What $1.5B From Oaktree Actually Cost Ishbia

In early '21, Mat Ishbia told me he was certain UWM would be crowned 👑 the largest originator in America in '22. I wasn't sold. Rocket's refi machine was pumping out extraordinary numbers at the time — $351B(!) in mortgages in '21 — & seemed unstoppable.

But Ishbia was proven right. Rates skyrocketed in '22 as inflation crippled the economy & purchase became the only durable origination strategy. UWM's broker army & ops wizardry 🧙 were far better equipped for that market than Rocket's call-center-heavy 📞 operation, & UWM has held the title for nearly four years.

But even the most talented operators get punched 👊 in the nose eventually.

UWM gambled that its origination machine alone would be strong enough to outrun the advantages competitors were building while the Ishbia family collected sports teams. The failed pursuit of Two Harbors, a heavy debt load & a servicing strategy built around selling rather than holding, have left the Ishbia clan w/ very few levers to pull in a still-high-rate market.

We’ll tackle that — and the shifting competitive dynamics in wholesale — in today’s edition of The Mortgage Scoop. Plus, if you or your team are still relying on Zillow leads, I have some v bad news for you 😥…

That's the takeaway from Tuesday's earnings call. Zillow, CFO Jeremy Hofmann told analysts, has "designs to be one of the biggest purchase lenders in the country." He added: "That would not be possible in the legacy ad model."

The old business where you pay Zillow to find you borrowers…

Zillow is now finding the borrower, pre-approving the borrower, showing the borrower homes they can actually afford, & closing the loan itself via the Preferred program.

Numbers-wise, ZHL collected $84M in Q2 revenue, up 75% YoY, on $2.2B in purchase origination volume. That's up 95% from roughly $1.1B a year ago, in a quarter when the purchase market was flat. Let’s call it a $9B annual run rate. ZHL is now a top 25 purchase lender, & claims its average LO produces roughly 2x the industry average in purchase units per month (despite them not being sticky w/ agents). 

Now, Zillow isn't taking food off anyone's table yet, but it was half that 12 months ago. As I’ve previously reported, Zillow’s customer acquisition costs are a fraction of what traditional mortgage lenders pay & b/c the margins are much fatter than traditional agent advertising, they’re gonna do this shit at real scale. They kinda have to. Check this out: revenue per Preferred connection ran 23% above the old ad model in '25, & they're guiding to 35% by year-end.

All of which means your co-marketing dollars are getting less important to Zillow every quarter. They found a better way to monetize the borrower, & it happens to be the thing you do. And yeah, that sucks 😿 .

From the looks of it, Zillow is moving customers into its Preferred model pretty fast: 21% at the end of '24 —> 44% at the end of '25 —> 61% today —> 75%+ by year-end.

Asked whether ZHL is actually profitable, Hofmann told analysts that the per-unit economics are positive today across both fixed & variable costs, & that "ZHL is now at a spot where it can grow profitably." He said that the company expects mortgage economics to eventually look like what it earns on Preferred agent referral fees.

LOs — If you have lead gen MSAs w/ agent partners, you may want to ask them what their plan is here. And if they’re using Follow-Up Boss for a CRM, you might have to fight for that referral a lot harder than you used to…

A $451.9M net loss in Q2 & thinning liquidity — despite $39.7B in originations & a 133 bps margin — led Ishbia to the doorstep of Oaktree Capital's Global Opportunities Group, the distressed/opportunistic shop under Brookfield's flag.

Look, these are not the first guys you call when you need money. One source described them to me as "very fucking serious private credit/distressed dudes." They are excellent at what they do, & what they do is price desperation quite accurately.

I’ve got fresh details on how the competitive landscape in wholesale is shifting for paid subscribers, as well as some questions about the $603M hedging mistake 👇