Estimated Capital Needs for AI Data Centers

With Morgan Stanley’s European Leveraged Finance Conference underway, our Head of Corporate Credit Research Andrew Sheets joins Chief Fixed Income Strategist Vishy Tirupattur to discuss private credit, M&A activity and AI infrastructure.

Read more insights from Morgan Stanley.

----- Transcript -----  

Andrew Sheets: Welcome to Thoughts on the Market. I'm Andrew Sheets, Head of Corporate Credit Research at Morgan Stanley

Vishy Tirupattur: And I'm Vishy Tirupattur, Morgan Stanley's Chief Fixed Income Strategist.

Andrew Sheets: Today, as we're hosting the Morgan Stanley European Leveraged Finance Conference, a discussion of three of the biggest topics on the minds of credit investors worldwide.

It's Thursday, October 16th at 4pm in London.

Vishy, it's so great to catch up with you here in London. I know you've been running around the world, quite literally, talking to investors about some of the biggest debates in credit – and that's exactly what we wanted to talk. We're here at Morgan Stanley's European Leveraged Finance Conference. We're talking with investors about the biggest debates, the biggest developments in credit markets, and there are really kind of three topics that stand out.

There's what's going on with private credit? What's going on with the merger and acquisition, the M&A cycle? And how are we going to fund all of this AI infrastructure?

And so maybe I'll throw the first question to you. We hear a lot about private credit, and so maybe just for the listener who's looking at a lot of different things. First, how do you define it? What are we really talking about when we're talking about private credit?

Vishy Tirupattur: So, Andrew, when we talk about private credit, the most common understanding of private credit is lending by non-banks to small and medium sized companies. And we probably will discuss a bit later that this definition is actually expanding much beyond this narrow definition. So, when you think about private credit and spend time understanding what is the credit in private credit, what it boils down to is on average, on a leveraged basis, the credit in private credit is comparable to, say CCC to B - on a coverage basis to the public markets.

So, the credits in the private credit market are weaker. But on the other hand, the quality of covenants in these deals is significantly better compared to the public credit markets. So, that's the credit in private credit.

Andrew Sheets: So, Vishy, with that in mind then, what is the concern in this market? Or conversely, where do people see the opportunity?

Vishy Tirupattur: So, the concern in this market comes from the opaqueness in these deals. Many of these private credit borrowers are not public filers. So not much is well known about what the underlying details are. But in a sense, a good part of the public markets, whether it's in high yield bonds or in the public, broadly syndicated leveraged loans are also not public filers. So, there is information asymmetry in those markets as well.

So, the issue is not the opaqueness of private markets, but opaqueness in credit in general. But that said, when you look at the metrics of leverage, coverage, cash on balance sheet…

Andrew Sheets: Because we can get some kind of high-level sense of what is in these portfolios...

Vishy Tirupattur: Yeah. And we look at all those metrics, and we look at a wide range of metrics. We don't get to the conclusion that we are at a precipice of some systemic risk exposure in credit. On the other hand, there are idiosyncratic issues. And these idiosyncratic issues have always been there and will remain there. And we would expect that the default rates are sticky around these levels, which are slightly above the long-term average levels, and we expect that to remain.

Andrew Sheets: So, you may see more dispersion within these portfolios. These are weaker, more cyclical, more levered companies. But overall, this is not something that we think at the moment is going to interrupt the credit cycle or the broader markets dynamic.

Vishy Tirupattur: Absolutely. That is exactly where we come down to.

So, Andrew, let me throw another question back at you. There's a lot of talk of growing M&A, growing LBO activity. And that could potentially lead to some challenges on the credit front. How do you look at it?

Andrew Sheets: So, I'd like to actually build upon your answer from private credit, right? Because I think a lot of the questions that we're getting from investors are around this question of how far along in this always, kind of, cyclical process; ebb and flow of lending aggressiveness are we? And, you know, this is a cycle that goes back a hundred years – of lenders becoming more conservative and tighter with lending. And then as times get good, they become somewhat looser. And initially that's fine. And then eventually something, something happens.

And so, I think we've seen the development of new markets like private credit that have opened up new lending opportunities and then also new questions. And I think we've also seen this question come up around M&A and corporate activity.

And as we start to see headlines of very large leveraged buyouts or LBOs, as we start to see more merger and acquisition – M&A – activity coming back; something we've at Morgan Stanley been believers in. Are we really starting to see the things that we saw in the year 2000, or in the year 2007, when you saw very active capital markets actually coinciding with kind of near the peak of equity markets near the top of major market cycles.

And in short, we do not think we're there yet. If we look at the actual volumes that we're seeing, we're actually a little bit below average in terms of corporate activity. There's really been a dearth of corporate activity after COVID. We're still catching up. Secondly, the big transactions that we're seeing are still more conservatively structured, which isn't usually what you see right at the end. And so, I think between these two things with still a lot of supportive factors for more corporate activity, we think we have further to go.

Vishy Tirupattur: On that point, Andrew, I think if you look at the LBOs that are happening today versus the LBOs that happened in the 2007 era, the equity contribution is dramatically different. You know, equity to debt, these LBOs that are happening today [are] of a substantially higher amount of equity contribution compared to the LBOs we saw pre-Financial Crisis…

Andrew Sheets: That's such a great point. And the listener may not know this, but Vishy and I were working together at Morgan Stanley prior to the Financial Crisis, and we were working in credit research when a lot of these LBOs were happening, and…

Vishy Tirupattur: And I used to be tall and good looking.

Andrew Sheets: (laughs) And they were just very different. We're still not there. If you go back and pull the numbers, you're looking at transactions still that are far more conservative than what we saw then. So, you know, this activity is cyclical, and I think we do have to watch deregulation, right? You saw a lot of regulations come in after the Financial Crisis that led to more conservative lending. If those regulations get rolled back, we could really move back towards more aggressive lending. But we haven't quite seen that yet.

Vishy Tirupattur: Absolutely not.

Andrew Sheets: And Vishy, maybe the third question that comes up a lot. We've covered private credit, which is very topical. We've covered kind of corporate aggressiveness. But maybe the icing on the cake. The biggest question is AI – and is AI spending?

And it just feels like every day you come into the office and there's another headline on CNBC or Bloomberg about another mega AI funding deal. And the question is, okay, where's all that money going to come from?

And maybe some of it comes from these companies themselves. They’re very profitable, but credit might have to fill in some of the gaps. And you and some of our colleagues have done a lot of work on this. Where do you think kind of the lending story and the borrowing story fits into this broader AI theme?

Vishy Tirupattur: Our estimate of simply data center related CapEx requirements are close to $3 trillion. You add the power required for the data centers and add another $300-400 billion. So, a lot of this CapEx will come from – roughly about half might come from the operating cash flows of the hyperscalers. But the rest, so [$]1.5 trillion plus, has to come through various channels of credit.

So, unsecured corporate credit, we think will play a fairly small role in this. Of that [$]1.5 trillion plus, maybe [$]200 billion to come from unsecured credit issuance by these hyperscalers, and perhaps some of the securitized markets, such as ABS and CMBS that rely on stabilized cash flows may be another 1[$]50 billion. But a different version of private credit, what we will call ABF or asset based finance, will play a very big role. So north of [$]800 billion we think will come from that kind of a private credit version of investment grade, or a private credit markets developing. So, this market is very much in the developmental mode.

So, one way or the other, for AI to go from where it is today to substantially improving productivity and the earnings of companies that has to go through CapEx; and that CapEx needs to go through credit markets.

Andrew Sheets: And I think that is so fascinating because, right Vishy, so much of the spending is still ahead of us. It hasn't even really started, if you look at the numbers.

Vishy Tirupattur: Absolutely. We are in the early stages of this CapEx cycle. We should expect to see a lot more CapEx and that CapEx train has to run through credit markets.

Andrew Sheets: So, Vishy, there's obviously a lot of history in financial markets of larger CapEx booms, and some of them work out well, and some of them don't. I mean, if you are trying to think about some of the dynamics of this funding for AI and data centers more broadly versus some of these other CapEx cycles that investors might be familiar with. Are there some similar dynamics and some key differences that you try to keep in mind?

Vishy Tirupattur: So, in terms of similarities, you know, they're big numbers, whichever way you cut it, these numbers are going to be big dollar numbers.

But there are substantial differences between the most recent CapEx boom that we saw towards the end of the late 90s, early 2000s; we saw a massive telecom boom, telecom related CapEx. The big difference is that spending was done by – predominantly by companies that had put debt on their balance sheet. They were already very leveraged. They were just barely investment grade or some below investment grade companies with not much cash on their balance sheet.

And you contrast that with today's world, much of this is being done by highly rated companies; the hyperscalers or between, you know, A+ to AAA rated companies, with a lot of cash on their balance sheets and with very little outstanding debt on their part.

On top of that, the kind of channels that exist today, you know, data center, ABS and CMBS, asset-based finance, joint venture kind of financing. All of these channels were simply not available back then. And the fact that they all are available today means that this risk of CapEx is actually much more widely distributed.

So that makes me feel a lot better about the evolution of this CapEx cycle compared to the most recent one we saw.

Andrew Sheets: Private credit, a rise in M&A and a very active funding market for AI. Three big topics that are defining the credit debate today. Vishy, thanks for taking the time to talk.

Vishy Tirupattur: Andrew, always fun to hang with you

Andrew Sheets: And thank you for listening. If you enjoy Thoughts on the Market, please leave us review wherever you listen and tell a friend or colleague about us today.