They Grew Up in ‘Mortgage Families.’ Now These 21-Year-Olds Have Raised $4.1 Million to Put an AI Loan Officer Inside America’s Lenders.

The idea for Copperlane started with a phone call to Athan Zhang’s mother.

She had spent her career in the secondary mortgage market, on the risk side, and Zhang,  then a computer science student at Princeton, was struck by something she described: by the time a loan reaches investors, it has been flattened into a spreadsheet of numbers, and a lot gets lost along the way. To Zhang, that sounded less like a paperwork problem than a data problem.

That observation is now a company. This month, Copperlane announced $4.1 million in seed funding led by TQ Ventures, with participation from Y Combinator, US News Digital Ventures, Eight Capital and angel investors from Mercor and others. Zhang, 21, is chief executive; his co-founder, Brianna Lin, also 21, is chief operating officer. The two met through Y Combinator and bonded over a shared inheritance: both grew up in what they call “mortgage families,” with parents whose careers spanned Freddie Mac, Fannie Mae and the Federal Housing Finance Agency.

A business built for a brutal stretch in lending

Their timing is pointed. Originating a mortgage has become punishingly expensive, and, for a long stretch, unprofitable. Independent mortgage banks lost an average of $1,056 on every loan they originated in 2023, their worst showing in years, according to the Mortgage Bankers Association. The industry clawed back to a thin average profit of $443 per loan in 2024, the MBA reported, but the rebound was uneven: smaller lenders stayed in the red for a third consecutive year, and the fourth quarter slipped back into a loss. Production costs alone ran north of $11,000 per loan.

Copperlane’s wager is that a meaningful share of that cost is waste: hours spent chasing documents, re-reading bank statements and clarifying the same discrepancies on loan after loan, across the millions of mortgages Americans take out every year.

What ‘Penny’ actually does

At the center of the platform is Penny, which the company describes as an AI mortgage loan officer. Rather than the rigid, rule-based software lenders are accustomed to, Copperlane says Penny uses a general-purpose AI model to read a borrower’s file the way a junior loan officer might. At the application stage, the company says, it can scan bank statements, flag a large deposit that doesn’t square with a borrower’s income, ask the borrower about it and draft a compliant letter of explanation. This all happens before the file lands on an underwriter’s desk.

Copperlane pitches Penny as something lenders can dial up or down: a copilot that hands work back to humans, or, for those who want it, something closer to an autopilot. The company claims the assistant can compress a document review that takes four hours or more into minutes and sharply multiply a loan officer’s output. Those figures are Copperlane’s own. The startup has not disclosed how many lenders use the product or how much loan volume currently runs through it.

Founders betting on fit

What Zhang and Lin argue they have is fit. Zhang studied computer science and worked as a quant developer before starting the company; Lin studied computer science and finance at the University of Pennsylvania and traded on Wall Street. They readily concede that rivals know the mortgage business better than they do, and say they are trying to close that gap by embedding with lenders rather than studying the industry from a distance.

Their lead investor frames the opportunity in similar terms. The mortgage industry, said Schuster Tanger, co-founding partner of TQ Ventures, has been waiting for software that does more than digitize paperwork and instead “thinks through the complexity of a loan.” He called the founders’ mix of family insight and technical depth “rare.”

Lin casts the mission more broadly. Better technology for lenders, she said, “directly translates into a better experience for borrowers,” framing the company’s ambition as helping more Americans reach homeownership.

The harder questions

The bet is not without friction. AI in lending is now squarely in regulators’ sights. The Consumer Financial Protection Bureau has repeatedly warned that there is no “advanced technology” exception to consumer-finance law, reminding lenders that a black-box model must still give applicants specific reasons when they are denied credit. Bank regulators, including the Office of the Comptroller of the Currency and the Federal Reserve, have begun pressing institutions during examinations about how they use AI in higher-risk areas like lending, down to vendor controls and human oversight. And in 2025, the Massachusetts attorney general settled a fair-lending case centered on an AI underwriting model, a reminder that when AI touches a credit decision, the lender, not the vendor, is usually the one on the hook.

Copperlane says it keeps a human in the loop and is investing heavily in making Penny “safe and aligned,” with much of the new capital earmarked for engineering and the unglamorous work of integrating with lenders’ existing systems. For an industry that has spent years trying to modernize itself with mixed results, the proof will be mundane and measurable: whether files come through cleaner, whether loans close faster, and whether the defects that haunt the secondary market,  the very thing that started Zhang on this path, actually go down.

The idea for Copperlane started with a phone call to Athan Zhang’s mother.

She had spent her career in the secondary mortgage market, on the risk side, and Zhang,  then a computer science student at Princeton, was struck by something she described: by the time a loan reaches investors, it has been flattened into a spreadsheet of numbers, and a lot gets lost along the way. To Zhang, that sounded less like a paperwork problem than a data problem.

That observation is now a company. This month, Copperlane announced $4.1 million in seed funding led by TQ Ventures, with participation from Y Combinator, US News Digital Ventures, Eight Capital and angel investors from Mercor and others. Zhang, 21, is chief executive; his co-founder, Brianna Lin, also 21, is chief operating officer. The two met through Y Combinator and bonded over a shared inheritance: both grew up in what they call “mortgage families,” with parents whose careers spanned Freddie Mac, Fannie Mae and the Federal Housing Finance Agency.