Upstart Bank Charter Throws AI Lending Into Hot Seat

Upstart just moved one step closer to becoming a nationally chartered bank, and the real story is not the approval itself. It’s that the Upstart bank charter would pull an AI-driven lending marketplace deeper into the bank regulatory perimeter, where its underwriting models, funding strategy, and consumer controls face a tougher test.

The Office of the Comptroller of the Currency granted conditional approval for Upstart Bank, N.A., after Upstart submitted its application in March, according to PYMNTS. That conditional nod is a milestone, not a launch permit. Upstart still needs approval from the Federal Deposit Insurance Corporation for deposit insurance and from the Federal Reserve to become a bank holding company.

The thesis is simple: an Upstart bank charter could give the company more control over lending economics, product reach, and funding architecture. It also makes Upstart’s AI credit engine a supervisory issue, not just a fintech pitch deck.

Upstart's bank charter push turns its AI lending pitch into a regulatory test

Upstart has long sold itself as a lending marketplace powered by AI. A bank charter changes the frame. The company is no longer only matching borrowers, bank partners, and credit buyers. It is trying to own more of the regulated infrastructure underneath the loan.

That matters because Upstart says Upstart Bank would allow it to lower lending costs and bring its full product offering to all 50 states. The company also says the bank would not have physical branches, would originate consumer loans nationwide, would accept FDIC insured deposits, and would complement, not replace, existing funding partnerships.

“Upstart Bank will allow us to lower the cost of lending and bring our full product offering to all 50 states, advancing our mission to radically reduce the cost and complexity of credit for all Americans,” Upstart co-founder and CEO Paul Gu said.

XOOMAR analysis: the charter is best read as a control strategy. Upstart wants more direct command over origination, funding, product design, and national reach. That can make the business sturdier if executed well. It also gives regulators more reasons to inspect how the company’s AI systems behave when real borrowers, deposits, and bank capital are involved.

This follows our earlier read in Upstart Bank Charter Cracks OCC Door, AI Test Looms, where the central question was whether AI lending can pass bank-grade scrutiny rather than fintech-marketplace scrutiny.

The OCC conditional approval leaves two major gates closed

The OCC conditional approval lets Upstart keep moving toward Upstart Bank, N.A., but it does not mean the bank can open tomorrow. The company still has pending applications with the FDIC and the Federal Reserve, per Upstart’s release cited by PYMNTS.

That distinction matters. Conditional approval is a regulator saying the proposal may proceed if specified conditions are met. It is not the same as full operating approval.

At this stage, regulators will focus less on the marketing claim and more on the operating machinery. XOOMAR analysis: for a proposed national bank, the pressure points usually include:

• Capital: whether the bank has enough loss-absorbing capacity for its business model.
• Risk management: whether credit, liquidity, model, operational, and compliance risks are controlled.
• Board oversight: whether directors can challenge management and understand the AI-driven lending model.
• Consumer protection: whether borrowers receive fair, clear, and legally compliant treatment.
• Bank Secrecy Act controls: whether financial crime systems can support national operations.
• Operational readiness: whether the bank can run safely without physical branches.

Upstart’s proposed bank model also changes the supervision equation. A national charter can simplify some parts of a multi-state lending strategy, especially when a company wants to originate across state lines under one bank structure. But that simplicity comes with more direct federal supervision and higher expectations for governance.

Annie Delgado, Upstart’s chief risk officer and proposed CEO of Upstart Bank, framed that scrutiny as part of the bargain.

“We’ve been challenged extensively throughout the process, and that’s exactly what should happen when an institution is seeking the privilege of becoming a national bank,” Delgado said.

The numbers behind Upstart's charter strategy are about control, not just growth

The available numbers show why the Upstart bank charter is strategically attractive, but also why it raises the stakes.

Upstart’s 2024 company profile lists US$842 million in revenue, an operating loss of US$114 million, a net loss of US$109 million, US$2.37 billion in total assets, US$633 million in total equity, and 1,193 employees. Those figures describe a scaled fintech lender, not a small experiment.

A chartered bank model would not magically fix profitability. It would change the cost structure. Compliance, capital, examination readiness, risk staffing, audit systems, and board governance all get heavier. The trade is clear: Upstart may gain more direct control over lending and deposits, while accepting a bank-like burden.

Gu described the charter effort in March as the “natural evolution” of Upstart’s business as it grew in “size, scale and product offerings.” He also said the company wanted to “launch the first bank built from the ground up on AI.”

That phrase is the heart of the risk. If Upstart succeeds, it gets a cleaner platform for national AI lending. If it stumbles, the weak point will not be branding. It will be whether AI credit decisions can withstand bank supervision, consumer challenges, and credit stress.

Upstart's AI underwriting now has to satisfy bank examiners, not just loan buyers

Upstart’s core identity rests on AI-based credit decisioning. The charter process pushes that model into a stricter zone.

Regulators will care about explainability. They will care about repeatability. They will care about whether an automated system treats similar borrowers consistently and whether the bank can prove that outcomes are fair across protected classes.

XOOMAR analysis: the key supervisory questions are likely to cluster around model risk management:

• Validation: Can independent reviewers test whether the model performs as claimed?
• Bias testing: Does the model create hidden disparate outcomes through proxy variables?
• Adverse action notices: Can borrowers understand why they were denied or received certain terms?
• Data governance: Are inputs controlled, documented, and monitored?
• Drift monitoring: Does model behavior change when borrower profiles, macro conditions, or loan mixes shift?
• Audit trail: Can Upstart reconstruct and defend automated decisions after the fact?

Upstart’s pitch is that AI can reduce the cost and complexity of credit. That is a powerful claim. It is also exactly the kind of claim regulators will test.

Consumer advocates and examiners will not only ask whether more people get approved. They will ask who gets approved, on what terms, and whether the model’s reasoning can be challenged. A model that expands access while hiding discrimination would fail the public-interest test. A model that performs in normal periods but weakens under stress would fail the safety-and-soundness test.

This is where Upstart’s bank bid becomes bigger than Upstart. If the OCC, FDIC, and Fed allow a full launch, they are effectively saying an AI-first lender can sit closer to the core banking system, provided its controls meet bank standards.

Bank partners, borrowers, investors, and regulators are not watching the same story

Upstart says Upstart Bank will complement, not replace, its platform’s funding partnerships. That is an important line. The company is trying to reassure existing bank, credit union, and institutional fund relationships that a charter does not make them obsolete.

Still, incentives change when a marketplace lender becomes a bank. Some partners may value Upstart’s technology, referrals, or loan acquisition tools. Others may see a chartered Upstart as a more direct competitor in origination.

Borrowers could see benefits if the bank structure lowers costs or makes products available more consistently across all 50 states, as Upstart says. The risk is opacity. If automated underwriting decisions are hard to explain, a faster digital loan process can still feel arbitrary to the consumer on the wrong side of the model.

Investors get a different signal. The charter reads as a bid for long-term margin control and funding flexibility. But the same move brings bank-style expectations around capital, compliance spend, governance, and supervisory discipline.

Regulators are watching the hardest problem: whether AI lending can move from partner-dependent fintech operations into a national bank without weakening consumer protection or safety standards.

The broader lesson fits with other bank-risk stories we’ve tracked, including $8.5B Repayment Puts First Citizens SVB Debt on Trial. When financial firms take on more direct control, they also take on more visible accountability.

Upstart joins fintechs chasing charters for fewer dependencies

PYMNTS reported in February 2025 that some fintechs had opted to gain their own banking licenses, and that a push to overhaul the application process could make more fintechs consider that path. Upstart’s move fits that pattern.

The logic is not hard to see. Sponsor-bank arrangements can help fintechs move quickly, but they also leave strategic functions outside the fintech’s direct control. A bank charter can reduce reliance on third parties for origination and deposits, if regulators allow it.

Banking Dive reported that the OCC conditionally approved Upstart’s charter application about four-and-a-half months after it was submitted. The same report said Upstart was advised by Klaros Group, and quoted Klaros co-founder Michele Alt describing the approval as “speedy.” It also noted that Wise received the OCC’s first charter denial since at least 2010, while Nubank, Mercury, and Valt Bank have received conditional approval, with Erebor Bank receiving full approval.

Those details support a narrow conclusion, not a sweeping one. The door may be more open for fintech charter applicants than before, but conditional approval remains a high bar. Full approval and actual bank operation are harder tests.

For Upstart, the post-application phase is where the corporate story becomes an operating story. The company has to show that its lending model can run inside a bank without relying on the looser flexibility of a pure marketplace structure.

The next test is whether Upstart can make AI look boring to regulators

If Upstart clears the FDIC and Fed gates, AI lenders will have a sharper playbook. Seek a charter, bring more lending infrastructure in-house, accept heavier supervision, and prove the model works under bank rules.

That last part is the hard part.

Regulators are likely to demand a controlled launch posture: strong capital planning, careful product rollout, documented model governance, and proof that Upstart’s automated systems can be monitored and challenged. XOOMAR analysis: the evidence that would strengthen Upstart’s case includes clear final approvals, transparent operating conditions, disciplined growth, and no early signs of model or compliance strain.

The evidence that would weaken the thesis is just as clear. Delays at the FDIC or Fed, restrictive conditions, fair lending questions, weak credit performance, or signs that existing partners pull back would suggest the charter is more burden than advantage.

The Upstart bank charter is a door opening, not a victory lap. Upstart now has to prove that a bank built around AI can act like a bank first, and a technology company second.

Disclaimer: This XOOMAR analysis is for informational and educational purposes only. It is not financial, investment, legal, tax, or professional advice. It does not provide buy, sell, hold, price-target, portfolio, or personalized recommendations. Verify information independently and consult qualified professionals before making decisions.

Impact Analysis

• A bank charter would move Upstart’s AI lending model under deeper federal supervision.
• Upstart could gain more control over lending costs, products, and funding if final approvals are secured.
• The decision could shape how regulators assess AI-driven credit underwriting inside the banking system.