Bank of America just extended OpenAI a $520 million credit line, its first loan to the company, after turning down the same request not long ago. The reversal lands weeks before OpenAI is expected to push toward a US listing.
You don't usually watch a bank change its mind this fast. Bank of America told OpenAI no on borrowing, then said yes, according to Bloomberg, which reported the deal on July 8. The new facility makes BofA one of OpenAI's biggest lenders and adds to an existing undrawn line from other banks, pushing OpenAI's total available credit above $5 billion.
Brian Moynihan runs a bank that has spent years being the least adventurous name in AI financing. As recently as last year, Bloomberg reported, BofA executives were openly skeptical that companies burning cash at OpenAI's rate could sustain their business models. Skepticism doesn't pay the bills. An IPO does.
OpenAI is preparing a US listing, and Bank of America wants in on it. Winning an advisory mandate on one of the largest tech IPOs in years is worth far more to a bank than the interest on a $520 million line. Bloomberg's reporting notes BofA is also chasing an advisory role on Anthropic's planned listing, which means the same bank is now positioning itself on both sides of the frontier AI race at once.
There's a second reason this matters to Moynihan specifically. BofA's Merrill Lynch wealth arm cares deeply about being associated with marquee stock listings, since it's the kind of visibility that pulls in new brokerage clients. Credit is the entry fee. Advisory fees and Merrill's brand halo are the actual payoff.
None of this happens in a vacuum. Bank of America has helped raise nearly $500 billion in AI-related capital since 2025, per Bloomberg, roughly 60% of all such fundraising across investment-grade debt, leveraged finance, and equity markets. That's not a bank dabbling in a hot sector. That's a bank that has decided AI financing is its business now, full stop.
What changed isn't OpenAI's balance sheet. It's still unprofitable, still spending at a pace that would sink most companies. What changed is the calculus around risk. A bank that once treated OpenAI's borrowing needs as a red flag now treats them as the price of a seat at the table.
Anthropic is playing a different game #
Anthropic hasn't taken this route. It has leaned on massive equity rounds from Amazon, Google, and other backers rather than courting direct credit lines from banks still working out how to underwrite a company with no clear path to profitability. That's a more conservative capital structure, and it costs Anthropic less optionality with lenders like BofA right up until IPO season, when both companies will need the same banks to run their books.
Frankly, the interesting part isn't the $520 million. It's that a bank this risk-averse decided the reputational cost of staying out was higher than the credit risk of getting in. Direct loans to unprofitable AI labs were, until recently, treated as a bridge too far even by banks eager to underwrite their stock and bond sales. That line just moved, and every other lender watching OpenAI's IPO build toward launch now has to decide whether to move with it.
OpenAI hasn't announced IPO timing or an exchange. Bank of America hasn't confirmed which bank will ultimately lead the listing. What's confirmed is simpler: the bank that said no is now on the hook for over half a billion dollars, betting the listing arrives before the bill does.
Also read: xAI Launches Grok 4.5 With Pricing Built to Undercut Anthropic's Opus 4.8 • Prime Intellect Raises $130 Million for Enterprise AI Agents • China Will Let Its Top AI Firms Buy Nvidia's H200 Chips Again