We convened a Club meeting at Palm Beach Gardens on Saturday. We just didn’t know it. My wife, Lisa, and I were signing bottles of her finest Fosforo mezcal at Total Wine & More, and I can’t believe how many people showed. It reminds me that Florida has to be the venue for our next Club convention. We have too many fans to count. I wish I could say that we had 150 Tobala aficionados. We did have quite a few. But the first, and some would say, the only order of business was what I am doing out here in San Francisco right now, going to meet Nvidia founder and CEO Jensen Huang, the self-effacing genius I call DaVinci, because like the earlier version, Jensen has been thinking decades ahead. I’d say centuries but that might be a tad hyperbolic. He’s amazing. Jensen’s speech at Nvidia’s annual GTC developers conference on Monday will be a transformative moment in artificial intelligence as the website, complete with a countdown clock, boasts. There was this pretty great piece about Jensen from a couple of months ago in The New Yorker , where there are several references to how he doesn’t like public speaking. You won’t know that from Monday’s speech. He likes to break news, he likes to explain, and he likes to compete. This speech will be all that and more. Will it propel the stock of this AI chip powerhouse that’s up a staggering 77% year to date after more than tripling in 2023? I don’t think I would ever say this about anyone on else, not even Lisa Su, CEO of Nvidia’s rival Advanced Micro Devices , another competitor from the same coaching family tree. But will you give the man a break? I think Jensen deserves one day a year when he can trace out a vision without a per share attached to it. One day when he doesn’t have to explain how he can handle the transition from the H-100 to the H-200 or if he can keep those gross margins at 70 plus. I remember once going to hear the late Intel CEO Andy Grove speak at an event after the chipmaker had just crushed it in the 486 to Pentium transition and the guy didn’t even speak. What do we want to hear? Jensen is not the predictions sort of guy, but he’s about big think as anyone who walks on this planet, including Tesla CEO Elon Musk, and I think he wants to stress sovereign AI, health care, and the planet — maybe even in that order. If you look at all of the panel discussions at GTC, you will see plenty about how to train and how to infer and how to put accelerated computing and generative AI to work for you and your organization. Content generation by machines who are just better at it than we are. Or will be soon enough. The practitioners he’s got speaking are the real who’s who, not the sightseers, and I am old enough to admit being overmatched by the discussion. I will need titles that explain what’s being said, something that the machines could give me if I had only asked earlier. Now, I want to give Jensen a break so he can outthink all of us and get us right with him. But I also have to address how the stock and the semiconductor group have been acting subpar. Remember, I said Jensen deserves his moment, not me. I have to suit up just like every other day. Let me give you a little Nvidia zeitgeist. Right now, I think Wall Street is real nervous. Almost everyone has a buy on it. The price targets keep getting bumped just because the stock has overrun the price targets, not because there’s something new, something different, something additive or crazy positive. We’ve already had that. There’s a sense that this just can’t continue. Sure, plenty of people in line Saturday would, correctly I think, say to “own it, don’t trade it.” My mantra for Nvidia (and Apple ). The expectations are too high. The margins too high. The prices too high — of Nvidia’s graphic processing units (GPUs) — not the stock. The stock is actually cheap on forward numbers, which also makes people uneasy. They maybe lack the history: the stock has not been one that has a lights out multiple, because too many analysts have learned that, in retrospect, the stock was cheap. So maybe it will be again and they don’t want to sell it. They don’t want to miss another huge quarter. But I want to point out that if you looked at this stock last year from July to January it did nothing. Just sat there. All I heard was that it had a fantastic but unrepeatable quarter. Sure, I was sweating “own it, don’t trade it.” Who wouldn’t? NVDA 1Y mountain Nvidia 1 year I think it’s entirely possible that it does the same thing again, the only difference being that it is at a much larger market capitalization: nearly $2.2 trillion. Nvidia’s market cap on July 3, 2023 was $1.05 trillion. That, by the way, is the gravamen against the stock. It’s just too big versus Jensen’s opportunity. Why? Well, let’s go there and explain the naysayer case. First, we really have no reason to believe that any one company could have such a hammerlock on generative AI and accelerated computing. It’s just not possible. Second, Nvidia might be backordered now, but we know what happens when the channel is filled, right? Third, it may be a great company, but for it to have gotten into the trillion dollar market cap class so fast makes no sense. It’s a gaming chip company for heaven’s sakes. Fourth, companies that grow 40% to 50% a year tend to crash and burn. How can this company grow are more than 200% and not suffer the same consequences? Fifth, who the heck is this guy with the black leather jacket anyway? Let’s take them head on and knock them down. Artificial intelligence has long been considered a real loser of a gameplan for any company. There’s a big graveyard for those who have tried. It was a graveyard for Jensen, too, we just didn’t know he was toiling in it because he was doing so well with gaming chips. The whole time he had a decade’s worth of thought and development he just couldn’t get it commercialized. I think it’s because the whole thing is so crazy amazing that it almost seemed chimerical. Then along came Sam Altman, an old friend of Jensen’s and CEO of OpenAI, the company behind ChatGPT. People fell in love with the ChatGPT concept and realized that it was all based on supersonic chips that only Jensen made because everyone else was smitten by Moore’s law — named for Gordon Moore, the late co-founder of Intel — and a belief that transistors would get smaller and more powerful every couple of years. That was the Intel way. Still is. But it has lost its utility. When Gordon Moore first came up with it, Intel was run by him, Robert Noyce, Intel’s other co-founder, and Andy Grove, who went on to be CEO. Intel back then was every bit as insanely cool as Nvidia is now. There was this remarkable press conference where Noyce held up a pen and said that Intel was developing a chip that size that would be more powerful than an IBM computer. Some belligerent blowhard laughed and said that who would pay $2 million for a piece of hardware the size of a ballpoint pen. To which Noyce retorted, they won’t. They will pay $200. Could be apocryphal — but wow, can you imagine the moment that the microprocessor was born? And, can you imagine if you owned but didn’t trade Intel from 1990 to 2000? That’s where we are with Jensen. He has the field to himself just like Intel did because nobody was as smart as the people who worked at Intel and no one is as smart as Jensen. Even Musk, who left the Nvidia fold to develop his own chips, had to eat a crow morsel — like he would allow himself more — when he couldn’t do it and went back to paying for Jensen’s product. The ironic difference is that while Moore’s law is dead. (Don’t tell current Intel CEO Pat Gelsinger, who thinks it is still alive.) Nvidia’s chips are heavier and bigger and aren’t meant for the personal computer. They are meant for the data center so who cares how big they are. I know Jensen teased a product with HP Chief Executive Enrique Lores two weeks ago, one that AMD has more than teased already. But Intel got so hung up on the size of the microprocessor that it stopped thinking about speed and about comprehension and learning on a grand scale. The fact that Jensen is the leader in both accelerated computing and generative AI — meaning the ability to generate content, cogent, usable, smart content, explains why the stock is where it is. The man is two for two in the biggest, most disruptive changes in the history of computing because everything that is currently about computing is just second-rate compared to Nvidia’s wares. You could say that Jensen can’t maintain his price point. I would come back and say two things: he’s already upgrading his chips, but big companies are happy to take whatever he gives them and if you look at the total cost of ownership (TCO), it’s actually pretty cheap. He’s got the figures, I don’t but I have only seen one person publicly confront him on that, Frank Slootman, the recently retired CEO of Snowflake, and Jensen demurred. No sense fighting with Frank. I think his price point stays or even goes higher. Why not, supply and demand matter. I think that there isn’t anyone close. AMD’s Su is selling GPUs, but Jensen is selling a platform with software and product has learned things so it can go to work for you immediately. It seems like everyone in the enterprise software world is writing for Nvidia’s product, so speed doesn’t matter. Inference, the output that comes from feeding the beast of a supercomputer, belongs to Nvidia. It’s not a tie goes to the runner situation. It’s a little more winner take all, loser takes something. How about the double and triple order phenomenon, and how once the channel is filled there will be a collapse in price? Lots to refute here. There are thousands upon thousands of potential customers out there that will be customers in the next few years. They just don’t know it yet. Nvidia doesn’t sell its chips to unknown customers. There is no channel. Nvidia has teams of engineers that work with clients and Nvidia knows what is out there. It can’t meet demand but, more important, there are gigantic companies that would probably take everything his company makes if they could. He’s rationing because he has no choice not to. When he comes up with the H-200 in scale the H-100 will still work plenty well. No channel. No destock. No shortfall. The size? This is a tough one because it is awfully sudden. But it does all go back to Jensen’s legendary missed quarter, the one where he did $5 billion in revenue when he was supposed to do a fraction of that. Once you realized he was on a totally different run rate you had to adjust what you thought his company was really worth, so you had that mistaken value to the downside. It had to come up. Now here’s what’s hard to figure. It’s incredibly difficult to value a company that has a huge trajectory. But, as I always like to tell people who ask about stocks that trade in crazy patterns, you have to value it at something you might as well value it as something that’s a measurement you can grasp versus others. The amazing thing here is that you aren’t doing this one at a price to sales or some adjusted free cash flow metric. The darned thing is actually inexpensive on 2025 earnings estimates, maybe because no one believes or maybe because people think it is a peak number. I believe and I can’t see how it can peak with all of these verticals not even scratched yet. The crash and burn crowd has not met Colette Kress, the CFO, who has a command of the numbers that is about as fine as any chief financial officer I know. She has a firm hand on what could go wrong. What could go right take care of itself. Finally, Jensen’s wife dresses him. So does mine. You have a problem with that? Bottom line So now let’s go full circle. Tech has had a good run. It feels toppy. Not the first time this year. There are no real new data points on the horizon for now. We are nowhere near earnings season yet. There are plenty of Nvidia camp followers that have no idea what it does, and they are itching to sell. I think they will. I say let them. I just don’t have a case to dump it. The naysayers just don’t have enough to scare me out of it, con me out of it or talk me out of it. So, my take? Just stay the course. And, if it really gets hammered? Buy some. (Jim Cramer’s Charitable Trust is long NVDA, AAPL. See here for a full list of the stocks.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
We convened a Club meeting at Palm Beach Gardens on Saturday. We just didn’t know it. My wife, Lisa, and I were signing bottles of her finest Fosforo mezcal at Total Wine & More, and I can’t believe how many people showed. It reminds me that Florida has to be the venue for our next Club convention. We have too many fans to count.
I wish I could say that we had 150 Tobala aficionados. We did have quite a few. But the first, and some would say, the only order of business was what I am doing out here in San Francisco right now, going to meet Nvidia founder and CEO Jensen Huang, the self-effacing genius I call DaVinci, because like the earlier version, Jensen has been thinking decades ahead. I’d say centuries but that might be a tad hyperbolic.
He’s amazing.
Read the full article here