00:00So I feel like we have to start with AI.
00:03You know, we were chatting in the commercial break about the AI build out and you were comparing it to
00:08the railroads.
00:09And it only feels like, you know, a slight exaggeration when you think about the force this has been when
00:14it comes to both the private and the public markets.
00:17It's really exciting if you step back and think of the numbers, which are kind of extraordinary.
00:22There's a number being thrown at $100 trillion of investment needs to go in in the next, you know, by
00:272040.
00:28And you think about that evolution and that amount of capital going to support growth and innovation.
00:36It's super exciting.
00:38And you could start with thinking that are the capital markets deep enough to support that kind of capital?
00:44And the great news is, yes, everyone is backing this.
00:49And you're seeing it in the private markets, the public markets, the equity markets, the debt markets.
00:55And there's a lot of innovation as well, which we're really excited about.
00:59When we created the Capital Solutions Group about a year and a little bit ago, we were really thinking about
01:05just this type of opportunity,
01:07which is pulling together our asset-backed financing, our credit finance, our best risk management,
01:13and our corporate client relationships to really solve problems, both for our clients on the issuing side,
01:19but importantly on the investing side, these are majorly interesting investment opportunities for our own asset and wealth management division,
01:26but also for our clients.
01:27So it's a really exciting time.
01:29And there's a lot of consternation about it.
01:32But I think we at Goldman Sachs fundamentally are just really excited.
01:35Yeah, absolutely.
01:36And, you know, another comparison I've seen is when it comes to this AI data center build-out in particular
01:42is comparing it to the fiber build-out.
01:44And I do wonder, you know, how you approach the question of whether or not there could be a risk
01:50of overbuilding here.
01:52I think in any disruptive technology, there's always a risk, a fervor.
01:57But what we see, again, if you just take a look at what needs to be built to support these
02:03models and to support this innovation,
02:05you know, I don't think we see it.
02:07But, you know, we keep an eye on it.
02:08And I know there were some articles written about bank exposure to the asset class.
02:13I know we at the firm are very comfortable with our exposure.
02:15But we obviously keep an eye on it, for sure.
02:18I mean, you talked about whether the capital markets are deep enough to kind of fund this AI and technology
02:22build-out.
02:22But kind of building off of Katie's question, are they smart enough?
02:25Because, I mean, there's a lot of money flowing around already.
02:29And I am curious if you think it's being put to the best use.
02:32Well, I think when you look at the structures that people are focused on,
02:35first of all, we take a look at just investment grade.
02:37It's a very liquid, deep market.
02:40It's a rated market.
02:41It's very well understood.
02:43When we look at the amount of capital that's come into the asset class, we'll call it AI writ large.
02:48Year to date, it's about $500 billion.
02:51That's private and public.
02:52It was $450 billion for the entire 2025.
02:57You know, private credit also does deep, deep diligence.
03:01And so if you look at the time that goes in to structuring these transactions, it is months and months.
03:10And there's a lot of negotiation around the leases, the real estate, the power.
03:15And so what we've seen is, yeah, people are being very, very diligent.
03:19There's an interesting transaction that we did at the end of last year for one of the fastest-growing energy
03:25companies.
03:26And they came to us and said, look, we've been in the private credit market,
03:30and we're interested in more innovation and potentially more flexible capital.
03:35What can you do?
03:36And we came up with a $5 billion package that went into what I'll call the classic markets, asset-based
03:42loans and high yield.
03:44They're back into us now, looking at preferred.
03:46And, again, we're doing deep diligence with these companies before we bring it to market.
03:51What about the duration mismatch for some of these yields?
03:54The idea that the technology is moving so fast so that, you know, we're talking more on the debt side.
03:59So if you're issuing debt with, obviously, a fixed time frame on it, but the technology, let's just say, accelerates
04:04and shifts a lot faster than that longer-term duration,
04:08is that a concern or is that just in my head?
04:10Well, I think if you think about from a data center perspective, I haven't read anything that would suggest we
04:16don't need data centers, right,
04:17and we don't need the technology of that infrastructure to support these models.
04:22I think the software and the language models are changing rapidly,
04:26but I don't think the infrastructure that underlies that and, importantly, the power required is going to change.
04:33And a lot of this investment is going into the data centers and the power and the compute.
Comments