Welcome to Pathways, a Macquarie Asset Management podcast where we provide fresh perspectives and insights for institutional investors and consultants about real assets, private markets, and macroeconomics.
Daniel McCormack:
Hi everyone and welcome to the Pathways podcast. I’m Daniel McCormack, MAM’s Head of Research. Private credit is one of the most nervously watched asset classes in the market right now. AI is disrupting software; cracks are appearing in direct lending; and geopolitical and economic uncertainty are adding to the disruption.
So today, we’re going to dive into this asset class, and I’m very pleased to say I’m joined by Harlan Cherniak, who’s our Head of Infrastructure Debt and Co‑Head of Private Credit in the Americas. He’s got over 20 years of investment experience in the space, so it’s great to have Harlan here.
Harlan, let’s start with the big picture. What are you actually seeing in private credit markets right now?
Harlan Cherniak:
Good morning, Dan. It’s great to be back here on your podcast. As you know, private credit has grown from almost $500 billion to $1.7 trillion over the last decade. That velocity has created both capital and mistakes. Our job today is to tell you which is which, and I think it’s a bit of a tale of two cities.
You’re starting to see the emergence of stress in certain pockets of the market, as well as real opportunity in others. This market is not monolithic. We believe that we’re in the great bifurcation.
Real assets and hard assets, those credits are holding up well. You’re seeing some signs of stress in leveraged software, as well as other pockets of technology, and those cracks are quite visible.
The wall of capital is meeting the wall of reality. We’ve seen spreads compress, structures have loosened, not everywhere, but in specific pockets, and the leverage loan default rate is approaching 5.5% as of January.
And across infrastructure credit, historically we’ve been below 1% across every cycle, prior to 9/11, through the GFC, and through COVID.
And covenant light loans seem to be meeting their moment. The structures that evolved during 2020 through 2022 are now being stress tested. And infrastructure credit, which never went down that covenant light path, that structural discipline matters.
Daniel McCormack:
Harlan, why don’t we dive straight into the sector that’s dominating the headlines, which is, which is software.
What is actually happening in private credit’s biggest single exposure?
Harlan Cherniak:
I think software may be having its Yellow Pages moment.
This is an opportunity where you have to double click on your portfolios, double down on your underwriting, and both credit selection and credit managers, they matter.
Software represents about 20 to 25% of the private credit market. And that compares to roughly 16% weighting in the leveraged loan index, with more than half of those credits rated B‑minus or below.
And today there’s about $25 billion of software loans that are trading below 80 cents on the dollar. It also represents almost 30% of all the distressed loans across the index.
And you have to roll back the tape and think about how we got here. Fueled by private equity, sticky annualised recurring revenues, businesses with high margins, long duration contracted cash flows, and loans that were often attaching at dollar zero and detaching between five and six times debt to EBITDA.
On paper, this was textbook private credit collateral. As a result of some of the adjustments to EBITDA, some of the challenges to run‑rate revenue, and some of the churn that we’re seeing across the SaaS landscape, as well as the pressure and or the perception of AI disruption or AI displacement, there has been a spotlight placed on this sector. And it’s a real question as to how AI is going to impact every software credit across the space.
We’ve looked at both the vertical and the horizontal impact. Deeply embedded industry operating systems that have low AI disruption risk relative to generic tools that have wider deployment and are at risk for being disrupted due to marketing automation. And those are the credits that you have to be highly careful of.
Need to think about the mainframe versus the Yellow Pages. Air traffic control software – nobody’s ripping that out. A generic analytics dashboard – AI may replace that in six months for ten percent of the cost.
And this is an opportunity for everybody to re‑underwrite and to understand which credits they own and the potential expected value range of outcomes and how those losses might impact portfolios.
Daniel McCormack:
Economic history is certainly littered with Yellow Pages‑type moments. I can think of several other companies off the top of my head that have gone the same way. You know, it’s not like we haven’t seen this before, but it is certainly a bit of a worry.
AI is disrupting a whole range of sectors, and I wonder, as an economist, what comes next after software. I’m sure there’ll be something.
But Harlan, perhaps we could just pivot and, and make this very real for people. We were talking offline, and you were talking about the app that we used at our MAM conference last year.
So, could you just talk to the audience about that and what you think is going to happen with that kind of software, and then, do you have an example of software that might be durable through this kind of period of disruption?
Harlan Cherniak:
Sure, I think both of those, Dan, are, are worth highlighting. As valuation multiples in the space from both a private equity, from a going private, and from a strategic perspective, some of them approached over 30 times enterprise value to EBITDA.
And so one of those examples was a portfolio company called Cvent. It was a $4.6 billion take‑private, over 30 times enterprise value to EBITDA.
And while leverage today is roughly 6.8 times using 2027 estimates, which sounds manageable, the peers that we tend to look at from a comparable company perspective and from an equity margin of safety standpoint are now trading down closer to nine times enterprise value to EBITDA.
Think of an Adobe or a Salesforce — blue‑chip software services assets that are facing the same potential headwinds as it relates to artificial intelligence. And so that equity cushion underpinning those credits has compressed dramatically. Terminal values have been revised down.
And so today, whether you’re looking at enterprise value to EBITDA or revenue multiples, the collateral values have been cut in half.
And you’re also seeing signals from other large private credit managers who either proactively or reactively have been reducing exposure to software across the space.
On the flip side, there are other software assets that are irreplaceable. Think about something like an Ancestry.com. It literally requires your DNA, your family tree and its history is embedded in the software. Three generations of relatives have uploaded all of their historical information. The switching cost is almost infinite. And there’s no AI start‑up that can replace 30 years of your family’s genealogical history.
So, what’s the lesson here? Software is not a monolith. The market is pricing it like it is. And that creates opportunities for the investors who are willing to do the work, and to look through their portfolios and ensure that there’s the appropriate measures of diversification, and the appropriate measures of control.
Let’s think about how we got here. You’ll remember last year, it all started with two credits, First Brands and Tricolor. Those were two specific incidents of credit fraud. And then came the cockroaches: When you find one, people start to look for others, whether it was the messaging from Jamie Dimon at JPMorgan or others.
Then it was AI. Well, AI software is the largest private credit sector. Simultaneously, all these challenges and all these nerves started to come about.
And then earlier this year, geopolitical tension, geopolitical shock, whether the US and Israeli attacks in Iran, and now with the closure of the Straits of Hormuz.
The impending chain of events between rising oil prices, consumer confidence, federal and fiscal monetary policy, they have a multitude of risk factors colliding at once.
And so, I think we have a question here is, are we in a garden‑variety default cycle, or are we in a systemic collapse like we saw during the GFC?
I think we can debate, you and I, about whether this is concentrated and not systemic, unless perhaps, we dive into some of your more recent views around the potential risk of oil.
Daniel McCormack:
Harlan, you’ve talked about before that AI is both disruptor and opportunity at the same time. Why does that matter?
Harlan Cherniak:
It’s a real‑world proof point. We here at MAM are using Anthropic’s AI tool, Claude. I believe we have about 500 licenses. Across the platform, I think we’ve tried to expand that to over 1,500 and we couldn’t.
So, there’s clear capacity constraints, and I think it’s clear that we’re still in the very early innings of both the adoption and the ability to really lean into this AI technology to complement, not substitute, the assets that we have and the processes that we’re implementing to make investments. I think this is one of the fastest‑growing enterprise software companies in history, and they’re supply‑constrained, not demand‑constrained.
And it’s really that physical infrastructure —around power, data centres, around cooling—and as you know, that’s exactly what we’re focused on, on the infrastructure credit side of the business.
Dan, your team has done some real work on whether we are in a bubble. I think the audience would love to hear your verdict, whether we think about it as a bubble, or whether we don’t.
Daniel McCormack:
Yeah, there’s a lot of concern out there in the market that we are indeed in an AI investment bubble. And you’re right, I mean, we did some work comparing what’s going on now with what happened in the mid to late 1990s with the internet and IT revolution that took place then. And, in short, our conclusion from all of that was we may well be only in the fairly early stages of this investment boom.
So, it’s a little bit complicated because US investment has undergone structural change. Back in the 90s, for example, tech investment was single‑digit percentages of non‑residential investment. Now it’s about 30 to 35%.
Tech has become a much bigger component of the overall investment picture, so its ability to move the macro needle is much more significant, and so its ability to garner headlines is commensurately much more significant.
But if you look at growth rates, for example, it looks like we’re in late 1997, early 1998, in terms of tech spending. And that applies too if you just track tech investment in level terms, right?
So, if you index to 100, starting when ChatGPT was launched, and you index to 100 at the start of 1993, which is when that investment boom really took off, what you find is you’re in 1997, 1998 territory. And the boom went all the way to 2000, depending upon exactly how you measure at the end of 2000. I would layer on top of that that I think AI is probably more powerful than the internet and the technology revolution that happened back then. It’s more powerful for productivity. It could be more powerful for corporate profitability.
And it justifies that investment even more so than the technology that we had last time. Look, it always pays to be a little bit cautious when you have an aggressive run‑up in investment like we’ve had recently. But if we compare to the 1990s period, to use a baseball analogy, it’s the third or fourth innings here.
Harlan Cherniak:
Yeah, I think some of the statistics are staggering. The capex commitments that have been announced from some of the largest hyperscalers aggregate, close to $600 to $700 billion for 2026 alone. Amazon at $200 billion. Google at $175 billion.
It’s remarkable when you think about the AI infrastructure —the data centres, the GPUs, the networking, the power that’s required. And that’s on top of the cloud‑based computing and the data storage required to fuel, as we’ve talked about, the internet of things.
And when you think about these growth rates, between the last fundraise in December 2024 and the most recent fundraise that Anthropic raised, over a 15‑month period their ARR grew from $1 billion to $19 billion. I can’t think of another precedent in enterprise software history that’s experienced that type of explosive proliferation and just the sheer magnitude of the dollars that that’s contributing to the ecosystem for infrastructure.
Daniel McCormack:
Yeah, absolutely. I’ve been working as an economist and investment strategist for almost 30 years now. And economists, we always look for structural trends, right? We love a structural trend. And I would have to say that this digitisation trend is probably the most compelling structural trend I’ve seen in my entire career.
You talk about some of those technologies that are coming over the horizon. They’re going to drive data growth, in addition to AI and cloud computing. It’s a very reliable kind of trend that we are facing here. So, I think you can have a high degree of confidence in terms of where we’re going. And it’s pretty exciting, really.
Let’s pivot a little bit. We talked about the problems in private credit. We’ve talked about the disruption coming from AI. Could you just tell me where you think infrastructure debt sits, within that broad spectrum?
Harlan Cherniak:
I think what would be really helpful is to make infrastructure tangible. The way we think about it, infrastructure is financing the essential assets and the essential services that everybody needs from the time they wake up at 5:00 a.m. to the time that they plug into their morning Zoom at 8:00.
At 5:00 a.m., my alarm goes off, fueled by the regulated electricity grid. Take the dog for a walk, throw out the trash, hop on your Peloton, take a shower, log into your Starbucks app to pick up your coffee on the way to work. Log into Spotify to listen to some music or talk to Siri on the way to the train. Or the subway. Thinking about those that drive to work, they may have charged their Tesla or any other electric vehicle overnight. They get to the office and they swipe into their smart access controls or the elevators, and then you log into Zoom for your morning meeting.
And if you think about all of the essential assets and all of the essential services that are required to fuel that morning, these are cell towers that have long‑term tenancy with AT&T, Verizon, or T‑Mobile. Thirty‑four million Starbucks rewards members, and those sales that are flowing through those apps. The 5G towers, the fibre backhaul, and every node across the infrastructure that’s needed. I think there’s almost 751 million Spotify users that are consuming data through Google Cloud. Data centres powered by renewable energy power purchase agreements, and all of the superchargers that are connected globally. This is just a great example of digital, physical infrastructure at work.
The way we like to think about infrastructure is it used to be defined by roads, bridges, and tunnels, water, gas, and electric utilities. The utilities of today are digital and invisible, and they represent some of the most durable infrastructure opportunities in recent history.
Daniel McCormack:
That’s a great story of what exactly infrastructure is and, hopefully, that really brings it to life for the audience.
In my team, we, we’ve done a lot of work looking at infrastructure equity and what it tends to offer investors is kind of two things: One, defensiveness. So, when GDP growth is weak or you go into a recession, infrastructure still gives you a very respectable return, while other equity can fall away in terms of its performance.
And then secondly, its inflation‑hedge aspects. And this comes through a range of channels. It comes through regulation. It comes through pricing power courtesy of a monopoly or a monopolistic position. It can come through contracts as well.
Infrastructure just has this ability to pass on cost increases much more reliably than other equity does, and that’s the foundation of that inflation‑hedge aspect.
The world that we’re in right now is pretty exciting in many ways, but it’s pretty volatile. We’ve talked about AI, but we’ve got a lot of geopolitical and economic volatility and uncertainty as well. So, having something in your portfolio that is defensive and offers a bit of inflation hedge, I think is a good thing.
But how, how do those traits that I just talked about, how does that apply to infrastructure debt? How does it feed in? How should investors think about infrastructure debt in that context?
Harlan Cherniak:
Dan, you and I have talked about this. I think a word that I’d like to throw out there, or an acronym that I think is really appropriate, that’s getting a lot of attention, is the HALO effect around infrastructure. And by that I mean hard asset, low obsolescence. We’ve started to talk about some of these attributes that you’re seeing for infrastructure resiliency.
Equity investors are focused on control. Our colleagues can control the board, management teams, capital allocation, and the alignment through compensation.
On the infrastructure credit side, what we’re really focused on is the asymmetry in the return profile. Credit agreements and covenants. And it’s these covenants that are so well structured that allow us to both monitor project progress and use those covenants as the early warning and monitoring or flashing signals when there’s a challenge. And it’s the balance of those two that allows us to potentially earn these very attractive risk‑adjusted returns in this environment.
From both a macro and from a micro perspective, infrastructure credit is a rather inefficient market. And it’s benefiting from some of the same trends that we’ve seen across the broader private credit or direct lending corporate credit landscape since the emergence from the GFC 17 years ago. One of my colleagues likes to say that infrastructure debt today is where direct lending was 17 years ago coming out of the GFC. And we’re benefiting from the continued retrenchment of the investment banks, whether it be project financing or opportunistic financing, as they’ve all approached certain regulatory or risk‑based capital thresholds that has forced them to either pull back and or find strategic partners in the space.
Daniel McCormack:
Harlan, perhaps we could just spend a few minutes now on the BDC landscape because there’s a lot of headlines out there about redemptions. If you turn on CNBC or Bloomberg, you’ll certainly hear about that. It would be great to separate the theatre, or what we see on TV, from the reality. From your perspective, what’s actually happening here?
Harlan Cherniak:
To be clear, we’re very bullish on credit, both across traditional direct lending or corporate credit and infrastructure debt. But you have to look at the scale first. Today there’s north of 165 BDCs. They represent about $500 billion of AUM. So, there’s been significant growth over the last couple of years, both across the institutional channel as well as the retail channel and, in this environment, repricing its risk appetite is what you’re seeing through some of the data today. It’s not necessarily a specific credit event. It’s a bit more of a sentiment event, or some of the fear‑mongering that you’re hearing about in the theatre.
The gross numbers are dramatic in terms of the repurchase, the tender, or the redemption requests. But you have to think about the BDC market in terms of why they’re structured this way. Most of these corporate credits have an average tenor of five years, which means 20% of the portfolio should roll off naturally every year, and thus 5% every quarter. I don’t think it’s a coincidence as to why the fund‑level gates were set at those thresholds.
But what the media is not telling you is over the last few quarters, you’ve seen significant net inflows that have offset some of those significant gross outflow figures. They’re only focused on the redemptions.
I’m not here to tell you that we don’t expect those redemptions to increase. And I’m not here to tell you that certain managers have other levers, or others have played the approach differently.
Some have put up their gates. Some have honoured pro‑rata redemptions. Others have transferred assets or sold assets to secondary market purchasers. Others have taken the approach of doing what we would do, which is to act responsibly on behalf of all of our clients as fiduciaries.
And so, while we may expect elevated redemptions, the BDC market is an opportunity. We’re pretty bullish on the BDC opportunity set. Through the diversification of both corporate credit, which as you know is an LBO M&A pro‑cyclical product, and infrastructure, which is a bit more idiosyncratic and has much more robust controls and covenants, and might be a bit more of a project‑specific or thematic‑specific opportunity set. They both offer diversification. They both offer yield and income.
And particularly for those in the retail channel, whether it’s through their 401(k) or defined benefit plans, or for those that are seeking long‑term compounding opportunities, these are highly qualified managers in both private equity and private credit, with the ability to compound those returns over time.
Daniel McCormack:
That’s great, Harlan, thank you. I do have a couple of questions left, and one I definitely wanted to ask you on this podcast is the one about systemic risk in private credit. I remember the global financial crisis. There were a lot of complicated leveraged products that really unwound through that period. You might be worried that some of the cockroaches that we’re seeing in private credit is a leading indicator of systemic risk. What’s your view on this? Is there real systemic risk here or is it overstated?
Harlan Cherniak:
It’s a great question, and one that we both debate internally and with our sophisticated investor base regularly. The word that I would use to describe what we’re seeing, or the view from the Macquarie Asset Management perspective around credit is orderly.
While there’s some redemption stress in the BDC space, and it’s real, and there’s likely to experience some speed bumps – whether it’s maturity wall or whether it’s from an underlying credit or default cycle perspective – we don’t believe it’s systemic.
If it starts with the BDCs, the 1940 Act and the regulations that provide the guardrails for investors are designed to prevent that. They’re capped in terms of how much leverage they can have in the system.
The SEC requires that, whether you’re a publicly traded BDC or a perpetually private BDC, every quarter you have to report every single asset on a schedule of investments with footnotes that detail payment in kind, non‑accrual, eligible portfolio company, or RIC status.
And so, there’s an enormous amount of data out there that both shows the diversification within an individual BDC but also talks about the correlation across all the BDCs given how much club financing or these private credit financings have driven an opportunity for multiple funds or multiple managers to co‑invest or to partner together. And so, overall, I think we’ll navigate through this quite well.
The tools exist: The gating technology, secondary sales, co‑investments. And more importantly, the underlying loans, they continue to perform. You have to remember that even these software credits will not be displaced overnight. There may be some pressure on terminal values or multiples. And you have to go credit by credit, manager by manager.
On the infrastructure credit side of the equation, we really believe that the illiquidity of the infrastructure credit market is an opportunity. When we look at the dissection of returns between base rate, plus credit spread, plus complexity, plus illiquidity premium, and some of the incremental fees, we believe there’s an attractive risk‑adjusted return profile embedded there in the infrastructure credit markets.
And if you look over the long‑standing Moody’s, Cliffwater, S&P data around the probability of default and the severity of loss, the historical default rates and the historical loss rates for infrastructure credit are very attractive and provide investors with shelter from some of the storms that are out there today.
And so, is it systemic risk? Is it real or is it overstated? Our view continues to be that indirect lending and infrastructure credit, that there’s very limited systemic risk at this opportunity. And there could be inefficiencies or mispriced risk out there in the market.
Daniel McCormack:
Great, Harlan. What I’ve heard today, one of the key points, is private credit is not a monolith. There’s a lot of different components to it. We talked about infrastructure assets and all the different types that are there.
Just to bring it together, what would be the two or three key takeaways that you would want listeners to walk away from today’s podcast with?
Harlan Cherniak:
We’re constructive on credit across both direct lending and infrastructure credit. That being said, you have to know what you own.
There is real stress in the market, and it requires you to re‑underwrite a handful of scenarios. And while some of the risk is concentrated in specific sectors, specific structures, or specific vintages, there are opportunities. We’re very bullish on infrastructure debt due to the essential services, the essential assets that they’re financing. The collateral is real. The contracts are durable. And we think the pricing is quite attractive. They’re physical, hard bricks and sticks. And we think they have a long, both useful and depreciable life irrespective of how AI investment and how this cycle runs its course. So, know what you own. And if you don’t, ask for help.
Daniel McCormack:
Fantastic. Well, that’s a wrap. Thanks very much, Harlan. Some wonderful insights today. I think we touched on some asset‑specific things and we also touched on some high‑level macro things, such as, is there systemic risk here in the market. So, fantastic and hope to have you back on the podcast soon.
Harlan Cherniak:
I appreciate that, Dan. As always, I love reading your research and all the work that your team puts together.
Daniel McCormack:
Well, to our listeners, thanks very much for joining today. I hope you found that as informative and as interesting as I did. If you would like to go deeper into some of these subject matters, please reach out to your relationship manager or visit our website to download some of our research.