In Round #46: KKR on Trading Loans; Apollo & Blackstone's $36B Loan; Loans to PE-Backed Companies in Decline; CDS on Hyperscalers; Defaults Reexamined; Faster Runs on Banks; Notable Deals; Bulletin Board.
In the long term, Google will eat Nvidia's cake with it's TPUs. The are specialized for matrix multiplication which is was AI and ML algorithms are all about under the hood.
Excellent content. So many worthwhile threads here, but I will focus on a few points.
1. Your interview with Ron Kahn at Lincoln was top notch. Many podcasts have a low takeaway ratio. This one was choc a bloc with nuggets. And you did an excellent job of keeping it on track and pinning Ron down on the critical questions. It was a quality look into the valuation process. It made me slightly more comfortable (especially the part where he said the client nearly always uses Lincoln's valuations - albeit a range is provided and Lincoln allows input from the client - so somewhat more comfortable, but there is room for games).
2. A lot of Athene's growth has been funded by Funding Agreements lately, so they have become less reliant on annuities. This is hot money that can turn on a dime if Athene runs into trouble. It may also present an A/L mismatch risk to the extent the FAs roll off faster than the assets.
3. The Broadcom backstop on the $36B Apollo/Deal. As I have written about on the Meta data center financing, the rating agencies are allowing the guarantee to influence the rating of the data center debt, yet ignore the impact of the guarantee on the rating of the guarantor (eg Meta, Broadcom). I'd be curious to see how they justify this in the Broadcom case.
And yes, the only easy way Apollo can clear that kind of debt is using Athene's balance sheet. How else can they compete with Wall Street's distribution engine? At some point the concept of Glass-Steagall will be applied to insurance companies.
Though the debt may be disappeared in one of Apollo's nesting doll vehicles like AMAPS or the Fox Hedge LP. (I wrote about these recently).
4. I have a BDC tool (which I won't share here). It's currently free. DM me if interested. A lot of time spent developing additional data points and normalizing the data.
5. Regarding CDS being used to hedge AI debt - on its face, this looks like a tool that is serving its intended purpose. What NOBODY is talking about, is the regime shift in how CDS are cleared since the GFC. Trillions of dollars of risk are being transferred and nobody is talking about how these transactions are cleared and settled and risk managed.
1. Appreciate the kind words. I also agree on the game part. I didn't want to push harder because PR teams might hold up the release of the episode if they felt something came across too unfavorably. I also liked what I heard, although I wasn't a fan of the idea that a private credit firm can come back with "additional information" after seeing the marks. But it's all fair.
2. I thought FABNs typically have 3–5-year maturities, with only a handful of 1-year deals. Isn't that the case?
3. So, in theory, META could guarantee an unlimited number of deals and it still wouldn't affect its rating?
4. Is it LoanEdge? If not, I'll send you a DM.
5. Is it common for investors to buy CDS protection without actually holding the underlying position?
It's my own BDC platform, not Loan Edge. I can drop the link here or feel free to DM me. Don't want spam your comments with my own website. We are in the process of converting all footnotes to fields (eg pledged assets, non-qualified assets, affiliates).
Yes, the FABNs generally are 3-5 years is my understanding as well, but they also have FABN CP and also I believe longer term. Given the length of some of their underlying investments, 3-5 years isn't long enough to match asset duration. But on my list is to more precisely quantify the A/L mismatch. I have Athene's entire debt holdings downloaded and need to match to FABN schedule (and annuity).
I think you're right on CDS, though I'm not aware of the data to confirm. To the extent they hold CDS without the underlying it's either a speculative position or a more macro hedge. Either way, my point is the focus on the central clearing process and whether margin process is fit for purpose.
And yes on the Meta guarantee, though maybe S&P has a line where they may impose an impact. $30B should have been enough to cross that line.
Thanks. Spent a lot of time normalizing the data across time and BDCs and creating new value added fields, most recently extracting more footnotes into actual fields (not released yet to production).
Interesting to see the high % of assets that are pledged and the type of assets that aren't pledged.
Not sure, it wasn’t discussed. Maybe that B tranche is some combination of mezz/equity, or Anthropic itself holds equity? Unlike Meta’s data centers, where holding the equity tranche is actually a solid idea given its strength, Anthropic is technically a cash-burning machine, so it would kind of make sense to have them put equity in. But those are just my guesses.
In the long term, Google will eat Nvidia's cake with it's TPUs. The are specialized for matrix multiplication which is was AI and ML algorithms are all about under the hood.
Excellent content. So many worthwhile threads here, but I will focus on a few points.
1. Your interview with Ron Kahn at Lincoln was top notch. Many podcasts have a low takeaway ratio. This one was choc a bloc with nuggets. And you did an excellent job of keeping it on track and pinning Ron down on the critical questions. It was a quality look into the valuation process. It made me slightly more comfortable (especially the part where he said the client nearly always uses Lincoln's valuations - albeit a range is provided and Lincoln allows input from the client - so somewhat more comfortable, but there is room for games).
2. A lot of Athene's growth has been funded by Funding Agreements lately, so they have become less reliant on annuities. This is hot money that can turn on a dime if Athene runs into trouble. It may also present an A/L mismatch risk to the extent the FAs roll off faster than the assets.
3. The Broadcom backstop on the $36B Apollo/Deal. As I have written about on the Meta data center financing, the rating agencies are allowing the guarantee to influence the rating of the data center debt, yet ignore the impact of the guarantee on the rating of the guarantor (eg Meta, Broadcom). I'd be curious to see how they justify this in the Broadcom case.
And yes, the only easy way Apollo can clear that kind of debt is using Athene's balance sheet. How else can they compete with Wall Street's distribution engine? At some point the concept of Glass-Steagall will be applied to insurance companies.
Though the debt may be disappeared in one of Apollo's nesting doll vehicles like AMAPS or the Fox Hedge LP. (I wrote about these recently).
4. I have a BDC tool (which I won't share here). It's currently free. DM me if interested. A lot of time spent developing additional data points and normalizing the data.
5. Regarding CDS being used to hedge AI debt - on its face, this looks like a tool that is serving its intended purpose. What NOBODY is talking about, is the regime shift in how CDS are cleared since the GFC. Trillions of dollars of risk are being transferred and nobody is talking about how these transactions are cleared and settled and risk managed.
Now that's how you leave a comment! :)
1. Appreciate the kind words. I also agree on the game part. I didn't want to push harder because PR teams might hold up the release of the episode if they felt something came across too unfavorably. I also liked what I heard, although I wasn't a fan of the idea that a private credit firm can come back with "additional information" after seeing the marks. But it's all fair.
2. I thought FABNs typically have 3–5-year maturities, with only a handful of 1-year deals. Isn't that the case?
3. So, in theory, META could guarantee an unlimited number of deals and it still wouldn't affect its rating?
4. Is it LoanEdge? If not, I'll send you a DM.
5. Is it common for investors to buy CDS protection without actually holding the underlying position?
It's my own BDC platform, not Loan Edge. I can drop the link here or feel free to DM me. Don't want spam your comments with my own website. We are in the process of converting all footnotes to fields (eg pledged assets, non-qualified assets, affiliates).
Yes, the FABNs generally are 3-5 years is my understanding as well, but they also have FABN CP and also I believe longer term. Given the length of some of their underlying investments, 3-5 years isn't long enough to match asset duration. But on my list is to more precisely quantify the A/L mismatch. I have Athene's entire debt holdings downloaded and need to match to FABN schedule (and annuity).
I think you're right on CDS, though I'm not aware of the data to confirm. To the extent they hold CDS without the underlying it's either a speculative position or a more macro hedge. Either way, my point is the focus on the central clearing process and whether margin process is fit for purpose.
And yes on the Meta guarantee, though maybe S&P has a line where they may impose an impact. $30B should have been enough to cross that line.
Good to know all this. Please drop the link here, in case if other readers want to try it out as well. Don't worry about the spam.
Here you go. Some new features are in the staging version (e.g. footnote extracted categories).
https://www.tpehub.com/bdc-intel
You guys really put a lot of work into this!
Thanks. Spent a lot of time normalizing the data across time and BDCs and creating new value added fields, most recently extracting more footnotes into actual fields (not released yet to production).
Interesting to see the high % of assets that are pledged and the type of assets that aren't pledged.
Certainly Micheal Burry would answer yes to #5. It's a viable way to bet against the credit.
Lol
Who is funding equity tranche in the Anthropic deal?
Not sure, it wasn’t discussed. Maybe that B tranche is some combination of mezz/equity, or Anthropic itself holds equity? Unlike Meta’s data centers, where holding the equity tranche is actually a solid idea given its strength, Anthropic is technically a cash-burning machine, so it would kind of make sense to have them put equity in. But those are just my guesses.