00:00 Josh
The first pillar, not surprisingly, has investors excited, Tom. It will help capture this market by 2025. I’m curious how you see the AI industry and trends evolving in 2026, Tom.
00:30 tom
Good afternoon, Josh. Thank you very much for the interview and I wish you all a Happy New Year. Yeah, so, we described the four pillars of this rally, and as you said, the most important and the first pillar is artificial intelligence. It’s true that a rising tide lifts all boats, not just this year but over the past three and a half years. But we think this situation will change somewhat in 2026. Importantly, this doesn’t mean it’s bound to go away, on the contrary, I think enthusiasm for AI will look more like Q4 than Q2 and Q3. I mean it’s broken. So we’re not going to let every tech stock run away. It’s going to be a series of winners, like your Micron, and then there’s probably going to be some losers, like your Oracle and to some extent Qualcomm. I’m not talking specifically about stocks in 2026, but as an example, as artificial intelligence matures, we are going to have to be selective.
01:41 Josh
Well, let me ask you, is AI going to be more divisive next year, you know, if I were a viewer or an investor, I would listen. Tom, in this broken world of AI that you describe, how do I identify the winners and losers?
02:00 tom
So for me, I see AI as a cycle that’s maturing, right? In fact, the first stage is the epiphany stage, right? Oh my gosh, we can do it. In order to do this, the first thing we need is a semiconductor. Now we’re going to run out and buy all the semiconductors we can, and obviously Nvidia, Broadcom, TSMC are all benefiting a lot from this. I think it’s now evolved into those companies that are really leading the way. My top choice for AI in 2026 is probably Google. I like Gemini 3. I think their advantage in search gives them huge leverage against Open AI. So I like Google from a search perspective. I also like memory because that’s the next phase of data center construction, right? We need computing power. Now suddenly we need all that memory to power everything we need. So I think it’s more of an infrastructure play, moving towards memory on the infrastructure side, but also from a sort of hyperscale perspective, I really like Google and YouTube and then Gemini 3.