저자: TomLee · 외국인
요약 보기
Tom Lee는 AI 트레이드가 구조적으로 유효하며 엔비디아는 장기 보유 대상이라 주장한다. 소형주는 연준 금리인하 기대로 향후 12개월 강세가 유력하다고 본다. 엔비디아 실적 쇼크 후 조정은 매수 기회이며, 2026년 성장 우려가 주가를 눌렀다고 분석한다. 블랙웰 마진 소폭 하락은 영향이 제한적이라 평가한다. 다만 출연자 발언 중 일부는 제목상 다른 영상 구간이 섞인 것으로 보인다.
다듬은 글
we saw big comeback for NVIDIA we saw uh really some continued flexing strength from the likes of Apple which is now I would say firmly on top in the market cap race such as it is and then we've seen small caps struggle more than some might have hoped though today is an exception which is more uh influential in this market um they're both important and they reflect different you know investor views I think you you still always want to own the best companies in the most important themes I mean that's Nvidia and that's Apple um and they're going to really work over any cycle but when the market begins to believe the FED is going to move rates back towards neutral and that's both cost of money mortgages auto loan rates you know these are really big Tailwinds for cyclical stocks and small caps in particular so I think that's why I'd still want to make a very educated bet with high probability that small caps could do very well the next 12 months and even between now and your end it's been a while since we've talked about your granny shots uh remind us what your thesis is behind some of these stocks that you like yes uh well granny shots is built on the idea that there are themes like multi you know secular drivers of an economy and our original granny shot model looked at AI cyber security energy security um you know Millennials as a generation and then we we want to own the best stocks in those themes and that's really what was the Genesis of the granny shots and the names like Nvidia Apple I Nvidia has been in that list since 2019 so I think it's always important for investors not to be trading the best names but to really find the best companies and hold them for the cycle okay good reminder Tom Lee thanks for joining me here thank you this return to AI as a result of Jensen Long's comments at Goldman the Oracle exclusive interview outside of the White House killer everything we you and I went through of course Larry Ellison had a say on that Oracle conference call and everything that came out of the Oracle numbers Microsoft as well presenting at uh Goldman Sachs and communic copia also so focused in terms of AI and all the opportunity there uh and so you know are you more aever you've been a skeptic well no I've never been a skeptic in the power of AI to transform everything what I concerned about and continue to be is what it's going to mean for society in general I'm not a skeptic on on on what's going to ultimately happen there there are still different points of view one is in one is held more by the seoa camp which we all started talking about earlier the summer with that first report from seoa talking and questioning about the reever the return on invested capital from all these and whether it would all find its way into Apps eventually or how this is all going to play out from a value proposition but whether or not this is going to trans form everything no I'm I'm not a skeptic about that Jim you're not good because and I haven't been you know that yeah it's funny when you speak to Jess Wong what he says is you know this has been alcohol for like 18 months and it's supposed to cure cancer I mean I ask myself does it cure cancer and you're like you pause for a second to explain that that's not the point it's it's about getting data instantly that would take years to put together I like that well sist skepticism came from recent examples in which doctors were still able to detect uh cases of breast cancer better than AI right now no doubt about it but it's a right now situation just to point out uh as he pointed me out that he has no problems with how well inidia is going to do his question is the spend and how much now Oracle Ellison comes out and says the spend is going to be endless because you had to keep up with the neural networks there's no end to the spend it's like will they build that's one of the key questions this sort of idea that at least embraced by sequa there will be day when it's kind of done you won't need to spend as much on gpus all the data will be taken in the generative the models the large language models will not be complete but be largely complete in some way and you'll move on to the next phase but it won't necessarily be value that's created as a result of of uh of those that are running the models right but I do think that one of the things like you look at what Tesla is talking about I mean Tesla is talking about robots is talking about how to make real self-driving we talk about self-driving car like some of the applications I mean it's kind of like when you're when you took chemistry which I managed to drop because I knew it hurt my transcript what what I did was I recognized that this is about trying to do basic versus apply the basic's being done by Jensen the applied is being done by Elon and the applied is huge and but chips as well which brings us to Jensen's W Jensen Wong's appearance at the White House yesterday talking about Ai and the buildout Nationwide how much of a public private partnership they'll be with the government are Megan cassella caught up with them if you miss to take a listen we've got to make sure that that everybody understands the needs coming the opportunities of it the challenges of it and um and do it in the uh the most efficient and uh scalable way we can absolutely do you see the need for a federal role here in terms of federal investment in combination with private investment well there's probably going to have to be uh Public public private collaboration this area because because the uh the rate of uh growth is really quite High um all although although artificial intelligence uh consumes a lot of energy to train it also saves a lot of energy when you consume it uh because ultimately it's not about training the model it's about using the model covered a lot of ground yeah look that's really interesting when training using I mean that's just really important that I I think the world of lisu uh and AMD but they don't have as much training they have more inference and what matters is that you just kind of have this vast catalog of things they that is embedded when you get an Nvidia chip well it's not a chip there's seven chips on a black well we got to really get it's an 80 lb machine supercomputer we got to stop calling a chip cuz people just say oh it's put them all together and you get and that's what you get and obvious back to Ellison's comments though and his as well we talk about it every day power power consumption on the need need for it the fact that I mean Ellison actually did talk about having a team of EX utility Executives at Oracle who were figuring trying to figure things out given how many data centers they're building recommission referring as well to the power consumption we're recommissioning nuclear power power we have we have to we have no choice Nvidia hit a rough patch after it posted results in mid August but this week the stock is up roughly 15% joining us right now for a more on the company in the sector is Joseph Moore Morgan Stanley senior semiconductor analyst and uh let me just ask you Joseph what happened this week what what's pushing the momentum One Direction or the other on the stock yeah I think for me the bigger surprise was what happened in the prior couple of weeks you know the quarter was really good uh they beat by $2 billion which has never happened until Nvidia started doing it a year and a half ago and I understand they wrestling with these expectations issues um you know and there's some gross margin questions that came up but I was really surprised at the selloff this is our top pick demand is very very strong and you know at this point focusing on the sort of product cycles and things like that it's very important we want to see the company execute to that but but really the key issue is just the strength of the business and we heard that from Jensen this week at a conference we've heard that from a number now of significant Executives on the customer side how much they're committed to these workloads uh so I think this is still um you know a very compelling story and again to me I I was a little I don't know exactly why it was up this week I don't know why it was down so much last week but the stock is very reasonably valued uh with very good growth prospects here going forward Will Blackwell have lower gross M margins versus Hopper I think maybe a tiny bit um I think that when the product was launched in March they had sort of discussed some price points um you know on CNBC at one point and that's a little bit unusual I think they're very focused on can we continue to provide value can we continue to grow the market and I think they're probably more focused on that than the specifics of gross margin there's going to be a little bit of pressure here in the next couple of quarters which is really just the early stage of that we're ramping it up there's going to be low yields initially um I think they've sort of said over the full life cycle the gross margins will be similar but they're also expanding the scope of products there's going to be a microprocessor with Grace uh there's going to be an nlink switch that they sell into these racks there's a lot of things that they're selling that could have a little bit lower gross margin so we have gross margins down by about a point next year uh I don't think it's that big a deal in the context of How High the operating margins are and how much operating leverage there is but I think it is the answer to question is yes it is just a tad lower um but you know that doesn't change the fact that the gross margins are still very good so the stocks just under 120 you've got a $150 price Target on it what's the time frame for that because that's another 25% increase even if you think that they could have a a little bit of a rougher time over the next few quarters yeah look it's a 12- month price Target and you know I think that they have a very strong 12-month window ahead of us you know the new products that they're shipping Blackwell are effectively uh you know accounted for through 2025 so you know the question becomes kind of growth thereafter and that's when that's the debate it gets a little bit harder for the St work for me I like to stock best when there are these near-term issues that they're kind of working through because I have a lot of faith in just how strong this business is and so you know you've seen this a couple of times now over the last year and a half you saw lead times come down in the early part of this year people the people sold the stock on that when there was concern that Blackwell was almost too good and we'd have this kind of pause in in Hopper sales that Drew down the stock then we had the delay all those things sort of are great setups because you know you're selling pick and shovels into a gold rush and the strength of the Gold Rush is kind of the main thing here uh it doesn't you know the new products are really important and I want to get them out but you know it doesn't really change how good these numbers are so for me tactically we've liked it better when you have these issues you still have these questions overhanging the stock so I'm not necessarily calling for 150 right away but you know I think these are a good buying opportunity these issues that have been dragging on the stock of late Jim you're an owner of Applied Materials Nvidia Qualcomm perhaps even amongst others why did the why does the chip trade still feel like it could have upside from here uh two reasons the cycle and the secular story uh and when those two come together Dom it's pretty powerful so the secular story here is obviously the AI Advent on chips and that's where you're seeing me in Nvidia Qualcomm which is for the edge Computing of AI applic ations and Applied Materials which is filling out the equipment for these new semiconductor fabrication plants that need uh to be building these AI chips but on the other side the cycle I think is intact this is what you just said Rob I completely agree um we've had a volatile summer Dom uh you know that in terms of chips up until July mid July these were really powering the market higher it's been a little bit weak since then people questioning the AI trade I don't think that's where the weakness has come from frankly I think it's simply explained by the Yen carry unwind which still has been going on today it may sound trit to say that but I'm pretty sure that's what the source of the weakness is both the cycle and by that I mean the global economy powered by the US is very much intact for the chip sector and you've got the secular story on top of it this is not a place you want to be underweight I'm overweight big picture the AI story all the Thematic elements of it um has it caused you to reassess anything or not not really I mean and in part if you believe in where AI is going and the idea that AI in at some point whether it's 3 four 5 years out is going to provide us with you know superhuman intelligence um you know all of the the catex that's being spent today is wholeheartedly justifiable because the opportunity to monetize that AI at that period of time is is beyond I think what anyone has really contemplated in the market so far and let's I mean just to point to Nvidia in particular obviously there have been these periods before where it kind of you know consolidates for three months and people wonder if you can actually extrapolate the demand growth into next year and Beyond um what have you heard from them uh recently and I guess how are you mentally modeling it out in terms of how long this can last yeah so I think there's many different ways and many different vectors of of growth for NVIDIA um and the slope can change in any any given year however so so let's just go back to what happened at the at the earnings call earnings call the numbers didn't go up and it was the first time in I think seven quarters that the numbers haven't gone up for NVIDIA the stock comes in as people are saying 2025 is Peak Goldman Sachs conference um which was just this past week Jensen gets on stage and talks about how Coos capacity has to grow in 2025 but it also has has to grow in 2026 that basically marked the bottom for the stock and the stock went from 104 to 120 where it is today yeah so you know it is very sensitive to the growth trajectory of 2026 and whether or not 26 will be a growth year now in our opinion given how much capex needs to be spent 25 is not Peak capex we think 25 26 27 and potentially 28 are going to be big capex years for all of the hyperscalers really um it was interesting to me to hear Larry Ellison from Oracle talk about AI on on their conference call uh where you know he kind of had some impatience about this question of how are we going to get paid for these products and he essentially says it's going to be everything it's like everything we do is essentially going to be built around these AI Technologies um which is interesting to me except it also goes back to this idea well technology always gets better and faster and cheaper so how is this different right now which just an accelerated days of it yeah and we have to have like an hour to have that discussion um but I think in in a nutshell and and I'm going to this is going it's going to sound a little bit Terminator esque yeah so bear with me but if you can have a coworker that has equivalent intelligence to you and can be effectively your foil and do everything that you need it to do and then some um how much would you pay for that coworker so if Microsoft for example opens up that opportunity for for the world or for Enterprises how much are you willing to pay for that coworker is it $40,000 a year $50,000 a year right and and that coworker Works 247 yeah so today Microsoft monetizes at a rate of about $1,000 per seat so the opportunity is 50-fold so when you put it in in that kind of context um you know the the big picture on AI is that we're about it's a it's a complete transformation of the technology and how we use it and I think that's the part that sometimes gets lost um in in this AI kind of are We what is it a trade are we a peak are we because the end point is very very different um than I think the market thinks up about 0.5% and Nvidia was basically flat today and same thing with Tesla which finished the day up 210 of a percent Tesla is still trading inside of the supply zone so I would be cautious about adding until we have a clear break above and close above the supply Zone that's what I'll be watching for in the days ahead I still think Tesla has upside ahead as long as the robo taxi event is not delayed again and as long as the rest of the market can simply remain calm speaking of the rest of the market next week we have a very important fomc meeting the Federal Reserve is expected to cut interest rates for the first time in years as they begin a cycle of quantitative easing I think the most likely outcome of the September meeting is a 25 basis point cut but we could potentially see a 50 basis point cut either way I think fom seen next week could provide a Tailwind for stocks for the next couple of weeks however I think it is reasonable to be cautious near the end of this month and once we enter October because at that point Market participants will be fully focused on the election in November and we could see some volatile and erratic price action near the end of this month that lasts until after the election looking at nvidia's chart we did get another close today above the 50 Ma on The Daily time frame which is encouraging the 50 ma is the purple line on your screen now I think it is possible for NVIDIA to retest this resistance trend line at some point in the near future but there are two things I want to point out regarding the shorterm outlook on Nvidia first as I mentioned we have F andc next week and so the rest of the market will play a big role in nvidia's price action next week also the histogram on The Daily time frame is still moving upward and it has even turned positive however the pace at which the histogram is moving upward appears to be slowing and that is somewhat concerning in the shortterm as this could be signaling slowing upward momentum so that's what I'm seeing in the short term now let me give you my thoughts on the medium-term outlook for NVIDIA so after F andc next week I think it is very possible for stocks including envidia to catch a positive Tailwind that lasts for one or two weeks however although I think Nvidia could trade higher in the short term and possibly trade higher later this month I am cautious on Nvidia until late October early November there are a couple reasons why the first reason is that attention surrounding Nvidia typically moves in Cycles this is a simplified graphic but I'm just trying to communicate the underlying concept here this cycle has already occurred twice this year with Nvidia reaching a relative low in between earnings reports roughly three to four and yes I know the most recent relative low in August was exacerbated by the unwind of the carry trade but Nvidia was trading lower well before early August and the stock likely would have reached a relative low during this time regardless of the unwind situation now 't always have to if this trend were to repeat itself a third time this year then we could potentially see Nvidia reach a relative low sometime in late October early November additionally there is an election in November and that will weigh down on markets especially so once we get to late October and early November so I think it is reasonable to be cautious with Nvidia over the next month and a half now let me give you my long-term Outlook from a long-term perspective I am bullish on the future of Nvidia even if the stock trades lower in late October SL early November I still think this company has brighter days ahead demand is through the roof hyperscalers must invest in AI infrastructure now and Blackwell is on the horizon as I've mentioned in many videos recently if Blackwell had not been delayed by 3 months then Nvidia would have guided Q3 Revenue Higher by about $3 billion and therefore the trend of nvidia's growth rate acceleration would have remained intact this is very important because it means we could see invidious growth rate re accelerate in the future once Blackwell is shipping in volume I think one year from now and two years from now Nvidia will be worth substantially more than it is now I genuinely would not be surprised if this company doubles over the next two years given their current growth trajectory and their upcoming product Cycles so that is my long-term view on Nvidia before we end this video I just realized that I never thank you all for 10,000 subscribers in a video I made a community post about it not too long ago but I don't remember mentioning it in a video so thank you all for 10K I really do appreciate all of you I hope these videos help you in some way and it means a lot to me that you watch so thank you with all of that being said I hope you all have an amazing weekend
Comments 0