Carl, thanks very much. Welcome to the Halftime Report. I'm Scott Wapner. Front and center this hour, techs continue to take off. That Nvidia hitting a new record high, the biggest IPO of the year coming to market today. We're trading it with the committee. Joining me for the hour, Josh Brown, Malcolm Ethridge, Bill Baruch, and also with us today at Post 9. We're calling him a special committee member today. He's Brad Gerstner. He is here. He's the Altimeter Capital founder and CEO. He was an early investor in Cerebras, which is going public today, and we're awaiting that first trade. I do want to show you what the markets are doing today. New highs for the S&P, new highs for the Nasdaq. The Dow has retaken 50,000 yet again. Nasdaq's on pace for its seventh positive week in a row. It is the first time it's done that since October of 2024.
Let's begin with the tech trade for obvious reasons. Nvidia, for obvious reasons. That new record high today. You did have that Reuters report. The US has cleared the H200 chip sales to 10 firms. May not be entirely new, but we are learning names. And as Jensen Huang is in China, we're wondering what he might come back with. The street remains optimistic. UBS has taken the price target to 275. Canaccord goes to 350. So there is a lot of optimism for a stock that everybody here owns. You get the first shot at this because it's one of your largest positions. It's so good to have you in the house. It doesn't happen that often, if ever. So we're so happy to have you. What should we make of what Nvidia is doing and what Jensen Huang could bring home on this trip?
I mean, the most unusual thing is Nvidia didn't do anything for 6 months as the rest of the semiconductor trade really catapulted to new highs. So over 6 months, remember, Nvidia hit $210 this year about 6 and a half months ago. And it was flat over that period of time. And so, it's our largest holding. It's almost 20% of our overall portfolio. But it's actually been a drag from a performance perspective up until recently. On the March 30th low, I think it hit 170 or 172 bucks a share. And so, the catch-up here really hasn't been about China at all, Scott. And by the way, China's de minimis at this point to Nvidia. They've announced a trillion dollars over the course of the next 6 to 8 quarters of demand on Blackwells and now Vera Rubins. And so, this is a little bit of icing on top of the cake to be able to sell their older H200s to China. I don't think that's what's moving the stock. They're going to report earnings next week. And I think what's moving the stock is everybody's looking at the entire complex and saying the most important company globally in AI, the best company executing today in AI is Nvidia. Why is it trading at 14 or 15 times fully taxed GAAP earnings when the rest of the complex is trading...
What's so crazy is this stock sat at 180...
Yes. Anyone who wanted it could have had it.
For sure. And it was at that price the cheapest multiple on forward earnings you've been able to buy Nvidia since before ChatGPT. So, it's just this astonishing thing where in plain sight Amazon rallied, Alphabet rallied, AMD, Broadcom, one after another. They just took off. And this thing was just sitting there.
Is there another one of these that's hiding in plain sight? As you look over your current holdings or stocks that you're not in, what do you think is another candidate that we're ignoring right now that has this type of potential?
You know, I love it. I love it when you come on the set and say, 'What is the thing that will go up 40% over the next year?' Welcome to the market that we're in. I mean, the fact of the matter is I'm not smart enough, nor are most retail investors at home, to pick it that closely, right? The reason we've been in Nvidia is because we know they are the market leader. We know that they are the cheapest in the trade, and so we just had to assess at some point in time numbers would matter. The economic law of gravity here is that if they deliver the returns that we expect that they're going to deliver, then one or two things is going to happen. Either they'll buy back all their own shares because the market won't value them, or the market will buy them. And I think the market at this point in time is saying this doesn't make any sense, but if you said to me, Josh, Brad, what is the catalyst that finally got this thing moving? I couldn't tell you.
Yeah. Here's my question, though. As we sit here and talk about Nvidia being at all-time highs, I personally own the stock, so I have a vested interest in it going up and to the right. But as I look at a company like Cerebras getting ready to IPO, what they do is proposing to come for Nvidia's lunch, right? Like their approach to inference is completely different from what Nvidia does. It's completely different from the Vera Rubin cluster that fills the rack that we're supposed to get excited about this summer after we hear about what it's going to change and what demand for that looks like. So maybe that explains why the stock had such a hard time blowing past 180, and now we're seeing the excitement as the entire semis space has moved up, but realistically after earnings, a little bit of air gets let out of that balloon because we have a chance to learn that even if it's not Cerebras, somebody like that is coming for those 80% margins.
Sure, but I would just say before Brad answers, I mean, Brad and Jensen Huang himself have been so bullish and optimistic about what inference is going to mean. Two years ago, you and Jensen were sitting on the podcast together talking about this very issue, how they both saw what this was going to mean. I want you to watch the clip and we can talk on the other side.
Inference is about ready because of chain of reasoning. Right? It's about...
It's about to go up by a billion times.
Right, by a billion X. That's the part that most people haven't completely internalized. This is that industry we were talking about. This is the industrial revolution.
My take from that is that in an industrial revolution in industry there's not going to be one winner, not one big winner. There are going to be many big winners and this company's going to be right along there, right?
It's so extraordinary watching that. You know, Josh, you and I were just talking. Sometimes these things are hiding in plain sight. Jensen told us. Like we knew inference time reasoning was going to change the game. We knew the world was moving to inference. He told you it was going to 1 billion X. And the reality is not many people listened. He went on to buy Groq for 20 billion dollars. Cerebras is going to go public today. These are inference factories, purpose-built chips in order to deliver low-latency inference to all the people in the world. And for the folks at home, here's the one thing to remember. There is no intelligence. There's no consumer ChatGPT. There's no enterprise intelligence. There's no cloud code without the production of tokens, right? It's the production and consumption of tokens that is intelligence. And what Cerebras does such a good job at, Groq as well with Nvidia, is producing very low-latency inference tokens. There is room for more than one winner to your point, Scott. This idea that Nvidia was going to lose share, you're right, is the thing that capped Nvidia at 180 bucks. But I think the world woke up to the fact that notwithstanding Cerebras' success, notwithstanding Broadcom's success, notwithstanding the success of Trainium and TPU, Nvidia was still going to sell everything.
Companies in the S&P 500 that specifically said on their conference calls this earnings season they can't believe the cost of compute and tokens. That is now becoming a material factor in the profitability of the enterprise users of AI and all of these products. Does a company like Cerebras or many companies like that, is what we're seeing the race to solve that problem so that the price of compute doesn't outrun our willingness to do more and more things with AI?
Right. Remember in the early phases, '23 and '24, we were solving for pre-training of LLMs. That required a lot of flops, right? We weren't consuming a lot of inference. Now the models themselves to learn, they need to have an inference loop. Agents require inference, and ChatGPT exploded. So the world moved from flops to being able to solve what they called the memory wall, moving KB cache and moving these model weights around on a chip. And the way that Cerebras, remember, we invested in Cerebras 8 to 9 years ago. And the idea was a long shot that they could do something nobody had ever done and put memory right on this wafer-scale chip right next to the compute. They pulled it off, and the world moved to them. Now it really matters. They have an unlimited amount of demand for their product. So the challenge for Nvidia today, the challenge for Cerebras, is can we find the power? Can we find the data centers to stand up all of this in order to support it? The one thing you heard on all those calls, Josh, whether it was Amazon or Google, we are compute constrained. Our revenues would be higher if we weren't compute constrained. And Anthropic's compute constrained, OpenAI's compute constrained, Cerebras will help solve part of that constraint.
Speaking of demand, you priced at 185, the indication is 335. As we await the very first trade, you will see it here. I promise you that. If it happens within the next 50 minutes, you will see it, and we will discuss it. I want to get back to that in a minute, but the man sitting quiet to your right is the one who I have to get to today who has heard all this bullishness about Nvidia, and I don't think feels any differently, but trimmed the stock.
Yeah, I absolutely do not feel any different. This is position management. We loved it back, too, that Nvidia does not have the CapEx that the other hyperscalers have. This got to be an over 8% position in a portfolio or under an 8% position in a portfolio. Scott, you called me out on the show, active and tactical. We trimmed this down to about just over a 6% position. It's been a 50 plus percent move in the semiconductor space just this year alone and last month basically 30%. We're managing risk. So, we have a concentrated portfolio. It's better than a 10% position Nvidia still. So, I still love the name. But, for us, we've seen this run up and then we have the Trump-Xi meeting at the time of this trim, Jensen was not joining that party. And then we have next week we have Nvidia's earnings. So, I think what...
Try and hit recall, recall, recall.
I'm happy with the move. I'm not... I mean, when it made sense, with the market moving up what was that Tuesday night and the news changed. But, I'm happy still a top position at both portfolios. I'm happy with where we are. And it's been our top position since the back half of 2022. I mean, it's always been right there. So, nothing changes in our mind. But, we're a little more tactical here. Across the moves we've made selling things that haven't had momentum and kind of trying to lean into the semi space over the past month or two. It gets us in a really great spot. We have about 7% cash. I'm looking forward to over the next week, next couple of months to get to work.
So, you made the point as we again wait for the first trade here indicated far higher than 185 is where it prices above the range. Last check was around 335. So, we'll watch for that. But, you made the point as our viewers are watching this and this is the first of the really huge AI IPOs to come. It's the biggest since Snowflake. It's the biggest chip one since ARM. They're like, this looks really tasty for me. You made the point with us when you were on a week or so ago that companies like this maybe to a degree not as much as a SpaceX or an OpenAI or an Anthropic have been essentially already IPO'd in the private market. So, temper your expectations as a retail investor.
Yeah, I always try to shoot straight when I come on here and I see some of these retail investors placing bids without limits, Josh, at $375, $400, in order to buy Cerebras stock. This is not a get-rich-quick scheme. This is a generational company that I believe if you're in it for the next two or three years will compound much higher because we're on a generational transformational wave around AI and inference and they are delivering the inference factories to the world. With that said, I think you just have to be cognizant of the moment. You just described we've had epic moves in semiconductors. We have an all-time high today for Nvidia. So, we got a lot of excitement around here. There'll be plenty of time if you want to be on the train and compounding with Cerebras to buy this stock. It obviously moved up on the pricing. Didn't start at 185. We moved it up to 185, which is where it ultimately priced. I think the institutional demand for this stock is probably in the $250 range, $275 range. If you said, 'Brad, where are you getting excited about buying more of this stock?' I would say it's more in that range. That's what I'm hearing from the institutions. I think most of the demand over 300 is frankly retail. Listen, we've locked retail out of all of these great value creation stories. You know it's a pet peeve of mine. That's why I want to do Invest America accounts, the Trump accounts. Every one of these people should be able to participate in the upside of this compounding. So, I understand that they want to get in on the action and all I would say is caution a little bit today. Don't have uncaptured orders. Don't get this thing filled at some astronomically high level and then regret it and be selling it. Don't worry about the quick flip. Buy this on a fair price over a reasonable period of time and compound over the course of the next couple years.
Take that then from inference, if you will, to exuberance. Because that is where it feels like the market has taken us, Malcolm, of late. If you, yesterday Nebius and today you got Cisco today, which if you're in this space or claiming to be a big player in this space and you have a good earnings day, what used to be special, as I've been saying, like a 14 to 15% move, is now routine. It's just become a daily occurrence for many of these companies. So, how would you assess the exuberance that some are suggesting exists within this market?
This is Cisco's biggest gap up on earnings since 2002. Which is basically saying anything AI is going up. Anything that looks like it's attached to a data center is going up. But I have a little bit of a hard time understanding the exuberance on the side here that's saying Cerebras should be valued at 400 bucks a share when it IPOs considering what its technology is proposing to do, right? It basically strips away all the old ways of semiconductor manufacturing that Moore's Law taught us we need to get as many transistors as we can onto one chip and on smaller chips as time goes on. And says, 'Basically, give me the biggest chip you got and I'm going to put as many transistors on it as I can.' Which then means that we don't need as much data center space from a square footage perspective as we've been talking about. We're not talking about the level of power consumption. We're not talking about the amount of fiber optic cables coming from somebody like a Corning. So, all of the stocks that are going up and to the right because the amount of money that's being poured into this by the four hyperscalers spending 700 billion dollars, Cerebras's success flies in the face of that, right? To an extent, at least. And so, I think that the irrational exuberance on one side should be met with the cold water or the rationalization that everything can't go up into the right forever. And once something does come along that changes the way we've been doing this for this amount of time, what does that mean for me and the companies that I work with?
Okay, so that plays to I guess Josh the idea that Jim Chanos and I spoke about the other day. The idea that both the earnings and the spending are more cyclical than secular. Right now you can make the argument this is a fundamentally driven market. Sure, there's probably hype mixed in somewhere. Ed Yardeni has a note today where he says he's tracking the exuberance to which he says it's impossible not to recognize some of the more ridiculous anecdotes that indicate some folks are getting a bit over enthusiastic about AI. Companies are switching into AI from entirely unrelated business lines. Market's performance is dramatically narrowed and stock gains are boosting the bottom line for some companies, but at the same time tech earnings forecasts continue to rise sharply and at the end of the day that keeps us in the bullish camp.
Well, earnings and earnings were up 50% for tech. Right. So you look at a company, you look at SanDisk stock trades from $250 to $1,500 like in a couple of months. Okay, on the surface it's insanity. But then you look at their projected earnings and it's like, wait a minute, they're going to 8x their earnings over the next year. So what would you expect the stock to do? So that's the... So now the question becomes, all right, but are they going to do that in '27? Are they going to 8x their earnings again in '28? So then the bigger concern becomes lapping these miraculous, some would say, once-in-a-generation capex explosions. Can we lap it another year, another 2 years? Because ultimately the growth rates can stay high, but they can't match what we've just witnessed. That's my trouble with some areas of memory chips, semiconductors, but that doesn't apply to every stock. The semi index right now, even after having gone up 222% in total return from the liberation day low, is still selling at 27 times forward. 27 times forward in arguably the biggest boom for chip demand we've ever seen. It's not 57 times, it's not 100 times. I wish it were, be easier. Just hit the sell button. It's harder because the fundamentals are there.
Well, how about that? Your Taiwan Semi, Brad for example, says the global chip market is going to hit 1 and a half trillion by 2030 because of AI. That's a fundamental fly in the face of that this is all just hype and irrational exuberance driving this whole thing.
I mean, Josh just said, you have these companies that are growing at extraordinary rates. I mean, in the case of Nvidia over 70% and they were trading below market multiples, right? This is a radically different situation than what Cisco faced in 2002. And so, the people who have been propagating that myth over the last 2 and a half years have missed out on all of this upside. The question you're always asking is when is too much too much? But if you look at the memory stocks which we've been very large in for now 2 years, they're still trading at five to six times earnings. I'm going to have the CEO of Micron on the BG2 pod next week and this is what we're going to talk about. We are fundamentally restructuring these industries. So, if you said to me, what is going to be the signal? Right? What's the signal that something can go wrong? Okay, tell me the signal. Tom Brady launches a data center. You literally take the portfolio flat, take the rest of the year off. That's scary because the number of neo-labs or neo-clouds that were being pitched in our venture capital firm every single week...
A new neo-cloud, there are celebrities in these things. From Shaq to Shakira to Jay-Z. That's where we're going. But I would just say this, back to the neo-clouds and the biggest challenge we have right now is our ability to stand up all the compute. So, why might revenues fall short, Scott? For a company like Cerebras or anybody else? Well, if we can't get the compute to run all the things that we promised to run, we will have a problem. And you've heard the rumors and seen the headlines that 30 or 40% of compute now is delayed this year. So, keep your eye on power and compute. If delays persist, if we don't get these things stood up, it will roll back onto the revenue and the ROI of the capex spend, that could slow the whole trade.
Back to you in just a second. All right, Altimeter's Brad Gerstner is still with us. I said we had some more moves from Bill Baruch. And look, I get I don't like it's my pet peeve, I guess. I don't like portfolio management commentary when somebody makes a move because it's like, 'No, you know what? Like, you guys are portfolio managers, you're managing your portfolios. There are a lot of things that you could be doing around the edges.' So, I almost feel like it's used as more of an excuse to explain why you make a move. You trimmed Amazon and Alphabet. Now, I'm sure you're going to tell me exactly what I just said, but...
Well, the numbers speak for themselves what we put up. I mean, we outperform the S&P very well year after year. And the reason we do it is because we're able to take some stuff off the top. We're ready to put that cash to work when the market comes back. I'm not calling a top here, but we do see some cyclical highs that come in June from our business cycle analysis and other indicators. But what we're doing in Alphabet and Amazon specifically, I mean, look at it.
What do you mean from your business cycle analysis?
A business cycle analysis that overlays with wave analysis. And some of the proprietary indicators we're looking at with breadth, it does line up that we have a pullback. We're also in year two of a presidential cycle that this could be a bit of a not a blow-off but this rally, I don't know where it's going to extend to, but then we may not do much for the rest of the year and we may bottom out through the end of the year and look at it for a low in February. I want to have some cash because if we rotate or if things like that, I want to be able I don't want to just make no money for the rest of the year.
No, I understand but you know, blow-off top.
I'm not saying blow-off top. I'm saying this could be the top that we kind of...
This could be a blow-off. I said...
I don't think I said it. I don't know if it's going to be a blow-off but a run up that kind of...
He said he's a big boy. He knows what he said.
