Original episode:https://youtu.be/Q6PUFCBbiBU?si=KOdpOqkjrlfEEGuz · Timestamps are clickable — they seek the player in place
Steve Eisman and Chris Edson, Apollo's Global Head of Originations, discussed the current state of the private credit (Private Credit) market, potential risks, and Apollo's business strategy [00:41]. Edson clarified that private credit is not limited to direct lending for leveraged buyouts (LBOs), but is a much larger and more diverse $40 trillion market [03:03]. Addressing market concerns about the software industry being impacted by AI and affecting debt repayment, he pointed out that Apollo has extremely low software exposure due to its adherence to underwriting standards based on cash flow and value anchoring [09:59]. Edson detailed Apollo's business model of building a diversified ecosystem through 16 origination platforms to provide customized liquidity solutions for large, highly-rated enterprises in exchange for a liquidity premium [12:53], and responded to questions about conflicts of interest between Apollo and its insurance company Athen, emphasizing that Apollo's asset-heavy model deeply aligns its interests with the funding side [31:51].
[00:00] - [02:08] Host Steve Eisman introduces the topic, pointing out that media's negative coverage of private credit focuses on software exposure and redemption limits on retail funds, and introduces guest Chris Edson [00:41].[02:08] - [04:24] Edson defines private credit: the LBO direct lending that the media focuses on is only a $1-2 trillion subset, while overall private credit is a $40 trillion market that includes mortgages, infrastructure, equipment, and aviation finance [03:03].[04:24] - [08:52] Discussing the potential impact of AI on software company valuations and debt repayment [05:03]. Edson believes that software lacking barriers and proprietary data faces risks, but core ledger software deeply embedded in ERP systems remains extremely difficult to replace [06:34].[08:52] - [11:40] Deconstructing Apollo's extremely low software exposure (only a few percentage points of AUM) and the value-oriented underwriting logic behind it (financing based on cash flow rather than enterprise valuation) [09:59].[11:40] - [16:45] Reviewing the historical drivers behind Apollo's formation of 16 origination platforms: in the post-GFC era, credit spreads in the public bond market were extremely narrow (no excess return) [13:31], prompting Apollo to establish a direct lending ecosystem to obtain stable yields [14:37].[16:45] - [19:59] Explaining the underlying logic of the direct lending model: giving up asset liquidity in exchange for an excess spread of about 150 basis points [16:27], and providing customized capital solutions for enterprises [18:04].[19:59] - [24:47] Introducing Apollo's top three largest origination platforms: Atlas (a $50 billion securitized warehouse financing platform acquired from Credit Suisse) [20:42], Wheels (a fleet leasing platform managing over one million vehicles) [22:24], and PK Airfinance (an aviation direct lending platform) [24:09].[24:47] - [27:52] Explaining cross-platform synergies: breaking down departmental barriers of traditional asset management companies, and unifying coordination across the 16 platforms through industry experts to achieve cross-selling [25:32].[27:52] - [30:56] Discussing credit cycles and underwriting standards. Edson points out that because the private direct lending market involves bilateral negotiations, its due diligence and collateral covenants are often stricter than those in the public market [28:34].[30:56] - [33:45] Responding to questions about conflicts of interest between Apollo and Athen (an annuity insurance company): Apollo is not an asset-light model, but has invested $35 billion of corporate capital in Athen, sharing first-loss risk, which highly aligns their interests [31:51].[33:45] - [36:54] Warning that the core of the credit cycle lies in liquidity risk (asset-liability maturity mismatch) [33:28], and introducing the extreme stress testing Apollo conducts on its origination platforms [34:09].[36:54] - [38:59] Discussing the redemption mechanism of private direct lending retail funds, explaining that limiting the redemption ratio (usually to 5%) is to protect collective interests and fulfill established contracts [37:11].[38:59] - [41:44] Clarifying the transparency issue of private credit, pointing out that regulatory filings of Athen and others, as well as BDC holding reports, fully list the details of all loans [39:12].[41:44] - [44:39] Steve Eisman summarizes: Apollo's asset-heavy direct lending model, low software exposure, and high transparency enable it to defend well against the upcoming credit cycle [42:20].[03:03] - [03:44]
- Type: Fact[06:34] - [07:23]
- Type: Opinion[09:59] - [11:13]
- Type: Fact[16:27] - [19:53]
- Type: Opinion[33:28] - [34:04]
- Type: Opinion
- Note: This viewpoint was highly endorsed by Eisman.[28:34] - [29:15]
- Type: Fact[31:51] - [32:59]
- Type: Fact[39:12] - [40:18]
- Type: Fact[08:52] - [09:19] vs [42:56] - [43:33]: In the interview, Chris Edson tries hard to downplay the direct systemic destructive power of AI on the software industry, stating that this process requires a long cycle and validation through application implementation; however, in his personal summary in the postscript, Steve Eisman directly reiterates that the credit cycle will ultimately detonate first in the software industry, and that the exposed risk and bad debts will be far more violent than what Edson described, creating a tension in judgment between a veteran investor and an industry practitioner.
So let's talk about the viewpoints brought by Chris Edson. Everyone feels like private credit is currently a powder keg ready to explode at any moment, mainly because newspapers are writing every day about high-valuation software companies about to default, or some wealth management product restricting user redemptions. But the reality is that private credit is not at all the narrow gate of "specifically issuing high-interest loans to junk enterprises" that everyone imagines.
First of all, the scope of private credit is extremely broad. Mortgages for ordinary people, airlines leasing airplanes, enterprises buying truck fleets, money borrowed to build bridges and roads—as long as it hasn't gone to the public market to issue bonds, it counts as private credit. This is a massive $40 trillion market. The "software explosion" that everyone is worried about actually accounts for only a small fraction of it.
Why is everyone so worried about software? Because with AI like Claude, the barrier to writing code has plummeted, and many SaaS companies that were previously treated as myths might see their valuations shrink, making them unable to repay their debts. Edson believes this concern is reasonable, but it depends on the software's "moat." If a piece of software has no unique data, is easy to install, and has little industry regulation, it is indeed in danger; but if this software is a company's accounting system (ERP), or even runs on old mainframes from decades ago, that company will absolutely not replace the system easily. Because the risk of changing systems is too great and the cost is too high, their debt remains very safe.
The reason Apollo didn't get stuck in this mud is that they are relatively "traditional," preferring to lend money to companies with real assets and stable cash flows, rather than painting grand illusions for software companies based on illusory "enterprise valuations."
So, since enterprises can issue bonds in the public market, why do they still look to Apollo to borrow this kind of illiquid money, and even willingly pay an extra 1.5% in interest? The answer is "customized solutions" and "avoiding volatility." For example, if an enterprise needs to spend 5 billion on equipment in batches over several years, issuing bonds in the public market means they have to take all the money at once, which is very costly; whereas with Apollo, they can agree on a draw-down-as-needed structure like a line of credit. Or, when geopolitical conflicts or banking crises break out, the public market might shut down entirely, while funding sources like Apollo can provide stable underlying support.
Regarding the conflict of interest that everyone is most concerned about—the "left hand to right hand" transaction (Apollo lending money to enterprises and then stuffing the loans into its own insurance company Athen)—Edson's explanation is simple: if this were a scam, we would be the first to go bankrupt. Because Apollo put $35 billion of real money (its own corporate capital) into Athen, bearing the bottom-layer risk of loss. This is not an "asset-light" third-party asset management model that makes money out of thin air; it is betting their own money on the quality of their underwriting.
[03:03] - [04:12] The segment where Edson broadens the definition of private credit. In one breath, he lists mortgages, commercial real estate, trade finance, equipment finance, etc., breaking the public's narrow perception of private credit as equivalent to "high-risk LBO loans," which is highly valuable for conceptual clarification.[05:45] - [07:23] The segment where Edson demonstrates live using Claude to write a math game for his daughter and deduces the impact of AI on software moats. This features his extremely vivid personal experience and detailed industry descriptions of ERP systems as "green screen" legacy mainframes, packed with a high volume of detailed information.[20:42] - [22:01] The acquisition story of the Atlas platform. Credit Suisse was forced to sell this core securitized warehouse business due to rising interest rates and deposit outflows, reflecting the intersection of the fragility of the traditional banking model and the trend toward asset-heavy private credit.[33:28] - [35:04] The segment deconstructing liquidity risk and maturity mismatch. Edson's "ultimate stress test" on the platforms (assuming no one lends another penny, would the enterprise go bankrupt before recovering the loans) is highly intense and serves as a classic dissection of risk management.A faithful reconstruction and plain-language retelling of the episode, generated by PodLens.
This is one source-grounded reading, not a replacement for the original. Every point is anchored to its source, so you can check it yourself — and corrections are welcome.