Upstart secured conditional approval from the Office of the Comptroller of the Currency to create a national bank powered entirely by AI underwriting. The move pushes the fintech closer to becoming the first bank built from the ground up on machine learning models managing credit risk and loan pricing.
Unlike traditional banks that adapt legacy systems, Upstart Bank will architect its operations around AI, aiming to streamline loan origination and portfolio management. The application was filed in March 2026, and it took just over four months to get this key regulatory green light.
However, Upstart still needs the Federal Deposit Insurance Corporation to approve deposit insurance and the Federal Reserve’s nod for bank holding company status. These two remaining approvals will be key before the bank can fully operate nationally.
Currently, Upstart partners with banks and securitization markets to fund loans. Controlling its own bank means it could replace expensive third-party funding with cheaper deposits, potentially lowering capital costs significantly.
Leadership shifts are also underway, with Chief Risk Officer Annie Delgado proposed as the bank’s CEO and CTO Paul Gu in a key leadership role as the institution takes shape.
Other fintech players like SoFi and LendingClub have shown that owning a chartered bank can improve lending economics. Upstart’s AI-first approach could set a new standard in this evolving landscape.
