Is Any Company Worth $5 Trillion? (NVIDIA’s Moment) | Shimmy Says Ep. 31

Hey guys, it's shimmy. And you know what? I'm joined again.

I, I'm gonna have to start paying him to come if he keeps coming like this. I'm joined again by my friend Dan, Dan O'Brien, COO, president of Futurum, because this is something right in his wheelhouse today. What I want to talk about and what do I want to talk about, I'll tell you what I want to talk about, is any company worth 5 trillion, that's trillion with a T dollars.

It's crazy when you think about it. I, I remember when I, I think it was Apple was the first trillion dollar company, and, and that was a little, people were like, my God, we have a trillion dollar company. And that couldn't have been more than two, three years ago.

And here we are at $5 trillion. It's, it's a crazy amount of numbers, or it's a crazy number. Um, as, as I, I've heard it, I've heard it explained on some of the TV news shows, you know, that's not a number, it's a mood, right?

And, and I, I think that's true. It captures the mood, this, this feeding frenzy, maybe some may call it rational exuberance around ai. Danno, before we jump further, let, let me, uh, is this a revolution?

Is it a bubble? Is it something we've never seen before? Both.

What, what's your take a $5 trillion company? Uh, you know, I, I'd lean almost to revolution. I mean, it's certainly, you know, a, a fairly monopolistic player, you know, at the forefront of one of the greatest, you know, secular growth trends we've seen in the tech market in a long time.

Um, you know, do the numbers justify it? You know, you look at consensus over the next few years and, you know, let's put a little hedge on consensus too, right? I think, you know, Jensen just raised the, the next, you know, quarter to four quarters by 15 to 20% at the GTC event this week, right?

So, you know, wall Street, uh, has a habit of underestimating these things, and that gets corrected as time passes as well. But, um, you know, I I, I do think the numbers justify right. You know, I think the other thing we need to take into account here, right?

You know, $5 trillion, right? The dollar is the metric here. You know, uh, there's been a lot of talk about inflation and, you know, the devaluing of the dollar, you know, the way to kind of keep your dollars whole is to keep them in the equity markets, right?

But wait, but wait, let me, let me just ask you something. So you're telling me $5 trillion doesn't go as far as it used to go? Oh, it Used to be, Just say it's a mere bag of shells, 5 trillion.

You're right. But let, let me put it in context for you. It's roughly equal, maybe a little more than the, than the GDV of of Germany.

The only two countries in the world that have greater than $5 trillion and and significantly greater is, is the US and China. I think the US is at about 30, and China's at about 19, though that's hard to measure the way they do things. Um, And Nvidia knows no geographic ballots, right?

And, and, and it's not a nation state, but when you have a company that is on par with the leading nation states in the world, one has to ask themselves, are we all destined to work in, in the Nvidia company town? Right? That was something from the early industrial revolution, right?

Carnegie's and the Vanderbilts and these people, they would build towns where all their workers lived and worked and died and, you know, did their thing. I'm not sure we're gonna live in the Nvidia worker town, but, uh, you know, will we live and work in a place where the Nvidia ecosystem is, you know, is driving and is enabled, uh, you know, very, very well could be the place, right? You know, any, you know, any white collar work, you know, in, you know, as Nvidia says, they're in the intelligence business, right?

So anywhere where intelligence is gonna be deployed, um, they're in that space. And, you know, we, we gotta talk about this relative to the valuation, but, you know, physical ai, robotics, right? You know, as, uh, the robot revolution happens here, you know, I think that will start to touch all our lives in the next few years.

Well, I, I, so to me, it's not just the robotics or physical ai, as some people call it. Uh, it, it's quantum too. So you take that triumvirate of ai, quantum robotics that I think does initiate whatever, I don't know, are we at the fourth industrial revolution or the fifth industrial revolution that that does ignite the fifth industrial revolution, and I think it'll be the biggest change that we've seen in humanity maybe since fire, right?

I mean, I mean, it's, it's big. Um, but you know, certainly the street is obsessed. I I live in Boca Raton, it's a land of stockbrokers.

Um, I can't go to dinner. I can't, I can't even go to the gas station without hearing people talking about, you know, Nvidia the market and stuff like this. Why is the street so in love with Nvidia, Dan?

Well, it's obviously been an amazing story on its own. Um, but you look at, you know, the lift it's giving to companies across the economy and across the supply chain. You know, it's, it's kind of the beginning of the momentum on this flywheel that's really driving the macro economy right now.

Right? So everything from turbines to concrete to, you know, cables, you know, raw materials, um, you know, you've seen the entire semiconductor capital equipment landscape, the fabs, the memory companies, you know, it's really touching an enormous part of the economy. Um, and so, you know, I think therefore, you know, that's why you're hearing about it everywhere.

It's, it's, you know, it's kind of the, the spark that is igniting this fire, um, in terms of where the growth really is happening in, in the world right now. Sounds like a Billy Joel song. We didn't start the fire, but, um, but, and you know, when I look at Nvidia, the, there's two, there's almost the tale of two companies there.

One is a semiconductor company. They basically, you know, are the GPU market. They invented and through bad times, good times, all through the times, Nvidia, you know, really was a great GPU, the standard bearer.

And, and these GPUs are a large part of their valuation right now. But I think the untold story that I think not people like maybe you or I, who are a little bit more on the inside here, don't realize is they've created an amazing ecosystem of software. 'cause when you look at the evaluation, it's not a hardware company valuation, you know, revenue to to market cap.

It's a software company kind of valuation. And they say it's a Platform company valuation. Yeah, You're right.

But, you know, but when you look at their, one can make the argument, they're a damn good software company. Incredible software company. Yeah.

I mean, I think that's really what's, you know, what's really set this thing, um, you know, forward is that they've got a, a enormous developer community building on their platform, right? You know, they have have made it so that you can take this hardware and you can get it up and running and get it driving value for you pretty quickly. So, you know, the, the hard work was done over, you know, to decades prior to the, the hardware boom that we're seeing right now.

So here's something I think about. What really is their entry barrier to competing with Nvidia? Is it designing better chips, better GPUs, or better AI chips?

They don't necessarily have to be GPUs, or is it cracking their software kind of shell? Or is it both? To me, It's for the software side, right?

I mean, you know, um, a MD is probably, you know, really, you know, it's probably only two or three GPU makers in the world. Um, you know, Nvidia, a MD you know, you may wanna include Intel there, you know, with, with the Gaudi side of things. Uh, but, you know, a MD is clearly the most credible competitor to Nvidia, right?

They're bringing their initial rack scale solutions to market in 2026. Uh, but it's really the work that's been done on the, you know, in, uh, a MD software side of things that has made that platform, you know, start to become much more of a, a, a true apples to apples competitor with Nvidia, right? So, you know, everybody's using the same equipment, the same fabs to make their chips.

There's obviously still a lot of room for differentiation in and around, you know, the, the design of these chips and, you know, the floor plans and, you know, big decisions that people make about going, you know, triplet versus big monolithic ics. But, you know, I really think for a MD it's been the work they've done on the software side of the last couple years that's really closed that gap. Yeah.

And, and I'll tell you the truth, I don't know if people on the street realize that I, I think they're in love with the GPUs and the hardware. It's like, it's like boys in cars, you know what I mean? We could get ccs and how many, when I was growing up was how many barrel with rated?

Did you have a six pack as a carburetor versus it's a four barrel and, and you know, that kind of thing. But it's the software that I think really is their, their real moat, if you will. That's what builds their moat and castle.

Well, I think a lot of that Allen comes down to hardware is kind of objective, right? We can put it in a lab. That's true too.

You can test it Faster. Somebody's more power efficient. You know, there's lots of ways to kind of objectively quantify hardware performance software.

You know, it's, it's a lot, you know, it's a lot more subjective, right? Yeah. You know, I mean, it's, you see things like, you know, developers who go past Cuda and, you know, really optimize at the kernel level.

A lot of what deep seek really, you know, made some noise about earlier in the year. You know, software's a lot more subjective. You know, you can do a lot more, you know, a lot more things there to, you know, kind of twist the results a little bit one way or the other.

So, you know, I, I think people tend to focus on the object objective side of things. Uh, the subjective side gets, you know, a little bit pushed to the back burner because you know of the subjectivity. Yep.

So, Dan, I, I, you know, I'm usually a glass half full kind of guy. You Are. But, but I think we owe it to our audience to maybe look at the, the flip side here, which is, you know, sound economic theory tells you if you put all your chips into one basket, it's risky.

And are we putting too much into the Nvidia basket too much into AI in general? But if you're gonna put it in ai, we're putting it all into Nvidia. And, and that's a risk that right now it's right, it's floating all these boats, but if something goes wrong, well, could it sink all these boats as well?

Uh, there's certainly, you know, there's certainly a little bit of a risk there, you know, being so dependent on one company. 'cause I don't think people are putting so much, you know, weight into betting on Nvidia specifically, so much as they're betting on their ability to create business value in other areas with what Nvidia brings. Uh, there's obviously a lot of interest in divers diversification in the supply chain right now.

You've seen enormous deals happening with open AI and Broadcom open AI and a MD, um, you know, Intel, you know, is clearly a player there. Um, Google, you know, really seems to be a, you know, material, you know, play in the training space, you know, in addition to, uh, to Nvidia, you know, it feels like there's a lot more optionality on the inference side of things. You know, training really seems to be coming down right now to more Nvidia and Google with a and d you know, probably having a shot there a little bit longer term.

But I, I think the market is trying desperately to diversify its supply there. But, you know, right now in, you know, and NVIDIA's winning on Mary, And let me JI agree with you Dan. Let me just say this upfront to everyone out there.

Kudos to Jensen Long and the whole Nvidia team, they're kicking butt. I, you know, I mean, if I was in business school and, and doing, you know, business school case studies, this is a, this textbook case study, man, they're re forget it. They're rewriting the book, but there are some, there are some potential warts when you dig in vi right there.

One of them is, look, 40% of their revenue last month, or last quarter, excuse me, came from like two unnamed customers. Mm-hmm. Yeah.

A lot of customer concentrations for sure. You know, at the scale of which Nvidia is playing, you know, there's only a handful of companies that can really afford to invest in hyperscale data centers, right? I mean, we, we already saw that proven out with the cloud.

You know, AI infrastructure is proving to be no different. Yep. Another issue is, you know, look, president Trump says they're getting a deal done with China.

We've heard this stuff before, but geopolitical it, sovereignty, geopolitical risk, it sovereignty could rear their head here, right? And, and kind of put, I mean, it seems like the whole world is on this AI look emotive train, right? This high speed train to somewhere.

But I think it's foolish or naive for us to think that companies like a China or even some of our allies in Europe wouldn't like to see a competitor to, to an NVIDIA and the US for that matter in, in this, you know, race to the future, to the next millennial or the next century in the next industrial revolution. How, how likely Shy about that and, you know, they've really done everything they can to, you know, boost and prop up. Huawei is kind of their national champion on this side.

Um, but, you know, China has an intense demand for Nvidia stuff as well. Right. You know, I, I do think that ultimately, you know, through all the trade nego negotiations, that they will work something out there and, you know, it'll probably still be a performance limited, uh, Nvidia solution that comes over.

But, you know, I, I do think that China is upside for Nvidia, you know, as the market kind of views them, right? Absolutely. So I, I think there's a balancing act there that you pla you are walking a tight wire with no net.

If you, if you make it too hard for China to get the chips they need and of a sufficient, a sufficient quantity and quality, then you're spurring them to do their own research and, and come up with alternatives. Yeah. Constraint based thinking is a powerful thing.

You know, I I think it's sort of Necessity's the mother of invention, right? Absolutely. On the other hand, if you give it to them, well then there's strategic implications at play there and you gotta worry about reverse engineering and, and all of that kind of stuff.

So, but you use the right analogy. They're really walking the tight rope and, you know, the profit, try to give them enough that the, you know, internal efforts, you know, are, are somewhat slowed down. 8 trillion on, on, uh, data center, uh, in the next couple years, right?

And then video get a decent chunk of that as the CPU of or GP as the semi of choice, forget what it's called. Um, but that things could go south there too. You could have political change, you could have the nimby, you know, movement, you know, or, and this is something that I've thought about.

What happens if, and this has happened in the past, what happens if this whole AI thing just doesn't happen as fast as we think it's going to happen? You know, the repercussions here are, are big. I think they are.

Yeah. I mean, I, I I look at it as, you know, we are somewhat constrained for the next few years and it's, you know, kind of different constraints at different points in time in terms of just how quickly we can build out, right? You know, I, I think at a time, you know, there was, uh, you know, kind of raw wafer shortages, then it was more the packaging, then it was more the high band with memory.

I think really energy has moved to the forefront as the limiting factor now. And so, you know, there's a lot of these big commitments being made that play out, out over a number of years. But our ability to actually build out is somewhat constrained.

And so in the short term, I do see that supply demand imbalance continuing and you know, in theory the supply should continue to outstrip the demand. I do think that somewhat naturally limits how far we can overshoot in that, you know, if we get a few years down the road and the, you know, the demand side starts to wane a bit, we won't have put all the supply that we're talking about actually in place yet. Um, and so, you know, do we get push outs?

Do we get cancellations? You know, as you get into the later part of the decade, you know, maybe, and you know, obviously, you know, that would take some wind out of the sails of the whole movement here. But, you know, the next couple of years, I don't think there's a huge risk of that.

Uh, I don't necessarily disagree. You mentioned Broadcom an inference. And I think that is something, again, for those of you who aren't living this day to day, Dan said it before, we use the Nvidia chips and then Google and then a MD for training, which in the beginning of this AI revolution and generative ai, that's where the action was, right?

The data scientists building the LLMs training, the off the LLMs building, the AI's training, as we move into the use of AI will use, we'll see more and more inference taking place. And that's a different computational load and companies like Broadcom, right? Different optimization there.

Yep. Broadcom is, is probably one of the leaders, if not the leader in the world when it comes to the kinds of chips we're using for inferencing, is that They're really the partner of choice, right? I mean, they've, they've long been, you know, Google's design partner on the TPU, you know, we see them winning, you know, kinda really most of the battle.

Um, you know, for companies that really wanna build their own silicon. And this is largely a hyperscaler, you know, trend. You're seeing open ai, meta, um, you know, Google, AWS, uh, Microsoft, et cetera, all get interested in building their own chips.

And bright call's aren't working with all of them, but they're working with the majority of them. Um, and they seem to really be kind of winning the lion's share of the design opportunity there. You certainly see Marvell planning in that space.

You see media tech, all chip, you know, there's a few others in that space that are, you know, kind of clearly relevant as well. I think Intel, you know, definitely considering the custom AIC side of things with their ip, um, A MD could be a big player there as well, but Broadcom really is kind of the chip designer of choice for, you know, people who wanna build their own silicon to partner with. Yeah.

And, and, and so that bodes well for them. And that's good to hear. But I, I wanna come back to Nvidia and we'll end it here 'cause I know we're running a little long.

We're not talking, and I had this conversation on Textron gang with Mike Ard and team the other day. We're not talking about AI being part of the IT stack necessarily. It is part of the IT stack, don't get me wrong, but it's not just part of the IT stack.

So in other words, AI doesn't exist within this IT stack anymore. And I think that's really jenssen's message is AI is in essence a world unto itself. It's a civilization unto itself, if you will.

It, it, it transcends, it, it, in some ways it transcends tech. It's, it's almost like a, a humanity next evolutionary step kind of thing. And that if you buy into that five trillion's cheap right?

That's the intelligence business. Yeah. That and right.

That's the intelligence business. And I think that's something people need to take in into account as well. So look, let's end it up, Dan.

My shimmy take on it is, I respect the execution there. I just questioned the gravity. $5 trillion is a huge gravity.

Well, that's sucking everything, you know, into that. Well, and it's a concentration there that it's just scary. 'cause if it don't go well, man, someone's going to have to pay that, pay that bill.

Um, I think you're right for the next short term, I don't see anything on the horizon that darkens this star Numbers keep rising, right? They, they keep, uh, even the most aggressive estimates out there seem to keep getting beat. And you know, like I said, I, I think there's, you know, longer term growth plays they've gotten in and around things like fiscal and robotics that, you know, unclear how much of that's even in the valuation today.

Yeah. But again, to be fair, look at, at this valuation, things gotta break right for them. And Absolutely.

And, and that's what we gotta watch. Absolutely. So Dan, I'm going to ask the people out there, but before I do, I'm gonna ask you at 5 trillion you were buy, sell, or hold, Probably somewhere between buy and hold.

I mean, if you're not in it already, um, It's expensive. If you're not in, It's the, you know, a big part of the wave here. But, you know, 6 trillion, 8 trillion, I, I could see those being in play in the next couple of years.

You know, it doesn't feel like this is the top to me. Very cool. What about you out there?

We'd love to hear what you just tell it. Buy, sell, hold. We're live on LinkedIn.

We're live on X. You can comment, you can, whatever you want to do. We hope you enjoyed it, Dan.

Oh, I appreciate. I know you were busy today and you, you jumped out onto this for me. I appreciate you, man.

Happy to be here. Thanks Al. Thank you.

I appreciate you guys watching. This has been Shimmy says we'll have another one next week, but we hope you enjoy it. In the meantime, keep your eye on Nvidia.

It's the rocket ship that defies gravity. This is shimmy. We're out.