The AI Opportunity Everyone is Missing (It’s Not What You Think)

Yes.

I mean, as the listeners of this podcast know, I'm very bullish, but there are a lot of really good well-reasoned bearish arguments out there that i would like to just bring to the table so we can collectively discuss them and hopefully break them apart enough to dispel them fully now one of them is that you know semiconductor stocks obviously have been all the rage of the ai boom and the ai build out and lotty a lotty do but the big bear thesis on them over the last forever is that it's the most cyclical industry out there at least one of the most cyclical industries out there and not so much cyclical in terms of valuation but cyclical in terms of earnings their earnings shoot up like crazy they peak and then they go down like crazy and it's typically a multi-year cycle so right now we're in let's call it year three and a half maybe year four of the ai infrastructure semiconductor super cycle So how much longer does this one last? Should we be worried about a peak in the next, let's say, one to four quarters? And if so, does that mean the trade peaks 12 months before that? So that's what I'm going to toss off to you guys.

What do you think about that, the cyclicity argument behind the semiconductor super cycle that we're seeing right now?

I'll take it first and glad to be on the show.

I think if we think about where we are in terms of semiconductors and how much further we have going out, if you just look at a couple of things, number one, you can look at the earnings expectations by analysts.

And even as we speak today, they have been revised higher.

So I think just in the case of Micron, you look at SanDisk, you look at some of these other storage players, The numbers just continue to go in favor of the bulls.

The other point I would throw out is while historically we do know that memory, whenever they get these low PEs, that tends to be the top.

That's where they're at their peak earnings.

But that just doesn't seem to be the case.

It seems, I hate to say it, that they might be less cyclical than what they were in the past.

And you know what? If earnings and backlogs continue to be out to four to five years, which is what a lot of CEOs have been saying, these stocks look really, really cheap here.

I mean, Micron is under seven times.

Sandus is about the same.

So I think if anything, you've got the earnings picture going.

You've got the.

got the narrative that is that is out there a lot of people are spooked but i don't really see this train slowing anytime soon and the last point i would say is that money continues to prop these stocks stocks higher and so as long as you've got earnings going up you've got money flows pushing them up by institutions uh the safer bets probably to stay long

yeah and you know to add to uh luke downey's point um CEOs coming out and saying multi-year backlogs, bottlenecks, et cetera.

Jensen Huang has come out and said they're having a memory issue for at least two years.

You have your major memory producers from Samsung and from Micron saying, yeah, they're saying at least till the end of 2028.

That's just for memory issues.

As you pack more and more GPUs into clusters, you need more and more silicon.

You need more and more semiconductors.

Semiconductors, it's in the name, right? You can regulate how much electricity you can transmit through the material.

Pure silicon manufacturers, there aren't that many of them doing what they do well, and they're all spoken for, for supply for years down the chain.

Yes.

I mean, as the listeners of this podcast know, I'm very bullish, but there are a lot of really good well-reasoned bearish arguments out there that i would like to just bring to the table so we can collectively discuss them and hopefully break them apart enough to dispel them fully now one of them is that you know semiconductor stocks obviously have been all the rage of the ai boom and the ai build out and lotty a lotty do but the big bear thesis on them over the last forever is that it's the most cyclical industry out there at least one of the most cyclical industries out there and not so much cyclical in terms of valuation but cyclical in terms of earnings their earnings shoot up like crazy they peak and then they go down like crazy and it's typically a multi-year cycle so right now we're in let's call it year three and a half maybe year four of the ai infrastructure semiconductor super cycle So how much longer does this one last? Should we be worried about a peak in the next, let's say, one to four quarters? And if so, does that mean the trade peaks 12 months before that? So that's what I'm going to toss off to you guys.

What do you think about that, the cyclicity argument behind the semiconductor super cycle that we're seeing right now?

I'll take it first and glad to be on the show.

I think if we think about where we are in terms of semiconductors and how much further we have going out, if you just look at a couple of things, number one, you can look at the earnings expectations by analysts.

And even as we speak today, they have been revised higher.

So I think just in the case of Micron, you look at SanDisk, you look at some of these other storage players, The numbers just continue to go in favor of the bulls.

The other point I would throw out is while historically we do know that memory, whenever they get these low PEs, that tends to be the top.

That's where they're at their peak earnings.

But that just doesn't seem to be the case.

It seems, I hate to say it, that they might be less cyclical than what they were in the past.

And you know what? If earnings and backlogs continue to be out to four to five years, which is what a lot of CEOs have been saying, these stocks look really, really cheap here.

I mean, Micron is under seven times.

Sandus is about the same.

So I think if anything, you've got the earnings picture going.

You've got the.

got the narrative that is that is out there a lot of people are spooked but i don't really see this train slowing anytime soon and the last point i would say is that money continues to prop these stocks stocks higher and so as long as you've got earnings going up you've got money flows pushing them up by institutions uh the safer bets probably to stay long

yeah and you know to add to uh luke downey's point um CEOs coming out and saying multi-year backlogs, bottlenecks, et cetera.

Jensen Huang has come out and said they're having a memory issue for at least two years.

You have your major memory producers from Samsung and from Micron saying, yeah, they're saying at least till the end of 2028.

That's just for memory issues.

As you pack more and more GPUs into clusters, you need more and more silicon.

You need more and more semiconductors.

Semiconductors, it's in the name, right? You can regulate how much electricity you can transmit through the material.

Pure silicon manufacturers, there aren't that many of them doing what they do well, and they're all spoken for, for supply for years down the chain.