Speaker 1: Joining us right now to talk markets, the Fed, and much more is Greg Fleming. He's the president and CEO of Rockefeller Capital Management. He's also the former president of Morgan Stanley Wealth Management. Greg, we always appreciate having you here.

Speaker 2: Great to see you, Becky. Andrew?

Speaker 1: Let's talk news of the day first, and that is what happened yesterday. This massive shift in technology prices and some of those stock prices taking a big hit. Is it a buying opportunity? What happened here? Were they overvalued before, or did the world really change that much in a day?

Speaker 2: You know, I think we need more facts on exactly what happened yesterday. I mean, did this all happen with many fewer chips, or did they stockpile chips and have access to chips? Is the mousetrap better? You know, that remains to be seen, it seems to be. Is it actually that much less expensive? But, you know, Becky, my view is that the space is so transformative, and there's so much investment going into it, that there was bound to be innovation and progress made outside the U.S. as well. And the fact of the matter is, if prices are lower, that's going to pull demand in more and faster, and probably lead to even more investment in the space. You know, that's away from the specific issue on chips. But broadly speaking, I have a hard time seeing this as being a major shift in the investment resources. And I think it's going to be a major shift in everything going into A.I.

Speaker 1: and what it's going to do going forward. I guess the question becomes, is it a shift in who the winners are in this paradigm?

Speaker 2: That's a good question. And I think that is the question. You know, I think it's really early on that, too. I mean, the reality is that it's a large language model. I mean, Microsoft's $84 billion goes way beyond, you know, the A.I.s and the language models. So, you know, is it... And, again, everybody jumped on this without even really knowing. Yeah. Is it substantially fewer NVIDIA processors? Is it, you know, did they have, you know, ultimately a similar number that they had either bought or stockpiled? Is the mousetrap better? Is it actually that much cheaper? What went into the numbers? There's a lot of opacity.

Speaker 3: So I think we don't know the answers, but in terms of the black and white extremes of the ends, you have to say it's likely somewhere in the middle. So let's assume that they have some of the highest-end NVIDIA chips, right? But not...

Speaker 2: Not as many.

Speaker 3: But not as many.

Speaker 2: So there's progress. There's efficiency.

Speaker 3: I think the bigger issue is that it's sort of the mother of invention. Because of the constraints on the Chinese, just by default, given that it is harder for them to get the chip, the structure of the code is around not just efficacy, it's around efficiency. And I think if you look at what OpenAI has been doing, you look at what Gemini and Google, all of these guys have been doing, they've almost disregarded the question about, can we do this more efficiently for now? It was always, let's just leapfrog and try to create progress and we'll deal with the efficiency question later. This code, I mean, just on a material basis, from what little I understand about it, was as focused, again, on this sort of efficacy issue, which, again, may not be 100% of what OpenAI is. But if it's 90%, it then just raises the question, once everybody starts to focus on efficiency, how that changes the dynamic.

Speaker 2: Although, first of all, wouldn't you expect efficiency at this point? Because there's been massive amount of dollars going quickly to a transformative technology, and you always have efficiency as a focus. So you'd expect it, once a technology starts to move forward. So you'd expect it. And if the efficiency is actually that much, that dramatic move, then that will pull in more demand quickly, because it's lower cost, you're going to have more users, and doesn't more demand in a space that's as transformative as this, doesn't that ultimately lead to even more investment over time?

Speaker 3: It may shift, and winners may shift. Except there's this very interesting question about the open source of it all. So the closed models, I get that. And we were talking about how, and this was sort of the undercurrent of what I was learning in Davos all last week. So many CEOs playing with OpenAI, playing with Claude, which is an anthropic service, playing with Gemini, and then saying to themselves, you know what? This is really expensive. How much of this can I replicate using Lama or some other open source model? And so you have a lot of them already moving in that direction. That is a sea change unto itself that I don't think people were picking up. And now that there's a service that I think bests Lama, but I think this goes back to the question. And the other issue,

Speaker 1: what that is, it's not just cost. It's also having control of it, not being beholden to somebody who basically can hold you up for ransom by owning all of it.

Speaker 2: That's where part of the whole Sputnik, this notion of a Sputnik moment, if this was a software, if this was a company in the UK or in Norway, would that have been the reaction from a Sputnik standpoint? But I think ultimately what Becky said is really the question that is going to come out of this, which is winners and losers, not space. I think the space will continue to attract a tremendous amount of investment. It's going to change the nature of every single business and how people do things.

Speaker 3: But what I understand why you're saying that part of it, I think it's a smart question. It's the question I appreciate. The answer, I'm more complicated. Because if you say to yourself, Apple, which for whatever reason decided not to invest in this space the same way everybody else did, and yet they can be a winner because they're just going to deal with the edge compute at the end of the day and they don't need to actually build the frontier model. Right. Other people are going to say the same thing.

Speaker 2: Yeah, but that's winners and losers. That's not what's going to happen in the development of AI will continue. There will be efficiencies and companies like this that push it forward. But the impact of AI on every business and everything that we do, and therefore the amount of overall level of investment it attracts, I think is still the same trajectory.

Speaker 3: And that's a little bit of what Satya was saying yesterday in terms of it may get commoditized, but the demand level is just going to continue to rise. And that always pulls.

Speaker 2: It pulls in more investment.

Speaker 1: And I'm sorry, when I meant held for ransom, I didn't mean DeepSeek doing it as the Chinese company. I meant a closed system, like an opening AI with them hosting it. Do you want to be in control of your own destiny and have a little more ability to see what's going on and running things from an open source perspective on these two? That's right.