The Six Five team discusses NVIDIA 3rd Most Valuable Company.
If you are interested in watching the full episode you can check it out here.
Disclaimer: The Six Five Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we ask that you do not treat us as such.
Transcript:
Daniel Newman: The third-largest company this week as we saw another massive rise. And Pat, this is a daily changing item. When I think I put this topic down, NVIDIA had gained another AMD. They were sitting at close to 2.4 trillion of market cap, maybe even 2.5. I’m trying to pull the chart right now because I don’t want to undersell just how absurd this was. So they peaked out this week. They were trading at $972, which they closed yesterday at $875. So it’s a hundred dollars a share change, Pat. So the company’s market cap had skyrocketed up to 2.4, 2.5 trillion.
But in that past week, even in the past month, this is probably just worth calling out. The company had soared past Google, it soared past Amazon. It then recently soared past Saudi Aramco, which was the only non-technology. And then the irony, I think I tweeted something, and it wasn’t that clever, by the way. But I tweeted something along the lines of, I guess AI is the new oil.
Patrick Moorhead: No, because how many keynote speeches have you heard? And maybe
Daniel Newman: Data is the new oil, right?
Patrick Moorhead: .. Did the same thing. Exactly.
Daniel Newman: Data being the new oil, now AI being the new oil. Well, NVIDIA quickly passed, and by the way, is really one big earnings beat away from becoming the second or most valuable company in the world. That’s the interesting data point of where we’re at. And it’s happened in such a short period of time. If you go back and you want to go on a chart and you go back three months, the company’s market cap has literally grown by a trillion dollars in just the past few months. That’s like Qualcomm, Intel, AMD and a Broadcom. That’s all of them put together. Now this is A, indicative of just how important and visionary Jensen Huang is, Pat. This is also frothy, unreasonable, unruly and needs some check in terms of how fast it can grow, consolidate. And the rug pull last week was really interesting. I’m hoping-
Patrick Moorhead: Hey, did you see?
Daniel Newman: Go ahead.
Patrick Moorhead: No, no, go ahead.
Daniel Newman: I was just hoping to see this rug pull last a little longer so I can go to that MarketWatch piece and do another victory lap about how I told you it was getting a little bit frothy up there. What were you going to say?
Patrick Moorhead: No, I was going to ask, did you see the Jensen interview at Stanford? I mean, it’s pretty good. He’s throwing out a few things like, “We had a 100% inference market share.” I don’t think he wanted to say that, given the regulators. He also said that the next generation of their AI cards, which you and I are going to both see at GTC are going to be water-cooled, which I’m not going to talk about NVIDIA being third most valuable company.
I did want to pontificate a little on this. Water cooling, at least historically for me, has always been the, it’s what you do when you run out of options. Meaning, graphics card, we’re going to make it go faster, or a CPU, you do water cooling. If you want to win in a benchmark war, you do liquid nitrogen. I did a couple of those when I was at AMD, because we were running out of gas, right? This is what you do when you run out of gas. So Daniel, it kind of makes me wonder, have we pulled every lever so far? We’ve done node shrinks, right? We’ve got disaggregated designs with chiplets. We’ve optimized packaging, and whether that’s COAS, EMIB, Intel’s 3D packaging, I always forget the code name for that.
Daniel Newman: Foveros?
Patrick Moorhead: Yeah, Foveros. But what’s next? Are we going to literally start liquid nitrogen cooling these things? Are we going to move to more optical interconnects between GPUs as opposed to copper? Do we do a hard move to ASICs? Is it faster networking that’s better distributing the workload across more nodes? We are running into a freaking wall here, and I’m really interested to go to GTC to see what they have in store. How are we not hitting a wall?
Daniel Newman: How do we get more?
Patrick Moorhead: Yeah, how do we get the next big bump in performance at a reasonable power and heat, right? You saw that Amazon bought a freaking data center next to a nuclear power plant. I mean, that’s just so awesome, right? I mean, I’ve got nothing against water cooling. It’s funny, Steven’s response to me was, “Oh, nothing wrong with it.” I never said there’s nothing wrong with water cooling. I’m just saying, what rabbit do you pull out of your hat after this to get the next big bump in a year and a half or two?
Daniel Newman: Well, I mean, first of all, when you talk like that, it just makes it sound like you know something about semiconductors.
Patrick Moorhead: I don’t know.
Daniel Newman: I don’t think you do.
Patrick Moorhead: I mean, some people may have heard that I know nothing.
Daniel Newman: I just wanted you to know I don’t think you do.
Patrick Moorhead: Yeah, I just read about it.
Daniel Newman: Read and regurgitate.
Patrick Moorhead: Selfie analysis is what I like.
Daniel Newman: I had to scratch my head. But no, listen, I was talking to Pat Gelsinger a couple of weeks back and he was talking about, by the end of the decade, we’re talking about a thousand times order of magnitude more powerful AI silicon. A thousand times or more from where we are today. And so how do we get that performance? How do we do it sustainably? Again, practical sustainability, meaning we can’t have 60% of the world’s power being consumed by data centers. It will not work. And by the way, 100% of inference, Pat?
Patrick Moorhead: He said a 100% market share in inference. Now, in the context, I believe what he means is in the data center for large language models. Because the example that he used was, every time you do a prompt, put a prompt in, that’s inference. He’s absolutely right. That was the context.
Daniel Newman: No, I get that. But then if it’s 100%, then what about that one I did on the Groq? Well, I’m not talking about Groq on Twitter. I’m talking about on the Groq chip. Because I’m inferencing an LLM on Meta’s Llama on that chip. So it’s got to be like 99.9999%. But I’d like to put an asterisk to make sure that we’re accurate here, because I did an inference on a non-NVIDIA chip on an LLM. It’s happened. I’ve done one. Just saying, dude.
Author Information
Daniel is the CEO of The Futurum Group. Living his life at the intersection of people and technology, Daniel works with the world’s largest technology brands exploring Digital Transformation and how it is influencing the enterprise.
From the leading edge of AI to global technology policy, Daniel makes the connections between business, people and tech that are required for companies to benefit most from their technology investments. Daniel is a top 5 globally ranked industry analyst and his ideas are regularly cited or shared in television appearances by CNBC, Bloomberg, Wall Street Journal and hundreds of other sites around the world.
A 7x Best-Selling Author including his most recent book “Human/Machine.” Daniel is also a Forbes and MarketWatch (Dow Jones) contributor.
An MBA and Former Graduate Adjunct Faculty, Daniel is an Austin Texas transplant after 40 years in Chicago. His speaking takes him around the world each year as he shares his vision of the role technology will play in our future.
