projects data center bit demand will increase approximately 40% in calendar 2027 and data center will represent more than 55% of the nan market in 2027 and on the earnings call leadership said they expect nan prices to increase modestly throughout the year. Overall these were very strong results from SanDisk. Right now is a difficult time in the market as it relates to sentiment. We heard from each of the major hyperscalers who all spoke about demand exceeding supply. They raised or reiterated their large backlogs and they reiterated their spending plans. Alphabet and Amazon both reported results in their core businesses that disappointed market participants which led to their stocks as well as many of the other tech stocks trading lower. At the same time, both companies posted very strong cloud growth and provided commentary about capex going higher in 2027. Microsoft's capex numbers came up short of expectations due to thirdparty capacity delays. So, I would broadly characterize hyperscaler earnings this earning season as being a mixed bag with multiple short-term speed bumps that are unique to each of the companies involved. And yet, while that is the case, the long-term thesis remains intact. All of them told us that capex is going higher. All of them spoke about being supply constrained. And all three of the major CSPs reported backlogs worth hundreds of billions of dollars, indicating that growth would be greater if they simply had more supply. In the short term, we have an important election next week and then a potential anthropic IPO in November. Given that backdrop, we're probably going to experience some choppiness in the market over the next 2 to 3 weeks, even as the long-term thesis remains intact. As I often say, remember to stay calm in this market. Remember to maintain a long-term perspective and do 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 dotcom 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. 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 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 spending 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