AI Earnings Growth Cyclical or Secular? Nigam Arora Talks MU, Chips & Corrections

**SPEAKER_1** (0:00)
Let's bring Marlee back into the conversation, turn our attention to a potentially market-defining question with our friend Nigam Arora, founder of the Aurora Report. Hi, Nigam, nice to check in with you. So just talk us through the key question that could define the stock market moving into 2028 You believe that it is whether this earnings growth is cyclical or secular. I turn to the memory trade to, I guess, start to wrap our heads around that question because that's got a massive question mark over it right now, given these long-term contractual agreements, given the pricing power that these companies have, given the supply constraints that we're seeing. Just walk us through your thinking here.

**Nigam Arora** (0:44)
Sure, Marlee, you set me up just right. Perfect. Memory trade. Let's talk about something like Micron. Micron trades at six times forward PE. And I know something about Micron. We bought at about $21 and we're still holding it for, you know, 3900 percent gain or something like that.
So if Micron continues to show this trend that we're seeing now for next seven to 10 years, somewhere here Micron stock is going to hit $3000.
On the other hand, if Micron earnings starts sputtering around 27, 28, Micron is going to go down to three to $400. So that tells you the range of the outcomes that can happen, right? So, cyclical meaning the earnings are not going to last very long, it's a cycle, it's going to be over. Secular means they're just going to go on for a long time. So, the real question is, are these earnings, the tremendous earnings that we're seeing, is it secular or is this cyclical?

**SPEAKER_3** (1:55)
So, Nigam, if that's the case then, and we accept that either argument could be correct because we don't have a crystal ball as far as I know and we don't know the answer, then what data is of the most importance for us to be watching to determine which camp is going to be right here?

**Nigam Arora** (2:15)
Well, the kind of things that I watch, I watch, for example, memory. I watch high bandwidth memory pricing. I mean, Micron doesn't tell me what they're negotiating with each customer, but there's a gray market where you can buy high bandwidth memory and you can find out how the prices are doing. I watch the lead times, you know, if I want so much of this memory, how long is it gonna take me to get it? Longer the lead time, meaning, you know, memory is gonna stay in demand longer. There's still a big shortage.
I kind of do some calculations based on all the capacity that's coming online, how much supply is that going to be? And then we project what the growth is going to be, and where do they come in equilibrium? And that's very important because it's not just a shortage of memory, it's what they're able to charge, right? So they're able to charge a lot right now. And if the demand and supply comes in equilibrium, then the prices are going to collapse. So those are the immediate things like memory, but overall, I watch that for all different kinds of semiconductors. And then, you know, where is the demand coming from right now? It's coming from our assumption that AI is going to add to productivity and the usage is going to keep on going higher and higher.
But then there, you know, there are things there too, you know. There are Chinese models which don't cost us a lot, maybe sometimes even 1 to 10 percent of the frontier models. Now, the market's calculations right now are, I think they're probably correct, that about 80 percent of the economic benefit will accrue to frontier models, and only about 20 percent to the open weight models.
But if that calculation is off, then that could have a big impact on how, you know, various semiconductors are used and what the earnings really are. And then after that, you know, there's a, so this is another thing that I've seen in my couple of decades plus several times. So we have a cycle right now, which I'll say is our chatbots and agentic AI that we've seen. Now, that's one cycle, but then there's another cycle that comes right after that, and another cycle comes right after that, kind of think of an ocean wave, one big wave after another. So various cycles can merge, and overall impact can be even though each individual cycle is cyclical, overall it turns out to be secular, right? So we have autonomous driving, we'll have to see the adoption rate, because that's a big user of semiconductor can drive up our links. We have robotics coming up, that's going to be a huge user of semiconductors, but we don't know what the adoption rates of all those technologies are going to be, right? We talk about space data centers, the quantum computing, and of course, I think defense spending is going to keep on going higher and higher, and it's all the modern techniques, the drones, the killer lasers and all that, and then again, it's all semiconductors.