Analysis
Vesta raised $30 million led by Conversion Capital, with Citi Ventures and Andreessen Horowitz joining and -- notably -- customers Pennymac and New American Funding investing alongside them, according to TechCrunch. The round brings Vesta's total raised to $85 million since it was founded in 2020 by CEO Mike Yu and Devon Yang.
Vesta builds AI agents that automate much of mortgage loan origination -- the paperwork-heavy process that currently takes about 40 days and roughly $11,000 per loan to close in the U.S. Some lenders now use Vesta's agents for underwriting decisions directly, with every action and piece of reasoning logged for compliance and audit purposes; lenders remain legally responsible for the calls the agents make. A personal-assistant product for mortgage issuers, handling task execution and workflow tracking, is in development.
“A personal-assistant product for mortgage issuers, handling task execution and workflow tracking, is in development.”
Yu says revenue is up 12x year over year, though he puts Vesta's share of the mortgage-origination market at under 5% -- a reminder of how fragmented and manual this industry remains even as AI adoption accelerates inside it. Vesta competes with legacy player ICE Mortgage Technology on one side and AI-native rival Xpanse on the other, putting it in the middle of a market shifting from software-assisted to agent-automated faster than most lenders' own compliance teams are used to.
Customers investing in their own vendor is a meaningful signal distinct from the check size itself -- Pennymac and New American Funding aren't just buying Vesta's software, they're betting on its roadmap. That alignment cuts both ways: if Vesta's agents make a bad underwriting call, the lenders who invested in the company are also the ones holding the regulatory and reputational exposure, since the compliance framework puts responsibility on the lender rather than the AI vendor.
Mortgage origination has drawn a wave of AI capital this year as lenders look to compress the 40-day, $11,000 cost structure baked into most loans, and Vesta's $85 million in total funding since 2020 is modest next to some infrastructure-heavy AI bets elsewhere in fintech. The bet behind this round is narrower and more mundane than a foundation-model play: that a document-heavy, compliance-bound workflow like mortgage underwriting is exactly the kind of repetitive, rules-based process AI agents are best suited to take over first, before harder judgment calls follow.
