'There's A Lot More To Come' In The AI World Says Aliaga

Added: 2026-05-30

[00:00:00]All right, Stephanie, so break it down for us.

[00:00:01]What are the big trends you're seeing driving this investment and why it's still a good prospect. It's been a huge inflection in the demand towards AI that is really come with this new era of a genetic AI.

[00:00:13]But what we're seeing in the markets today, despite the fact that it you know, the stock market performance has been remarkable.

[00:00:19]The semiconductor complex is up 80% since the end of March, which is just remarkable. The exceptional exceptionalism has been in earnings. And this is really due to the fact that as everyone is getting ramped up around their genetic processes, they are racing to secure compute. But the thing about this investment today, relative to the investment we saw a year or two ago, all of that investment was really around training models, and we didn't know what the ROI was going to be around training models. But now the demand and the investment is around inference. And we've we've been proven with a genetic coding tools with like cloud code that these tools are going to be very productivity enhancing. So now this CapEx has a lot more runway and that, you know, we're now suddenly looking at CapEx potentially approaching $1 trillion as early as next year. How do you know we're going to see the ROI when it comes to the investment for inference?

[00:01:14]Not everyone's going to have the same ROI.

[00:01:17]And I think that's definitely going to be evident.

[00:01:19]You know, there will all be all winners and losers.

[00:01:21]But the fact that you have, I mean, the agency that has been achieved around coding in and of itself is transformative.

[00:01:28]And it's not the only thing that is going to be automated by I.

[00:01:33]It is just the first real proof point. We have these tech firms essentially writing no code themselves. I'm vibe coding myself.

[00:01:39]I avoided every Python class they tried to throw at me.

[00:01:42]Um, and you've been successful in vibe coding?

[00:01:44]Yeah. I mean, I think the question is like, I've been successful on the research side.

[00:01:49]You know, the challenge for my employer will eventually be like, or if employers at large is, what is that ROI look like at the bottom line?

[00:01:57]And that is a lot more to do than just curious employees playing with these tools, but actually a more systematic rethinking, re architecture.

[00:02:07]Ring of labor. Yeah.

[00:02:09]And that is really yet to come. The reason I asked about the outcome of the vibe coding is because I did some, but I tried to do some vibe coding.

[00:02:14]The vibes were bad. I do the disco and it didn't.

[00:02:17]It didn't survive there. Mediocre at that.

[00:02:19]So I'm still doing this whole show. It is a new skill to learn.

[00:02:22]And so I think I was I wasn't using the right software, I was just using like an enterprise version of ChatGPT. What I need to use is like clock code.

[00:02:29]Yeah. Okay.

[00:02:30]Wait. But to that point, when we did the show together a little while ago, you were utilizing some AI tools.

[00:02:35]I was not, and and we were kind of doing it side by side.

[00:02:38]Yeah, a lot of times I was able to find the answers faster just going the old school way. And I'm wondering if you can talk to us.

[00:02:45]I'm seeing a little bit of a disconnect now between what Tim is just talking about the actual ROI for the product here, because we've got some reporting.

[00:02:52]You know, Microsoft has canceled a lot of its licensing over Uber.

[00:02:56]Excuse me? Uber burned through its entire 2026 AI budget in four months. Uh, you've got a fortune 20 CEO ordering token spending to be dramatically slashed.

[00:03:07]It seems like there are places it works. But as these companies are discovering, there's a lot of places it's not working.

[00:03:12]And yet we're seeing this spike in investment on the hard and on chips, on storage, on memory. So I want to take a step back.

[00:03:18]And, you know, as an economist, like every new technology that enters the scene and it has this period to this is the J curve, right?

[00:03:26]It goes negative in the beginning because you need to spend more money on the technology than it's making for you. People need to figure out how to use the technology that takes time, self-discovery, business discovery.

[00:03:36]All of that should be expected. What?

[00:03:38]We're just in a market where, yes, so much is relying on AI.

[00:03:41]So we're rightfully being very, you know, inquisitive and questioning what that ROI is going to be. But we gotta let this thing play out a little bit. You know, you when you first use a new tool, right. You're not going to be immediately more productive and effective with it until you build your own skills around clod, until you sit down and workshop agents that work for you and your workflow and your audience. But all of that takes either your own time or it takes your businesses time to actually engineer all of that.

[00:04:08]I think what is challenged these budgets right now is this, uh, trend of token maxing, you know, because I was so heavily subsidized.

[00:04:17]You could just use as much as you can. And, um, that was the badge on a wall and throw it all at the wall, which is great in this in the context of we do need to experiment with this. Everyone needs to just get on the, on on the bandwagon, you know, but it's not efficient when we're in a world of very tight compute capacity. So it raises the question about that.

[00:04:37]I ask everybody because nobody knows the answer to this, including Fed chair, former Fed Chair Jay Powell when he was asked about this all the time, which is the productivity gains that that we will see and in how people and economists are quantifying those right now.

[00:04:51]How are you quantifying those in terms of the aggregate data?

[00:04:54]Yeah, we don't see a lot of evidence, which is also to be expected.

[00:04:58]We don't see a lot of evidence for productivity gains.

[00:05:00]Well, I guess in terms of the broad, you said, Christy, you know, we're thinking in terms of the broad macro because this is the thing again, economist hat.

[00:05:08]Right. When you're in the early innings of a transformative technology, you're going to see pockets of significant productivity. You can see it.

[00:05:15]You hear anecdotes, you have companies. You have the new era of firms that are just founded by one person, and they're only one person.

[00:05:21]You know, all of those are examples of really significant productivity gains.

[00:05:24]Is that scaled across corporate America? I mean, the tech sector accounts for 50% of the S&P 500. 2% of the labor market and just about 5% of GDP. So, you know, there is a big disconnect there between what is actually diffused in the broad economy.

[00:05:41]All of that is to come. The early signals do tell us a pretty constructive picture of what is to come. And I will also say, beyond just the innate productivity of all of us being more productive at work, there is the capital investment wave underway, and that we think is the most tangible impact on the economy right now because it's going into the real economy, it's going to real hardware and infrastructure, and that is showing up in the data. So why do you think we're still seeing that hard investment on the almost the back end of this technology?

[00:06:10]Because they figure that we'll we'll figure out Tim and I will figure out how to use it at some point. But they want that infrastructure to be there when that moment comes. Well, I think the demand is already there for that infrastructure. I mean, every earnings season, the hyperscalers are saying, you know, our revenues would be higher today if we had the capacity to service it. So demand is accelerating massively.

[00:06:30]We're still pretty early footholds of usage of these tools.

[00:06:34]I mean, how much how many agents do we interact with on a day to day basis?

[00:06:38]How many do you think we will be interacting with in five, ten years from now? We haven't even gotten into the world of physical AI. Health care is often here.

[00:06:45]Jensen Wong, every time he gives an interview is like, it's going to be the one industry that is most impacted, and there's no killer health care AI stocks just yet. You know, at least in the broad scheme of it. Also, there's a lot more to come.

[00:06:55]Do you have any agents acting on your behalf yet?

[00:06:58]Um, I have a few, like, research agents I have to call on specifically.

[00:07:02]Yeah. Um, but they are there.

[00:07:04]They've been trained, you know, as I would train my own research, uh, analyst, uh, to prioritize resources. And are they are they, like, I've heard different, different feedback from how effective the work has been.

[00:07:15]Sometimes it's like, okay, I've heard people say, yeah, it's like kind of an intern, you know, you need to. You know, check its work.

[00:07:20]Yeah, you need to check the work. You check the work.

[00:07:22]But they're eager. Yeah.

[00:07:23]Uh, uh, what would you say? Like, how would you care?

[00:07:26]So I have there's this great chart in one of our decks, uh, showing hallucination rates by frontier battles. And they range from, like, 25%.

[00:07:35]The best you can get to 90% today. Okay.

[00:07:38]But the trick is, what is so great about these reasoning models is you can just ramp up the reasoning. You can ramp up the iteration so you have one response that it gives you. Right.

[00:07:48]And then I have it go through a number of skills around fact checking around voice correcting, um, around, you know, poking holes at what I've just written.

[00:07:56]And if you have it, iterate and iterate on its own before it even gets to you.

[00:08:00]What you get is going to look a lot better than the standard, you know, generative AI query that you were using and getting a year ago.

[00:08:07]So we have made a lot of progress. The question now is around orchestration, right. Because hallucinations will never be zero, nor are they zero for humans, you know.

[00:08:16]But I think the question is how can we manage that?

[00:08:18]I would like to say I hate all of that, but I do, I do.

[00:08:21]All right. It's a crazy world.

[00:08:22]I want to ask in the bigger scope. The U.S.

[00:08:24]is still really dominating the high momentum aspects of that I portfolio, but it's not a big club. Not that many countries are in this portfolio. Are there other places that you can see strategic competition coming up and what does that mean for the field overall?

[00:08:36]Yeah. Well, I will say even though it's not that many countries, the fact that it is more than the U.S.

[00:08:40]right now is quite important. I mean, in emerging markets, we've looked at this. And if you look at the tech sector weighting in Em relative to the U.S., IBM is now more weighted towards tech than the US because of the dominance that you've seen in Korea and Taiwan and just a few companies there. Right.

[00:08:58]But it is more than that. Um, the Em Asia complex is so central to so many pieces of the semi ecosystem. Um, and also all of the equipment that is going to be needed around, you know, the peripheral execution and implementation of these tools. Um, and then beyond that, materials, you know, if you're selling copper in this market, like you're pretty well, uh, positioned. Right.

[00:09:20]And a lot of those economies are in Latin America.

[00:09:22]So we have seen a global broadening out of eye.

[00:09:25]Things are still concentrated in the sense of it all, but it mean the world outside of eye is a lot less exciting than the structural growth dynamics that we are seeing right now. Is there an area geographically in this world you don't want to be invested in? Um, well, I think there's some areas that I'm definitely a lot more cautious. And I mean, look, you know, I mentioned Latin America just right now, like there's some like, you know, pockets of our resources. You know, resource exports do well.

[00:09:49]But then, you know, those economies have not been very well, um, uh, equipped at harnessing new technologies or even advancing where the puck is going.

[00:09:58]Um, a lot of political swing, a lot of political, social change policy and government support and import export, all that things.

[00:10:04]And I think look, the risk around, you know, the Strait of Hormuz is not just evaporated and there are some economies that are going to be more impacted.

[00:10:11]I think Europe definitely isn't as benefiting from being at the frontier of AI, and they're more sensitive to this. So some more caution there.

[00:10:20]Um, but I think, you know, exposure in the broad sense of it all makes sense in a globally diversified portfolio. Stephanie Aliaga, good to see you.

[00:10:27]Thank you. Thanks for coming in.