AI Native software startup Vesta announced on Thursday that it has completed a round of financing of $30 million, led by Conversion Capital.
This startup uses AI agents to automate a large number of steps in the loan issuance process. The founder stated that this helps lending institutions reduce the time and cost required to process mortgage loans. Also participating in this round of investment are three customers of the company, including Pennymac and New American Funding, as well as Citi Ventures and Andreessen Horowitz.
Vesta was jointly founded in 2020 by Mike Yu and Devon Yang. Yu is the company's CEO, who told TechCrunch that now is the best time for financing, as the demand for the product has grown exponentially over the "past year." He stated that the company's revenue has increased by 12 times year-on-year, and they have helped lending institutions issue over $100 billion in loans each year. To date, Vesta has raised a total of 85 million dollars in financing.
Yu said, "Although our business is growing quite well, our market share is still less than 5%. Now is the time to expand our team, seize the market, and invest in new product lines."
These new product lines include a personal assistant for mortgage lending institutions that can help with task execution and track work processes. In the United States, it typically takes about 40 days to complete a mortgage loan, with each loan costing around $11,000. Yu says, "Most of this cost comes from labor, and one of the main bottlenecks in the timeline is waiting for someone to review your loan."
In the vision of Vesta, humans can deploy a set of AI agents to speed up task completion. Clients can decide which tasks to assign to the agents of Vesta: “Many of our clients initially have the AI agents work under manual approval, then let them handle a portion of loans independently, and gradually expand the scope of their use,” Yu explains, adding that some lending institutions are even using the AI agents of Vesta to make mortgage loan underwriting decisions.
He said that regardless of the software or AI proxies used, the company is still responsible for the underwriting decisions; Yu also added that all operations and the reasoning behind the decisions will be recorded to meet compliance requirements and facilitate audit of AI decisions.
Yu indicates that this higher level of autonomy comes from the significant progress of the AI models over the past year. Prior to this, these models were not sufficient to support the complex, multi-stage tasks involved in mortgage lending operations. Previously, the company focused mainly on building an appropriate data architecture in order to use the most advanced tools to automate mortgage processes.
Yu said, "For us, the biggest breakthrough is [Claude] Sonnet 4.5. We have found that, compared to previous generations of products, it is better at following the instructions configured by users within the time frame we need."
In many aspects, the competitors of Vesta include not only traditional mortgage systems such as ICE Mortgage Technology, but also companies like Xpanse that also attempt to automate the mortgage loan process in a native manner using AI.
Yu says that, at least compared to traditional established companies, the advantage of Vesta is that these companies were not created for AI agency purposes; "It is very difficult to overlay AI agency services on top of them." As for the next step, Yu stated: "Our top priority is to win business from the remaining customers in the mortgage industry. Then, we will follow where the customers' needs take us."
