저자: VivekArya외CNBC출연자다수 · 외국인
요약 보기
엔비디아가 실적 상회 및 상향 가이던스를 발표했으나 시장 기대치(whisper)에 못 미쳐 주가는 소폭 하락. 애널리스트는 매출 성장률·저평가 근거를 들어 장기 낙관 유지. 블랙웰 과열 이슈는 일부 사례로 수습됐다는 견해. 수요가 공급을 크게 웃돌며 AI 인프라 buildout은 수년간 지속될 것이라는 주장.
다듬은 글
Nvidia reporting earnings at did beat expectations driven by surging demand for Tei chips but as we mentioned earlier also issuing an upbeat forecast for the current quarter despite signs of slowing growth and that's uh what's got people looking at this stock maybe a little differently we'll see V ARA is the um Bank of America Security senior semiconductors analyst what say you sir what do you think what what you think of last night's numbers yeah uh good morning uh I think that uh you know before we look at last night's numbers let's look at what they actually um reported and the track record that they have set so this is not a small company this is one of the largest in fact the largest company and it just put up a quarter where sales grew 94% and earnings grew over 100% And they're guiding to a quarter where sales yes they're slowing down but slowing down to a rate of over 70% and when I look at what they are able to accomplish next year we think they can grow over 50% and I was just looking at the other um you know so-call Max 7 Nvidia will probably grow four or five times that growth rate um next year and is trading at less than one times earning growth whereas uh the average of those companies is trading at over two times earn growth so we think the numbers are are exceptionally uh strong but uh you're right that uh they were not able to uh come close to the so-called whisper expectations I think part of it is they're just at the start of ramping their new Blackwell uh product uh Supply is just starting uh to get going but if I look look at the usual range of Beats they have versus their guidance I think that'll probably get us pretty close to where people were expecting the numbers to come in so I'm not worried I think this is just quarterly noise and this is to be expected in The Growing Pains of whenever they're launching a new product key is what do you think the disappointment is here then sorry can you say that again I said what do you think the disappointment is here I mean there was the disappointment is that yeah the supply chain um you know is scrutinized very closely uh it's followed by lot of uh people and every time the company uh reports uh expectations continue to uh build up naturally for this size and what the stock has uh done uh but there are always constraints uh when it comes to uh not just getting the supply but actually building the products getting them uh to customers and what we have seen from Nvidia over the last two years is that uh they are able to beat their own guidance by two to2 and a half billion dollars consistently you know last time they guided 32 and a half billion and they reported 35 billion so that's why I'm not surprised to see uh them not exactly come to the high end of where the whisper expectations are but if I look at the usual range of Beats they have we will come pretty close I think once they actually uh report the quarter so I think this is more a case of you know expectations in a transitions quarter uh it's nothing to do with the long-term fundamentals of the company and and your sense about this Blackwell chip is it's just going to knock it out of the park or you think it's going to is there anybody even in the vicinity comp competively at this point no I I I think they uh lead the pack for the reason that delivering this product is more than a ship right you have to first of all be able to put 72 of them in a single rack you have to connect it with very high speed Optical and copper interconnects then you have to have a system and software level approach then you need millions of developers who know how to do something with it and then you need to have a whole range of relationships and uh relationships uh not in the not only the supply chain but with relationships with the Enterprises with the cloud companies and I think that is UN unparalleled um the other thing to keep in mind is they are in an industry uh large language models where the technology is changing several times a year right and their ability to Del deliver a new product cycle they have accelerated from two Year to one year even that is not uh sufficient so we think there's another few years of buildout that is left to support this rapidly growing industry this is Maybe a longer term challenge if you believe it is a challenge but there is a debate taking place inside the large language uh model world right now about whether the the models themselves the scaling of the models is happening at the same Pace it used to uh meaning the sort of hockey stick uh angle is slowing down obviously you saw Sam Alman take to Twitter I think last week and say there is no wall meaning there is no slowdown um but what is your take on what What's Happening Here sure I think um you know when we uh look at the technology industry right we used to have Mor law and and I've been following this industry for a while and every year people have been pronouncing the death of Mor law and and you know here we are almost you know 50 years later and one version or another uh just keeps on uh going um when we look at the field of large language models it started last year um right so we are just kind of in the first one or two years but I think it's um it's it's expected that the industry is going to take a step back um in fact we uh saw a Microsoft CEO just address this issue uh two days ago he said that he's not seeing any kind of uh slowdown in in scaling and it's not just one company driving it you know this is a global industry so you have many different companies driving it in the range of in the direction of training in the direction of uh inference um so we do think that you know it's it's expected for the industry to you know pause and reflect from uh time to time and instead of changing the nature of these models three or four times a year maybe they change it one or two times a year right but the point is that the amount of computing infrastructure that is required just to address what is already there is going to take a few years to go and uh deploy so not wored what do you think the fair for those folks out there who either own the stock want to own the stock or thinking about what to do with the stock what do you think the fair value of the stock is yeah you know uh technology stocks tend to trade between one to two times their earnings growth rate uh we have their earnings growing over over 50% next year and the stock is trading over 30 times um next year so we still think that it's very compelling valuation even though it has run a lot so it really depends on on earning growth and as I mentioned when I look at the other so-called mag 7 in the industry um you know uh they are trading at over two times earning growth so using that specific valuation metric we still think Nvidia is very compellingly valued Nvidia reported third quarter results that beat expectations for sales and earnings and the AI chipmaker delivered a better than expected forecast uh earlier the stock had been down a little bit down by about 2 and a half% right now it's only down by about 8/10 of a percent joining us right now to talk about it is Ray won constellation research founder chairman and principal analyst hi Ray how are you hey good morning so did the decline in the stock price make sense to you is this just a case where Nvidia can beat expectations but they're going to it's almost impossible to beat the the highest whisper numbers on the street ever there are just such high expectations built into the stock every time this is not a big selloff it's down by less than 1% Becky you're absolutely right I think people were taking money off the table it was more than a mic drop moment it was a spike the ball in the end zone performance but the problem was everyone's saying hey what's going on right should we take some money off the table and come back and I think that's really what happened I mean these were some big numbers right 19.83% five billion in guidance that's huge yeah it's hard to find um many flaws or faults at all with the numbers that came through it does remind me of last quarter when they beat expectations again the stock was down immediately after the earnings but then it went on to grow something like 17% over the course of the last three months since then what you've still got a buy on this you're you're looking at I think a $180 price Target the stocks at 144 today I am and here's why when we look at the mag 7 they're basically growing exponentially versus the rest of the Fortune 500 so stocks are basically growing really quickly and Nvidia is basically selling to six of the seven of the mag 7 and it's a huge number it's about 50 billion a quarter in which these mag 7 are spending in terms of data center growth and so that's really what's driving the majority of nvidia's Revenue in the data center side so that growth from 3.8 billion to 30.8 billion in two years is huge right gaming doesn't even come close gaming was up at three billion and so what we're seeing right now still a massive build out on data centers that's going to happen for the next 12 to 18 months and that's what's driving it and then after that we're going to see something different which is in The Sovereign AI countries are going to say hey our companies can't catch up and we've got to invest at a country level yeah I I've heard people knock Nvidia because it's really just these these big companies like the metas and the Amazon that are buying all the chips right now but that's because they're first in line nobody else can get through to them um when you hear about the demand um that Nvidia will discuss Jensen Wang will discuss on the conference call call um it's hard to look at the supply versus demand as a as a problem it just means far more people want these chips than they can possibly produce at this point what if you were looking for a hole to poke because most the holes that I've heard people try and poking this argument don't make sense to me what's the best argument you can think of for why you might be concerned about the stock the only reason I'd be concerned about the stock is living in the valley what we're seeing is a ton of people what Nvidia retire they put in their five years they put in their 10 years they're leaving they're cashing out they're doing something with their lives and it's really the talent that's going to be hard to beat given that stock run up uh the other piece is really blackw to Reuben when it happens in 2026 the question is you know same questions people are going to ask you know are the chips going to be hot are they going to work uh I think that's where people are worried about but right now these chips are two and a half times faster and of course they cost 2x as much and people are willing to make that bet there were those concerns in the last week or so about the chips overheating the black oil chips overheating when put in some of these server arrays I I've read other reports that people are speculating it's already been corrected what do you think um from our conversations with people over there it has been corrected and I think the other piece that's happening is this is really complicated uh not a lot of companies have ever done this before and I think they've tried all different possibilities and combinations and I think they've figured out a couple Solutions and that's what we heard uh and and this shift is actually important because if you think about what is next right it's going to get even smaller and it's going to get even hotter uh and as you know the consumption and the power consumption on these chips hopefully come down but people keep stacking more chips in one place which is why they get really hot in data centers Ray I had not heard the the argument before about people retiring at Nvidia obviously if you've been there five or 10 years and you've got stock options or restricted shares you're sitting pretty you've made a whole lot of money in it it is pretty easy to walk out the door where does Nvidia get new Talent oh they're not having trouble finding new Talent uh I have a son that's a comsi Ecom major and as he tells me every comsi major today actually is looking for a job because AI is pretty much obliterated so mvidia is getting a lot of great talent from different companies from Young startups young startups uh who are actually folding in to you know experts in the Enterprise that have been around for a while so they don't have a hard time fighting talent but the existing talent pool is actually uh some of are starting to take a break even if there are problems with the black oil chips if there were issues around it my first thought was where do they go right now there's nowhere but who is um in pull position even if they're far behind who's in the next in the second place who might be able to come in and and have a different offering at some point down the road we all hope that AMD has something to offer but the challeng is the gap is so huge right they put so much money into development of this set of chips uh and the Capital required to get to the next set of chips and if you continue to go down that path very few companies are going to be able to pull that off and I think that's really the challenge uh but when people say hey what do you do after Nvidia the question is and we'll take a quote from uh the Wayne gresky Playbook Let's Skate to where the puck will be next and the question really is what are we going to do when we go from semis to Hardware to hyperscalers to software where does that move to the next piece and once again Nvidia has an answer they're in the space their Cuda layer actually allows them to build the software on top the chips and create that access point for the next set of ecosystems and so the layering that's going on in Nvidia is what really makes it a very tough company to beat because they've been thinking through all the possibilities and Jensen has even more to offer going forward let's get right to Nvidia moving higher after beating on both the top and the bottom line also nearly doubling sales year onye yesterday on the earnings call Jensen Wong did reiterate the strong demand for AI the age of AI is in full steam generative AI is not just a new software capability but a new industry with AI factories manufacturing digital intelligence a new Industrial Revolution that can be create that can create a multi-trillion dollar AI industry demand for Hopper and anticipation for Blackwell which is now in full production are incredible Jim everyone wants talk Blackwell but even some of the hopper guidance was surprising good yeah look I what you this is one of those situations where whatever you can get from them you'll take uh that that can be very Stark because you may not be able to get the first rank stuff uh I think that there are multiple customers I want to step back for this in a second I've never seen anything like this there are multiple customers who would buy everything they have when I say everything they have I don't mean like hey you know what I really want some I mean like I will take everything you have and the customer that we're talking about are customers that all we all know and they are not doing it in order to be able to say cuz the other guy may have it there's going to be a winner take all loser take none aspect but the main thing is is that everybody has something that they feel they can make 5 to one on I'll pay you 40,000 I'll get uh I'll make uh 200,000 and David when I say I've never seen anything like it what I mean is is that this is a different order of company that that we've ever seen in our years together cuz it's really Reinventing no it's not it's creating a whole new industry and in order to understand it you have to do things that you've never done before like for instance I put in J Jensen says something that's rather made I don't mean to go on too long but Jensen's talking about something that he like likes a Google product that he just you know he's just constantly on I mean good example same thing that Ben off said the other day I talked to Mark about it absolutely and what you do is like you input you know tell me what's so great about this conference call and you know here it is I mean I've got it tell this is it okay well this is I just all we asked this is Jeff marks we put in give us the highlights this is the product that that chansen said he loves to talk to oh you use the Google product to go over the conference he was using The Living Daylights out of it because he is thank you Carl he is because I used the Living Daylights out of it so I did it and I got five pages of the just the quick distillation if someone said pick your best five I I don't mean to ask too authoritatively on this I will say this that when the stock was at 14041 this morning at 330 I was going to town whatever they call it now just saying listen what do you clowns it's just short selling clowns with 100 200 shares because this thing is not anything I've ever seen um and of course the what you were using from Google is generating what it does on their yeah it's like kind of a a hle of mirrors but you when I SP spoke to the company for about a half hour and what they're talking about first of all the story that came out this weekend about burning too hot you know they they use the words some words that I was surprised they said it's a lie now that's a hard that's where you go to jail when when you lie about financials journalists whatever it's a lie I would not want to commit a lie against this company there was no heat issue you don't believe that cuz they deny it so you believe them and you don't I mean I I oh I believe even you said there are probably people within the company talking to the information yeah well that's what they do but if you if you go to to a reseller a a rapper a packager of which they're three right super micro and Dell and HP they did not report any overheating so it's entirely possible that you bought it that you bought something you got something and you souped it up and you overheated it I see but speak David it's a serious charge to say that story was a lie and they couldn't do anything about it cuz they were in quiet I do not want to be affiliated with a lie and anyone else who's affiliated with it I think is suspect yeah no you never feel good as a journalist when a company calls you out in that way for sure and something you want to avoid let's get back to the actual quarter itself and what it means for the future 375 billion in revenues in terms of guidance for fiscal uh fourth quarter of 25 um I mean let's just talk a little math here obviously yes we've never seen anything like it what do they do 94 billion in they're on a run rate of over 100 billion a lot of Reven um 94% excuse me uh increase first first time in five quarters so they didn't have more than a triple digit increase I disappointing there were there were clowns who were actually like clowns I know mean that I mean chhe heads but your point is simply the demand for Blackwell is undiminished it will take this product throughout its ramp there will be endless demand for yes and then every year they're going to get back on the every year of Reuben which is much more powerful than this but Ruben Ruben's next yeah from Philly Ruben got it um but I I when I say there was a lot of stories last night I just want to dispel I want to tell people why that stock got to 141 because there you there's a lot of people at home thinking how come it's a 148 well the story at 141 you know the stories that were run at 3:00 a.m.
4:00 a.m. 5 a.m. were how they missed guidance but there was no guidance so how did they miss it I think I mean to the degree there was hair on the quarter people are looking at networking down 15 that was a glitch it was a timing glitch it be right back up and righteous will tell you that at 10:05 today Ben asked a question about longer term guidance I love Ben that was a question not as well received by KCT Crest as I thought it being KCT Crest is the CFO who by by way David may I just say in terms of being intimidated by people yes I always feel like I'm back from Mrs tof's class in fourth grade when she gave me a b and my father told me I was a disgrace changed the whole course of my life explains a lot this is Tain a lot you've been trying to prove you've been trying to prove her wrong ever since yeah exactly you've almost done it Jim you're almost there no no my father still wouldn't approve even wherever the hell he is 10 years ago he died he's still with me he's still thinking in my mind was like Jim you didn't know how to speak to collect Crest call her back and tell her apologize that you weren't up to speed all right back to back to back to the CFO n sorry need I digress meanwhile $300 an hour for the psych for the psychoanalysis what do I get for it ends at 10:05 okay with you pay meanwhile some interesting pieces on the tape Jim about Jensen's World Tour we saw a piece of it in Japan and the notion that he is looking to dig into that Sovereign story and rely Less on the big four hypers scale here you know what the Dead Poet Society whatever that thing with Taylor Swift did that supposed to help Target this was much better now what's most interesting this is just a predilection that Jensen has cuz he doesn't want jingoism he doesn't want xenophobia he wants people to be be able to say okay listen we have our own in our own language uh the problem is is that how do you then therefore divvy up when Mark Zuckerberg says I need meta AI the best what do you say to Andy chassis CEO of Amazon yes when he says listen I won I got a I spent 8 bill on this Alexa and she's not conversational and you say listen FR you give me a for Blackwell when you say like have you ever tried to like say I want gerswin Symphonies on okay I like your gersman Symphony or American Paris and then it says it has to charge you for the movie American Paris that's over it will know if you want to ask for Beth's piano concerto and it starts playing the sonatas with the AL this example A lot it drives okay well look let's go with the mode Let's go with Mode start with the the Jupiter it doesn't know Jupiter for Mars the damn thing but now it'll ask you conversation and by the way when you come down I'll say Jim you have an 8:00 and then you have a 9:00 don't forget your psycho analysis with David uh and you can what's the matter and but it's conversational it's back and forth and that's what Jensen's talking about that's why one of the things we talk about anyone one we don't have conversational back and forth right now with Alexa she's disgusting I wouldn't I that's kind of rude okay stupid all right that might personal I get too personal but that's what happens that will change and it's a whole new industry because the industry now is B based on CPUs and coding now you can just talk to it one of the Great Moments was they have things that you know they're going to make it so that lawyers not clear we need them not clear we need them you don't need the lawyers well and not you don't need anybody between the first year and a third year associate those people cost 200,000 yeah but you're going to need them when they get older so you got to kind of keep them to go through the first second and third years oh okay so I'll let you go with that um somewhat a new industry of AI factories manufacturing digital intelligence Industrial Revolution these are things that are not constrained by a price earnings multiple I feel like an idiot price earnings multiple what do you like I'm small TV got small I'm here there's digital marketing intelligence like Riders how about hard for digital artist intelligence how about Runway I mean these things are like these are like these are but he says they're not disruptors they're addtive but when I read these if people in professions read this one paragraph all right but nobody has any idea what you're really trying to what you're telling them you're just reading bizarre names Harvey oh Harvey it's like to play with Jimmy Stewart you're not there no I'm not on you with that one you Harvey come on you know Harvey Jimmy Stewart no no no it's it's older it's older than we are then let's turn to you on Nvidia you write about uh constraints today and how important that that term is for NVIDIA yeah what's great about constraints is it means More's coming later and in nvidia's case they're the only game in town you can't you can't just plug in an AMD chip because the software won't work etc so you got to just wait until they can ship it to you so we like that that means it's the stocks buying getting bought on the dips you can see it it didn't stay down very long um it's probably Lish or slightly down now but the reason is is that we know more is coming later that drives sequential growth and margin Improvement as we go into the back half of the year and that's ahead of us and remember they're the only semis company not talking down the one Q that's a big differentiator in this market so I think a lot of people are going to be buying it on the dip what did you make of his push back on scaling uh walls so to speak well we wrote about that too earlier in the week and we think that that is just something we'll work through I was really impressed there there's a with Dario amade of anthropic and uh he kind of explains the synthetic data Trend and how machines are going to create data and then create data from scratch and that that will help feed these machines and keep them going usually in my experience and I'm no PhD and AI uh you know chip uh you know scaling but uh from what I can see the industry will work their way out of this you're not concerned then I because I I want to make sure our viewers understand it's this concern we raised a couple of weeks ago was sort of at the Forefront of it the idea that you're not getting anywhere near the the increases in terms of productivity from uh from the latest models versus the last Model you just don't you're not a buyer of that in terms of believing that this is not that there's a wall here that's being hit I think you have the smartest people in the world working on this problem and they're going to find their way through it and they're finding their way through things there there was other bottlenecks a few months ago a few months ago these are really smart f folks that are working their way through things but I think the positive surprise here is data is the new is the new gold is the new oil it as and it's really coming into its own we are creating data with AI we're creating synthetic data we're creating data from scratch and these guys are figuring out how to feed that into models also uh you know you have Elon Musk taking more and more video uh revolutionizing self-driving I think they've made a lot of advances through video and Jim talked lot about that but that's really just taken hold I I think we'll find our way through this and the data center is the new manufacturing Hub so to speak for AI as as Jensen said during the conference call let's get let's get back to Nvidia just from a your projections I mean what are what are your anticipations in terms of next year's revenues the growth rate and therefore back into a a multiple for me as to why you have a where your price Target is next year I have revenues growing 50% and I think it's conservative and then the next year I'm at 22% I who knows if it that 22 sounds pretty light considering Jensen's track record our Target is only 35 times our 27 estimate which is really calendar 26 so it's a PE to growth that's still very appetizing right now for this stock of course you know on any given day the stock can go down but I think given the growth that's out there for this company and given it's a systems company sort of like apple is for the consumer that's what this is for the Enterprise it's going to be bigger than people think still with more profits as a percent of the total industry than people think still and last longer just you know from my experience the vertically integrated companies that do it all that tend to hit it right tend to go on longer than people think I I'm interested when you're talking about your Revenue projections at 50% and then at 22% and how Nvidia executives are sort of walking this walk carefully talking about growth but you're probably not going to see 50% forever and still keep people excited I mean can they can they walk this carefully well what what needs to happen at some point here is this is a cash gusher too and Nvidia wants to go we're a growth company but at some point they have to buy back more stock because there's simply nothing they can do with it um the government probably can't let them buy anybody even close to you know you know uh a trillion of course so they're going to have to do tuck in Acquisitions and they're going to have to give stock back I think they are going to go through just like apple did that metamorphosis of more mature hey we have more recurring Revenue that's why they're building a software business that's why they're going to continue to do that and I think that they realize that but right now they're in growth mode and it's go go go but I think you know let's put some trust in Jensen and Colette that they know they have to build up those recurring revenue streams and return more cash to shareholders when appropriate and those days are coming and we'll see how they they do but I have confidence that this is a systems company which a little different than just a chip company who's running and gunning um like some of their peers you you even wrote you compared it to iPhone 6 era for Apple which I'm trying to think is that about the time Services got taken seriously over there well you know iPhone 6 was this huge uh big screen transition and then Luca mry uh yeah then the CFO really did a great job of saying hey guys this base we just did is we just built is going to have all this recurring Revenue TR you know here's it is more disclosures around it more now they're not ready yet in Nvidia but they said last night that 2 billion in software Revenue as we exit the year within a year or two that number is going to be way more significant way more tracked but also there's djx Cloud there's other things they're driving they're driving Nims they're they're enabling the market and um these kind of things are going to get more important as we go throughout and uh and I think a lot of companies especially Enterprises are not going to be able to live without these products as they adopt AI now joining us is Morgan Stanley analyst Joseph Moore uh Joseph obviously a lot of discussion so far on Nvidia it's been the talk of the week not just the morning what's your sort of overall initial read on the quarter does it make sense to you as well that you're seeing the stock PR price action lower here after that quarter posted and the expectations going forward yeah I think so I mean the narrative here is going to be the ramp of Blackwell right and the used these these descriptions of Blackwell of you know staggering demand insane levels of demand um at the same time your Revenue in the next 6 months is still driven by majority Hopper which is the older product and you know so I think you're going to have less upside for a couple of quarters here while we move into that Blackwell ramp and that's how we previewed it uh we didn't think there would be uh as much upside maybe as some as thought and and there wasn't but I mean it's still a very good quarter you beat Revenue guidance by $2.5 billion no Semiconductor Company other than Nvidia has ever done that and I think they are looking at a very strong Blackwell cycle into the back half but it is a transitional period as we sort of move from the the later days of Hopper into the ne the the beginning of the blackw generation there was so much to like in this quarter but nothing is without risk what would be potential concerns for you as you're looking ahead as black well skills yeah well they've talked about lower gross margin and they did guide gross margin in the first half of next year below our expectations I think that's largely a function of the complexity of Blackwell the the number of different skews that they're going to have out there the systems innovations that are required there uh and and wrestling with some of the supply constraints around that uh but they did mitigate that concern by talking about much higher gross margins in the back half of the year so actually if you look at our margins for the full year they actually stayed the same uh but we lowered the first half raised the second half and we exit next year at a healthier rate so I think that's one of the elements um and again I think you know when you have these transitional windows we are still talking about a very large number of Hoppers right they'll do north of $30 billion of Hopper Revenue this quarter which is a product that has already been effectively osbor by the Next Generation product so I think there's risk around all of that um but I still feel pretty good about it and I think we're we're looking at the the strength of Blackwell and and these kind of scaling concerns the debates that are out there I think Nvidia did a very good job of addressing that and talking about why this will continue to scale and grow into the back half of the year so you know some risk for sure but I think uh you know 3.6 trillion market cap for those of us who have covered the stock since the IPO is a surprising number but you know they're they're executing really well and the narrative is really strong and the uh the the sequential decline in networking got a little bit of a look uh last night and this morning um but then got ridden off his noise is it something well they have they talked about most of the core components of networking being up so infin ban was up ethernet was up Bluefield was up um so my sense is it is at the noise level I think there's some of the optical component business you know as there's a mix shift towards ethernet they probably have a little bit less of an optical sell through so that's a lower margin part of the business anyway so the core elements of networking continue to be quite strong and I think that um they talked about networking continuing to be strong into next quarter so I think you know there's mostly noise on that but it certainly is something that got people's attention finally the other thing that's getting framed a bit today Joseph is um I mean Jensen's been talking about Sovereign demand for a while now but this recent trip to Japan and this notion that I he trying to diversify away from the big four hyperscalers is he trying to graduate the perception of that pie what do you think yeah I think that there's there are a lot of sources of demand here right I think when you look at inference being close to 40% of Revenue probably over 40% revenue and when you look at Sovereign demand as you said they've talked about that as a low teens percentage of Revenue I think some of the Enterprise applications are sort of in a similar mid te percentage of Revenue um you know a lot of the focus goes to these Mega clusters and the the sort of race to have the biggest cluster and people spending tens of billions of dollars and that probably is the element that as you look out to 2026 has the most uncertainty but there are a lot of businesses within this that do have quite solid growth potential including Sovereign including Enterprise including uh some of the the inference parts of the business that you know I we we still have business plateauing somewhat in 2026 but still very good growth prospects you know even as we get past this period of the blackw product cycle quite an incredible company incredible run on the stock Joseph you um upgraded of course that Target price here today Joseph Moore thank you very much for being with us following nvidia's quarter Coast let's talk about Nvidia okay um your your read on the the quarter which was you know obvious ly a beat sales up 94% year on-year um that is a Slowdown from the prior three quarters which have just been astounding and 94% up year-over-year is amazing obviously but what's your take I mean let's let's say at the start it's down a little bit today because expectations were Skyhigh but this was a blockbuster quarter by Nvidia right what's clear to me now and what I think the biggest take away from the call is that the world runs on Nvidia as Jensen said AI is in full swing he said this is a multi-year event this is not just about training AI this is about moving all the world's x86 architecture all the applications that run on CPUs are going to be rebuilt over the course of the next three or four years on gpus that are running machine learning so I think that the the market is going to have to digest that this just isn't about training clusters in 2025 this is about GPU data centers that are going to be built in 26 in 27 and so for us our numbers came up you saw that I think 10 uh uh sell-side analysts took up their price targets for the stock we were buying it pre-market when it was down buying it this morning yeah so I you know I think it was a blockbuster report and I think you know it's clear to me that the intels of the world are going to continue to lose share and the nvidias of the world are going to continue to gain share and if you want to bet on the Tailwinds associated with AI you have to have Nvidia in your portfolio Okay so it's now you your number two position correct with buying a little bit more this morning Josh what was your take uh on the quarter you obviously well known uh for being in this stock for a very long period of time I think they were very Artful in the way they maintained everyone's uh enthusiasm for the potential for Blackwell where orders stand the fact that demand is still out outstripping Supply but they didn't ratchet up the excitement too much to put them in to a position where they just can't win the next time they report and that's really going to be the company's job going forward it is already the largest market cap company in the world it is already the most must-own stock in the world there's no need to continue to put pressure on themselves to outperform and outperform and outperform so it's harder than it sounds because they do have a really bright future and the revolution is real and the the dominance of gpus and data centers going forward is undeniable and their market share position within that Revolution is also undeniable so they have to now do this quarter after quarter give people realistic guidance execute ship products hit those numbers but not paint themselves into a corner that they can't escape from where the market expects them to double Revenue every year which of course like mathematically will be impossible so far so good I think Jensen understands the mission I think he's accepted the mission and I liked everything I heard last night the most important thing and I'd love to hear Brad react to this is this idea of the moat because competitors are coming in with new Solutions Amazon is ramping up production of their own uh uh trainum 2 chips like they will not have the whole markets themselves forever but they can maintain dominance and it might actually be a good thing for the ecosystem for there to be more products out there I think that's the the the tone that was set last night uh and I think other Nvidia Longs heard the same thing that I heard what do you think I mean that's all about the idea of the installed base and we've heard Jenson talk about that before almost in the same way and I think I've mentioned this in the past of thinking about that in the same regard as we do apple with its own sort of powerful installed base yeah I mean listen the stock started the year 4 9 bucks a share or 5050 bucks a share today we're at one you know 145 or something the asymmetry is closed and part of the reason for that is everybody had this wall of worry right custom as6 Amazon's going to have their own tranium chip I said on our recent bg2 podcast of course they're going to use teranium it's a cheaper price solution for certain customers but they also you know jasse has said he's going to buy as much Nvidia as he can get his hands on for as far as the eye can see they they know the world runs on Nvidia so um you know listen when we look at our price target for NVIDIA next year it's up 30% not up 100% like it was at the start of 2023 or at the started 2024 so when you ask me as a portfolio manager you know we took down the size of our position it was a really outlier position for us to start the year it was a normal siiz position for us coming into this quarter and we're adding a little on the pullback in this quarter but we think it's going to be a great compounder and a great bet on not only AI but also on the replacement the move from x86 to GPU based data centers for the next several years okay Bren what about you you've been a holder of this stock for a long time too yeah I mean I think what's so amazing and you talked about this moat and what this means for the amds of the world Etc is Jensen walked through how 64 Blackwell gpus can be can run gbt 3 compared to 256 h20s so they're even making chips the the Blackwell which not a chip they're making these gpus that are cannibalizing their own products and this is why they are so hard to compete with is they just continue to innovate so quickly and then when he talks about how Oracle is is announced this Zeta scale whatever that is AI Cloud cluster that can that can handle over 130,000 you know Blackwell gpus and so I think when When Brad talks about this shift going from modernizing from CPUs to gpus I don't think we are really great about understanding Jensen thinks it's trillion dollars over the next 10 years I mean the difference between a billion and a trillion a billion seconds is 32 years a trillion seconds is 32,000 years I think we're still early in this and this moat is going to last for a very long time but I do think they're going to be big Winners and big losers and right now Nvidia is the winner of the bunch and Liz I mean obviously the market still trying to digest it um if there was any bit of I don't even want to use the word disappointment I mean there's just so much to live up to as we said we had 11 or 12 price Target increases on the name I only see one firm talking about their guidance being below what they would suggest is the bull case I mean as you watch the market here react to it what do you think well I think the most important points that have been made so far are that expectations were so high Brad said that at the beginning of the show and then Josh about making sure that they can set realistic expectations going forward Ward this Market is hard to impress on stock price action I mean we're talking about situations where you've got beats and raises being punished in the stock price so first of all I think we're still in the infancy of this AI theme we're going to be in the infancy of it for a little while yet but as it matures what we're going to find out is whether or not the companies that have been the winners so far can continue to win along that vein and then who are the other ones that will come in and even the other industry groups that will come in and something that's been really interesting to me in the last month or so is just the performance of software and I think some of that has to do with the fact that okay we've established that this thesis will continue going forward with AI what else can companies start to spend on where can they actually find the profit and who are the companies that are going to help them create that profit and I think software is a big group that's going to benefit from that going forward let's start with Nvidia which is lower today despite delivering another solid quarter with Revenue nearly doubling year-over-year the guidance was ahead of estimates but the bar was very high and it could be somewhat of a sticking point for investors but come on are we going to equival about a 1% decline the stocks tripled this year uh my next guest kind of agrees uh he still has a buy rating on the stock and while he acknowledges some Supply constraints he expects them to be short-lived with Blackwell ramps accelerating in the first half of next year shy paduri is senior semi Analyst at Raymond James he's got a 170 price Target on the stock why do you think we're there's some selling off here in shini is it even significant given that last time around I think the shares would down 6 7% on the results and then they still rallied sharply into this quarter yeah Kelly we don't think it's all that significant first of all the demand itself is very strong right now one indication actually one big positive surprise in the quarter was that if you look at the hoer demand you know the current product that they're shipping in Q3 and also into Q4 you know that grew five billion sequentially that's the fastest growth in the last four quarters in in the last four quarters Hopper growth has averaged about 3 and a half billion and and in the in the in 3Q it was 5 billion and usually what happens is that when there's a new product coming out there's always this concern about a Slowdown in the older products but we're not seeing that so that tells me that the underlying demand is very strong and as you said you know Supply is the main issue for the inline is Outlook and that's something we were expecting so you know we think the demand is going to continue to outpa supply for almost in all of next year so if anything you know as long as cycle extends we actually view that as as as more of a positive so we're not worried about you know a small pullback here shini can we kind of kind of back out for a second and put this in context and maybe you can use you know what your direct experience throughout your career as a as a kind of the witness point it's so unusual for a Semiconductor Company to be worth three and a.
half trillion dollars and are we you know the valuation is not that rich they they continue to kind of match where they need to be with the market so it all seems real but is this sustainable as a a part of the US economy is it sustainable in terms of the semic cycle I mean I I would just love your thoughts on the significance of all of this yeah you know when you look at the market cap in in in a vacuum it does look a little scary and I'll be honest with you but at the same time you know you look at the company's profitability and and the valuation it doesn't look that bad you know historically this is a stock that traded at about 40 times on average right now it's in the low3s so the question is you know how long this you know AI related spending will continue uh just listening to their customers it doesn't look like you know we we we you know we need to worry about a Slowdown at least in uh you know for for the next you know 12 months uh of course you know there's going to be a pause at some point um you know we don't have visibility as to when that pause is going to be and the other question we get is that you know what's the ROI like on on AI why are we spending all this exactly you know amount right so I think I think you know the better way to look at it is that right now the competition is so intense among hyperscalers that's what's driving it it's not it's not so much about you know Roi I I think the reason is that in Tech you know there's always one winner right if you look at any of the you know previous you know Tech Evolutions there's always one winner so I think you know the competition is going to remain very intense you know before we know who the winners are going to be so that's not going to slow down so I think it's a little too early to think about Roi from that standpoint that's really interesting so basically because Google is battling Amazon is battling meta I mean because they're all still fighting to be number one it's benefiting Nvidia that that definitely makes sense and then of course there's the other thing Nvidia right everyone hated Nvidia think about the reasons people say don't buy this Market there were the reasons for the last 10 years not to buy Nvidia uh you know it faced all kinds of scorn from the financial media from Wall Street it was this it was that well today the financial Financial Wall Street has no choice they're all over this stock I want to bring in uh right now one of my favorite guests when we start talking about this from spar investment founder their CIO Ivana Dela Iana stop in here for a second look at this these are all the targets that was raised maybe more by now right so mizuo uh Goldman Bernstein City Deutsche Bank Morgan Stanley eore Barclays web buus truis TD cow and Barrett then you had a few reiterations right strong buy only thing I would say is kind of interesting to me the targets aren't that high not necessarily the highest we see is 190 from ever cor are you surprised that Wall Street has I mean they're saying buy the stock but they still seem to be playing it close to the vest well it's not necessarily surprising because that's how the sell side Works they really look at one quarter ahead and they model it if you want to be uh looking at the situation correctly you really need to take a step back and look at it three to five years out and model the next Generations of chips so the blackw is really the upside that we're expecting to see going into next year and then there is future generations of chips that will come after that so the key here is whether scaling laws will hold which will mean that people will have to invest and whether they're going to be able to execute on this next generation of blackw chips so let's let's look at though so here's nvidia's quarterly earnings going back to the fourth quarter of of 23 I mean just the mindboggling stuff data centers 3.6 billion to almost 31 billion it was a it was known as a gaming stock for a while remember in the the only the video game stuff uh that made a big move in fact every one of these categories data centers gaming professional visil uh uh visualization Autos all came in better than Wall Street consensus what's going on I mean they're they're executing in every facet of their business so really what it came down is the guide they guided to 37 a half billion for the next quarter which as you can see here is not as big of a jump right and I think what people misunderstand is this next generation is going to be very different than the hopper in that it requires actual physical infrastructure buildout so while the hopper was going into existing data centers right now you're going to need whole new data centers with all kinds of power requirements networking requirements so that slows it up a little bit here well that makes it a little more volatile but it's also going to make the upside larger so this is why you're seeing networking stocks power gen stocks really through the roof today is really because people are realizing hey there's going to be significant it's got to happen you like what you heard on Blackwell absolutely we've done Channel checks on Blackwell the demand as Jensen was saying say is truly insane every basically the prior generation Hopper people were ordering in the 10 20 30, thousands here people are ordering in the hundreds of thousands so here's a fun thing for you AI reviews nvidia's earnings results I love it right so AI says the negative Supply constraints uh gross margin pressure competition in China and maret Market digestion are any of these long-term issues for you well Supply constraints are short-term issue that we think are going to be resolved gross margin pressure I think people don't realize here that when you launch a new product you are trying to sort out the value chain and the supply chain and that's what's causing uh gross margin pressure as Blackwell ramps in the second half of this years gross margins will be up right uh competition from China that's not really a concern because we haven't really seen anything legitimate and Market digestion Jensen said that he expects that when we reach 1 trillion by the way Jensen's going to be on with Liz Clayman right after the show uh the AI data center value chain U this I love this it's a great uh piece uh is there anything on here we just we only have 30 seconds but is there anything else on here that we should be looking at where the next opportunity is so networking is really the most immediate opportunity that we're really focused on over in the networking area yep that's exactly right so this is basically like transceivers well this stock arista's up like 10 points today um um Marvel does these are some interesting names in that's right and then also the power and termal management side these are a little more boring companies so you're not going to see 10% B has been on fire though I mean somebody G berno I got phone calls on that one today real quick we as well we disclose our Holdings on our website by the way for people that want what's your website again spear invest.com spear invest.com you also like Nvidia snow and Tesla yep you've been rocking Godly give me hot five you've been rocking are you kidding me thanks a lot Ian all right joining us now of Fox Business exclusive is NVIDIA founder and CEO Jensen Wong Jensen great to have you on the show thanks so much for joining us nice to see you Liz it's great to be here great to have you the superlatives used to describe not just your earnings but your outlook ranged from I was writing these down full steam Full Throttle incredible but but the headline really emerges that you guys have a problem most CEOs would love to have demand for two of your products right both your Hopper chip and the orders for your new Blackwell chip far exceed the supply you have what's the most important step you guys are taking to speed up the process to fill these orders well first of all I appreciate that uh you know Hopper and Blackwell are two uh just magnificent Computing systems and and what they've done is these two Computing systems has really really reinvented Computing all together from software that was written by humans running on CPUs to now uh machine learning creating these neural networks that make possible artificial intelligence running on our gpus and so so as the world modernizes the data centers a trillion dollars of data centers around the world the demand is just incredible and so we're we're uh ramping both of these platforms incredibly hard we had a record quarter this last quarter we got it to a record quarter next quarter we're going to go from zero billion dollars of black well to many billions of dollars of black well and so it's it's going to be an incredible thing but the demand is just incredible yeah yeah I I'm so interested to know some numbers how big is the backlog for I guess both the hopper and the Blackwell system the successor that that a lot of people want to see are you seeing any customer Behavior Jensen that says you know what we're going to wait for the newer black whell model well you know people are modernizing their data centers and so they they need just needs to get going um and what I advise people to do is that uh and they are doing that which is we're going to create new systems every single year buy what you need that year there'll be something new that you can buy next year and then you'll just cost average your way through uh the way the reason why it's cost averaging is if we increase the performance every single year effectively you're reducing your cost for any amount of computation that you want instead of the classical old old data centers we're gonna you know several hundred billion dollars of capex each year will go into these um machine learning based and AI based systems and so we have we have hundreds of billions of dollars of of uh uh uh it systems that we're going to have to go build over the next several years I think the number's been put out there for all these companies whether they are Tech or just basic Enterprise ramping up and spending on AI is something like 350 billion but let me tackle the hyperscalers first meaning the 800 lb gorillas in the AI space who are desperate for your chips meta Google Microsoft SoftBank I would love to know how orders are filled meaning who gets preference I mean does it go by who ordered first or by the size of the customer well if you don't place an order it's hard for us to fulfill it and so most of them have placed orders uh we try we try to to uh uh be fair to with with everyone so that so that they can all get uh their AI infrastructure up and running um we also make sure that their infrastructure is ready and so between us and them uh there's no sense shipping computers if the data center and the power and and the Readiness of the system is not there and so so we work very very closely with all of our hyperscalers we've been planning for this for some time and um and you know I think I think at this point uh everybody has demonstrate everybody has stood up their AI infrastructure the grace Black Wall Systems uh I think uh oci has shown pictures of it uh Google has shown pictures of it Dell and cor we just showed a picture I think it was last night and and SAA at at his uh conference uh announced that they're going to have pre uh previews for their customers here pretty soon and they they've got their system up and running and so so everybody's really proud of the system they they've stood up and we're all just racing to get there yeah and and you know we're talking about hundreds of thousands of these and people are dying to get their hands on them can I just get some CL CL it on the question about the overheating this report over the weekend there was in the information this this report that said that some of these were overheating can you just tell me if that was an issue if it's true uh was it an issue basically with the Blackwell chip design a rack problem or was it simply that the data centers were not built to the right specs it's a it's a non it's a non thing it's probably it's probably somebody who's who uh heard somebody doing engineering integrating our systems into their data centers uh in the beginning there all kinds of engineering issues but these are normal things we're just doing engineering we know that the grace Black Wall Systems work and we know that all of the technology now works because we now demonstrated four of them standing up uh we have lots of them working at our company we have lots of them working within our customers companies and so so we're we're very confident in we're in good shape with with h Grace black o production ramp uh but it's still you know we have thousands of Engineers doing you know all kinds of things so I'm not sure what they heard but it's a non thing well it's it's a long way from 32 years ago when you founded the company with just a few employees and and a dream and as you've begin to really see this come to incredible fruition at least when it comes to generative AI I'm so interested to know about pricing how do you view pricing because if you're having trouble because you did say into few quarters for fiscal year 2026 that you would see some issues with filling these orders how can you keep the price the same and will it rise we always do pricing based on the value delivered and so we're very rigorous with respect to the simulation of these AI factories in a lot of ways the way to think about this Liz is is here's this Machinery this big huge machinery and you apply energy to it and what comes out of it is AI we call it generative AI in a lot of ways it's kind of like generating electricity in the last Industrial Revolution and they all these Machinery these large Machinery call AC generators and they produce electricity we're now producing this miracle called artificial intelligence you said that you're transforming and AI is transforming every industry every company out there all around the world and we see that we definitely see that but you have some of the loudest voices now in Silicon Valley I'm thinking Mark Andre and the venture capitalist from Andre and horo would saying AI has hit a wall it's plateaued it's it can't go further right now and he named two reasons he said due to both the lack of enough gpus and then the training data it's almost like he was pointing the finger at Nvidia uh that you guys aren't getting these out enough quickly enough what do you say to that well our demand is greater than our supply and so we need to get more out and and so um I we we work with with a thousands of startup companies and they're called AI native companies from the ground up they're built to be an company from the groundup this is a company that would develop uh capabilities and and new products and services that are made possible because of Aon and we're working with thousands of them and it's hard to to uh at this moment keep uh capacity and infrastructure available for every one of them and so we're we're working super hard to to get there as fast as we can which is the reason why we're ramping blackwall so hard um with respect to with respect to AI uh there's a fair amount of discussion about scaling laws and the scaling law basically says the more the larger the model the more data that you train it with and the more Computing therefore you need the more effective the AI becomes and so that's a we're we're kind of on our backf foot with respect to the amount of demand that's out there but we're running as fast as we can to get the supply out there well it feels like you're not on your back foot that you are standing pretty tall and holding ground because you are able to just crush it on earnings and your outlook is pretty impressive too uh you talked about growing customer bases in India and and obviously you will have a situation where you have to hold back on China depending on what the regulations are here in the United States but I'm interested to know if you're hearing from the Middle East they have had some real Ambitions here list in in just about every single country and it's sensible they now realize that their nation's data their language whatever their their their uh language happens to be their language their digital language encodes their nation and their people's knowledge history and culture and they they now realize that they have to take that digital dat that data and turn it into their nation's intelligence and that they can't they can't wait for someone else to come do it for them that they have to do it themselves the success that we've enjoyed is largely in the unit States but over time and very quickly whether it's in in the west or in the East uh we're going to see we're going to see just about every single country jumping onto this AI Revolution this age of AI as we finish up what do you say to people who are very concerned about this you hear them and they believe that AI is going to turn us into digital slaves and that the machines will rise and control all of us you've had these discussions before what's your gut reaction to that AI can do a lot of our tasks but they can't do 100% of our jobs that's the big idea the big idea the idea that that AI would would um uh somehow do 100% of somebody's job is very very unlikely and quite frankly impractical for any time near any any time in the Horizon what is very likely is that AI will do 20 30 40% of a 100 % of the jobs in the world so what that means is that AI is going to revolutionize and transform everybody's jobs and it's really important that everybody should learn how to use AI they're not going to lose their jobs to AI they're going to lose their jobs if they if it happens they're going to lose their jobs to someone who uses Ai and so uh that's that's the most practical way of thinking about it and for the foreseeable future that's absolutely going to be true and and um I jump on AI I would encourage everybody to use it and uh if you haven't had had a chance to use use chat GPT use perplexity use Gemini these are incredible use use Claud these are incredible AIS I use it every single day yes I would imagine you do just don't just don't replace newscasters and journalists please please J you're in great shape you're in great shape Liz we need you thank you very much Jensen we appreciate it all right I hope you're all doing great today the S and P finished the day about half of a percent higher and Nvidia finished the day up by about half of 1% as well I am very pleased with the results that Nvidia reported yesterday revenue for the quarter grew by more than $5 billion sequentially Nvidia beat on both the top and bottom lines and next quarter guidance was raised to $375 billion these are incredibly good numbers as usual the bar for NVIDIA was so extremely high going into earnings and so the stock traded slightly lower yesterday and after hours trading but as I said in yesterday's video I don't really care how Market participants react in the shortterm this is because these numbers are fantastic and the long-term thesis behind Nvidia and the build out of AI infrastructure remains firmly intact I've been saying for a while now that I would not be surprised if Nvidia were to double over the next 2 years and after seeing these results I honestly still think that it's possible given invidious it's also a relief to hear that the rumored overheating issue is essentially a non-issue if there is any overheating happening it sounds like it's just a one-off occurrence possibly coming from a unique customer act setup and it's not a consistent problem across all of the blackwall racks that is very good news but simply after this earnings report the long-term feis remains intact demand for invidious products is insane hyperscalers have yet to show any signs of reducing their capex spending on the buildout of AI infrastructure and we have a clear product road map from Nvidia for the next two years this year we get the launch of Blackwell then Blackwell Ultra in 2025 and later on Reuben in 2026 as long as everything remains peaceful in the Pacific and as long as none of the hyperscalers significantly reduce their capex spending on AI one to two years and things are looking really nice for q1 and Q2 of fiscal year 2026 as black wall production ramps up in volume Nvidia will report earnings for those two quarters in May and August of 2025
댓글 0