Thank you for the article. You do an excellent job integrating the Gotham and Hindenburg work into your own. PIK to cash interest is a big hurdle and the $1 billion Drive Time subsidy is a smoking gun. The one thing I have doubts about (and this is just vibes, I don't have expertise here) is that Ally may stop purchasing loans. Unless the subprime auto loan market totally collapses, I think Ally will continue to buy loans. I think they are just as addicted to the securitization flow of funds as the Garcias are.
Thanks. You're probably right that Ally doesn't walk away entirely, that was quite binary. The more likely scenario is they keep buying but pay less for the paper. They tighten terms, take the better credits, compress the premium. That's actually worse for the thesis than a clean break because it bleeds out gain-on-sale margins quietly with no obvious catalyst for the market to reprice around.
On the mutual addiction point, agreed, but that's kind of the problem. Everyone's incentives are aligned to keep the machine running until they aren't. The question isn't whether Ally keeps buying, it's what they pay. If that ~9% premium becomes 6% because delinquencies keep climbing, a third of Carvana's clean profit just vanishes.
Thanks for the detailed financial analysis, it’s genuinely strong work. One thing I’m not seeing anyone address though, is the biggest risk to Carvana’s model: the fact that buyer demand for the riskiest ABS tranches is drying up at the same time BLAST and N‑series portfolios are deteriorating. Losses are climbing the waterfall, Overcollateralization (OC) is being eaten away, and credit enhancement is shrinking fast.
With private credit already struggling on performance, it’s not clear who is supposed to keep funding these structures. The first real test was when Ally pulled funding, and the model failed immediately. Nothing has changed to prevent that from happening again.
A deeper look at BLAST and the N‑tranches — OC erosion, cumulative loss progression, and how close they are to cash‑diversion triggers — would add a lot to the conversation.
Incredibly detailed analysis. Great job!
Thank you!
“Pay no attention to the man behind the curtain…”
Great article
Thank you!
Thank you for the article. You do an excellent job integrating the Gotham and Hindenburg work into your own. PIK to cash interest is a big hurdle and the $1 billion Drive Time subsidy is a smoking gun. The one thing I have doubts about (and this is just vibes, I don't have expertise here) is that Ally may stop purchasing loans. Unless the subprime auto loan market totally collapses, I think Ally will continue to buy loans. I think they are just as addicted to the securitization flow of funds as the Garcias are.
Thanks. You're probably right that Ally doesn't walk away entirely, that was quite binary. The more likely scenario is they keep buying but pay less for the paper. They tighten terms, take the better credits, compress the premium. That's actually worse for the thesis than a clean break because it bleeds out gain-on-sale margins quietly with no obvious catalyst for the market to reprice around.
On the mutual addiction point, agreed, but that's kind of the problem. Everyone's incentives are aligned to keep the machine running until they aren't. The question isn't whether Ally keeps buying, it's what they pay. If that ~9% premium becomes 6% because delinquencies keep climbing, a third of Carvana's clean profit just vanishes.
You're an excellent writer, by the way. Nice work.
Thank you!
Thanks for the detailed financial analysis, it’s genuinely strong work. One thing I’m not seeing anyone address though, is the biggest risk to Carvana’s model: the fact that buyer demand for the riskiest ABS tranches is drying up at the same time BLAST and N‑series portfolios are deteriorating. Losses are climbing the waterfall, Overcollateralization (OC) is being eaten away, and credit enhancement is shrinking fast.
With private credit already struggling on performance, it’s not clear who is supposed to keep funding these structures. The first real test was when Ally pulled funding, and the model failed immediately. Nothing has changed to prevent that from happening again.
A deeper look at BLAST and the N‑tranches — OC erosion, cumulative loss progression, and how close they are to cash‑diversion triggers — would add a lot to the conversation.