Goodbye, Vishal 👋 & Chase is Coming for 850 LOs

There's a reason JPMorgan Chase led w/ "$750B" when announcing its “American Dream” housing initiative. It's a wonderfully splashy 💣 headline. But once you annualize the commitment & compare it w/ what the bank is already doing... 

The real takeaway may be Chase's plan to hire 850 LOs, a move that could reshape competition in key markets if it happens quickly. Today, we run the numbers & separate the marketing spin from the meaningful. Also, the most quotable man in mortgage — Vishal Garg — is out as CEO at Better 🐬. Plus: could a rival's AI hack yours & more.

Five days after receiving a board seat, hedge fund executive Daniel Lewis was named the interim CEO of BETR in a bombshell announcement.

That means Vishal Garg — founder, CEO, transit extraordinaire & the most reliably quotable man in mortgage 🐬 for the better part of a decade — "mutually agreed" to transition out. He keeps his board seat & is supposed to help Lewis w/ an orderly handoff.

(It’s not the best timing for Garg; he was scheduled to speak at HousingWire’s AI Summit next week.)

Better bundled the news w/ preliminary Q2 results & moved its earnings call up, to Aug. 6 from Aug. 10, which pretty much never happens.

The preliminary numbers FWIW aren't a total disaster: Funded volume of $1.67B, up 45% YOY. Revenue of $54.7M, up 28%. Net loss of $30.6M, which is actually narrower than the $36.3M loss a year ago. Adjusted EBITDA of -$14M (but plz note that includes a $6.5M benefit from a TRID reserve release on loans originated before June '22). Back all that out & you're closer to -$20.5M.

That said, growing 45% & still burning $14M a quarter is…

Cost reductions are now expected to exceed $45M annualized by year-end, up from a previously announced $25M target. It might partly explain the recent exec exodus, including Misti Snow’s departure.

Based on the press release sent out, Better is pivoting to a platform model "in which partners own customer acquisition." Lewis said the company will win by manufacturing mortgages efficiently, "not by outspending competitors on customer acquisition."

Of course, Garg’s whole original pitch, built on SoftBank’s money (thanks, Masa!), was direct-to-consumer disruption. All you gotta do is cut out the LO, own the funnel, buy the lead cheaper than anyone. Easy! Lewis appears to be saying they want enterprise partners to own the borrower, & Better will be the rails/infrastructure. He also flagged HELOC expansion & interest from IMBs & brokers.

What does that mean for Ryan Grant & NEO? Will Chad Smith be sticking around? What happens Garg’s significant stock position? Can Lewis — who has an activist hedge fund background — run a mortgage tech company?

I’ll have more for paid subscribers on Wednesday, including some behind-the-scenes drama 🍿…

JPMorganChase announced this a.m. that it intends to deploy more than $750B into American housing through ‘35. This is a decade-long, kitchen-sink number: multifamily debt & equity, affordable housing financing, grants, mortgage originations, philanthropy. 

Let’s start w/ some basic math: $750B over 10 years is $75B a year. Chase says that's a nearly 40% increase over what it deployed in housing the past decade, which puts the prior run rate somewhere around $54B a year by my count.

I think it’s important to note that Chase never really defines exactly what counts. Does the $750B include correspondent purchases? Warehouse lines? CRE term debt on stabilized apartment buildings that would've gotten written anyway? If it's all of the above, the "increase" is doing some real work here, particularly b/c Chase originated $53.5B in mortgages in '25 alone, & it's already done $32B through the first half of this year. Annualize that & mortgage by itself eats most of the annual number.

*The affordable-housing piece is easier to sort out. Chase has said it extended $50B+ in debt & equity for affordable housing since '21 & supported roughly 410K units in that stretch — call it 82K units a year. The pledge is a million units over a decade, or 100K a year. That’s real growth (roughly 20%), but not a crazy moonshot or anything.

*Same deal on the consumer side. 500K homebuyers over the decade is 50K a year. Chase did just under 90K transactions in '25 per RETR— but that count includes refis, & Chase doesn't break out purchase loans. RETR data shows it did about 48K purchase loans last year, though we don’t know how many were first-time homebuyers.

All of this is to say… I 100% believe the pledge is directionally real. Chase is already the largest multifamily lender in the country & finished '25 as the No. 1 depository originator & No. 3 overall by volume. As I’ve previously reported, Chase could absolutely dominate the resi mortgage market if it wanted to. 

But IMO, doing roughly $20B more a year in housing finance is not really that big of a deal for a lender the size of JPMorgan Chase, especially given that people have the memory of Nemo when it comes to this stuff years later.

“Remember when Bank Of America came out in 2019 w/ a $7.5B special homeownership initiative?” one IMB exec chuckled. “I remember running the numbers & it came to two loans a year per loan officer.”

OK, having said that, there are some things market participants should keep an eye out for. Chase says it is adding 850 loan advisors. Sean Grzebin, who runs Chase Home Lending, told American Banker that Chase thinks it has "a good value proposition for LOs" & is targeting markets where it doesn't have the coverage it wants.

It’s not a bad pitch. Balance sheet, jumbo, portfolio product, a well-known brand, branch referral flow, & a comp structure that sticks even when biz slows will be compelling for many. Of course, comp caps (roughly $12K, per sources), slower ops, a lack of non-QM & a super heavy compliance focus are drawbacks in this market for retail/broker LOs. 

The thing about the 850 number is the release doesn't say by when. Is that 850 hires by year-end '27, or 850 spread across a decade that ends in 2035? Adding less than 100 LOs per year is meaningless. 

“A lot of banks hired loan officers to meet CRA lending requirements, how did that go?” remarked one LO.

Another exec said he expected a relatively muted impact for IMBs & broker shops.

“Many times banks just scavenge off each others LO’s,” he said. “Much like IMB lenders scavenge off each other’s LO’s.”

As for why all this is happening now, ROAD to Housing became law July 11 — no thanks to President Trump — & dozens of provisions now need implementing rules. 

Chase is the third big bank in 6 months to float a fat housing number: Citi's five-year, $60B supply blueprint (250K units) came first, then Huntington's $80B in late July, half of it earmarked for mortgage & consumer lending. 

Anyway, nine years is a long time & I’d bet dollars to donuts that everyone who came up w/ this plan will likely have a new job by 2035 anyway. So really, who’s counting?

Days after OpenAI disclosed that AI systems breached its testing environment & broke into another company, rival Anthropic disclosed that its own AI models also hacked unsuspecting companies during testing. It got me thinking: Could CCM’s AI attack Rate’s? Could UWM’s AI system go after Rocket’s? Would ICE’s robots attempt to take out Vesta’s? I asked one mortgage tech exec if it was possible.

“Anything’s possible, but it’s very unlikely,” he said. “You'd have to be running it w/ a very open-ended set of tools the way that Hugging Face does evals, you'd have to hack around the security guardrails b/c it's not like OpenAI gives the public raw models w/o guardrails in their API layer, & you'd have to really have a very high token limit you aren't watching.”

Reassuring, right? (I have no idea what any of that means 😜.)

(🙏 If you like what you’re reading, tell a fellow mortgage junkie to sign up here.)