Liam Dalton, CIO of Willow Lake Strategic Advisors, joins me for Episode 4 of Mispriced Assets.

We talk about how Liam approaches markets through macro, fundamentals and technicals, why AI has become the dominant force in the current cycle, and where the next leg of the trade could come from.

Most of the conversation centers on the AI infrastructure buildout: the memory shortage, Micron, SK Hynix, Samsung, semiconductor equipment, hyperscaler spending, and the game theory around who finally breaks discipline and builds more capacity. We also get into why the next major opportunity could sit further upstream in the companies supplying the tools required to build that capacity.

Then we turn to the risks: rates, credit, energy, the real economy, politics, and whether one of the largest earnings booms in market history eventually has to be followed by an earnings bust.

Disclaimer

This podcast is provided for informational and educational purposes only. Nothing discussed constitutes accounting, legal, tax, or investment advice, a recommendation regarding any security or investment strategy, or an offer or solicitation to buy or sell any security or financial instrument.

Investing involves risk, including the possible loss of principal. Any securities, companies, industries, trades, or investment strategies discussed are included for illustrative and discussion purposes only and should not be considered recommendations. Past performance is not indicative of future results, and no representation is made that any investment or strategy discussed will be profitable.

The views and opinions expressed are those of the individual speaker and may change at any time without notice. They do not necessarily represent the views of Mispriced Assets, Willow Lake Strategic Advisors, or any affiliated entity unless specifically stated otherwise.

The hosts, guests, their employers, affiliated entities, clients, or related accounts may hold, buy, sell, or otherwise transact in securities or financial instruments discussed in this podcast, before or after publication, and may have economic interests that differ from those of viewers.

Information discussed may come from public sources or other sources believed to be reliable, but its accuracy, completeness, and timeliness are not guaranteed. No warranty or representation is made regarding the information presented. Viewers should conduct their own research and consult their own financial, legal, accounting, or tax advisers before making any investment decision.

Groupon CEO Dušan Šenkypl and CFO Rana Kashyap join me to talk about taking over a company they weren’t sure would survive—and what they’re building now.

We start with their backgrounds, why they stepped into the turnaround, and what it took to get the business growing again. Then we get into where they want to take it: an AI-native marketplace for local experiences, with Groupon’s deals available wherever customers are searching.

A lot of this conversation is about AI from the perspective of people actually implementing it. We cover Project Foundry, the company’s data infrastructure, AI assistants, and how smaller teams are changing the way Groupon operates.

We also spend time on the investment case: rebuilding customer trust, improving recommendations and booking, reaching younger customers, and making marketing dollars go further. The part I’m most interested in is getting a customer’s first purchase to become a second, third, and fourth—and what that could mean for earnings and free cash flow.

Disclosure: I own Groupon shares and am bullish on the company. This conversation is not investment advice.

Recorded September 10, 2026. This was not an earnings call or an update to financial guidance. Management’s forward-looking statements reflect its views on that date; actual results may differ.

Aznaur Midov, author of DEBT SERIOUS, joins me to talk about where private credit is actually starting to crack — and where the bearish case may be getting ahead of itself.

Aznaur spent years lending to sponsor-backed software companies, so we spend a lot of time on the part of private credit I’m most worried about: software. We get into AI disruption, the 2027–2028 maturity wall, amend-and-extends, recoveries, CLOs, BDCs, leverage and what happens when lenders and private equity sponsors eventually stop being on the same side.

We agree on a lot, but differ on how ugly the next few years get. Aznaur thinks sponsors and lenders have more ability to work through the maturity wall than I do. I think AI creates enough uncertainty in software that a lot of credits simply won’t be refinanceable at anything close to today’s marks.

A good debate with someone who actually underwrote this stuff before everyone decided private credit was an asset class you had to own.

In 1989 a federal regulator let a savings and loan count a certificate as capital. It was worth nothing, and everyone who signed it knew it. A few years later the same regulator declared it worthless, and overnight the thrift went from well-capitalized to one the government could padlock on any morning. Rod Dubitsky was inside it. He called the certificate phantom capital.

He is watching the same instrument today, on the balance sheets of the life insurers behind millions of American annuities, at something like a hundred times the size. That thread runs through the whole conversation, and it is where we open the show.

Rod is the closest thing markets have to a control group. His first day on a mortgage desk was Black Monday, October 1987. He worked the savings-and-loan wreckage from inside the regulator. He was at Moody’s when the instrument that would take down the world was invented two desks away, and he was the Credit Suisse analyst whose downgrade call ran a year ahead of the agencies. Same method each time: build the data nobody else would, then take apart the model everyone else trusted on faith.

We walk his whole career and land on what he and I are both staring at now, private credit, the insurers holding it, and the manufactured capital underneath.

A FEW THINGS YOU’LL HEAR

* The mortgage desk that opened the same morning the market crashed.

* The rating agency where he watched single-B bonds turn into AAA on a different floor, for five times the fee.

* The line he has repeated for twenty years: the absence of loss is not the absence of risk.

* The meeting where a trader begged him not to scare AIG, and he did.

* Twelve bonds rated AAA while the loans underneath them had already defaulted.

* Why he thinks there are no A-rated insurers left, only B through F.

* Why this one is bigger than 2008, and why being early feels exactly like being wrong.

Late in the conversation he says the people running this would have bought the Dodgers with your money if they could. I just wrote about two insurers that did.

Watch the whole thing. Ideally right before bed, then send it to anyone who needs to sleep well.

❧ ❧ ❧

For those who prefer YouTube:

A companion to the first episode of the Mispriced Assets Podcast, in conversation with Rod Dubitsky. The judgments here about particular insurers and rating agencies are often shared but our own respectively, drawn from the episode and from our published research; the firms named would contest them and are entitled to. Dubitsky writes at The People’s Economist. None of this is investment advice.