SoftBank Wants to Borrow $40 Billion to Buy More OpenAI. What Could Go Wrong?

Masayoshi Son has never been accused of thinking small. Now SoftBank is reportedly seeking a loan of up to $40 billion, mainly to help fund its investment in OpenAI. If your first thought was “that is a lot of borrowed money”, welcome to the AI economy in 2026.

The Reported Loan and the Announced Investment

Bloomberg reported on March 6, citing people familiar with the matter, that SoftBank was seeking a dollar-denominated bridge loan of as much as $40 billion, with a term of around twelve months. JPMorgan was named among four expected lenders.

That is a report about financing being sought, not confirmation that the full sum has already been borrowed or spent. It also should not be confused with the amount of SoftBank’s newly announced OpenAI investment.

SoftBank’s February 27 agreement is for $30 billion in additional investment, planned in three $10 billion tranches on April 1, July 1 and October 1, subject to closing conditions. On completion, it expects cumulative investment of $64.6 billion and roughly 13% ownership.

Two enormous numbers, two different jobs: one is the reported size of a potential financing facility; the other is the new investment commitment. The distinction is less exciting than the headline, but the accountant insists.

What the $730 Billion Valuation Means

OpenAI’s February 27 announcement describes $110 billion in new investment at a $730 billion pre-money valuation. The announced contributions were $50 billion from Amazon and $30 billion each from NVIDIA and SoftBank, with additional investors expected as the round progressed.

Pre-money means before adding the new capital. It should not be described as the company’s valuation after receiving all $110 billion, nor should an announced multi-stage round be treated as every dollar arriving at once. We covered the broader OpenAI funding announcement separately.

A private valuation is also not cash an investor can automatically withdraw. The value eventually realised depends on the investment’s terms, subsequent financing and whatever opportunity exists to sell.

Why Use a Bridge Loan?

A bridge provides money now while a borrower arranges longer-term financing or other sources of funds. SoftBank’s announcement says it expects to use bridge loans and other financing initially, then replace them over time using existing assets and other measures.

That is more specific than assuming the entire plan is “wait for OpenAI to go public”. An IPO could be relevant to the investment story without being a confirmed repayment event on a fixed date.

A twelve-month loan and a long-term investment live on different clocks. If the investment takes longer to deliver value, the borrower still needs a way to address the financing when it matures. The calendar does not accept “the technology is very promising” as a payment method.

The Risk Is in the Assumptions

The attraction of OpenAI is clear enough: a major position in a company central to commercial AI. The difficult questions concern future competition, the cost of delivering its services and the returns that can justify such a large commitment.

Borrowing adds financing obligations to that uncertainty. It does not make a successful outcome impossible, but it changes the consequences of delays or disappointing results. Evaluating the exposure requires the actual loan terms and the borrower’s wider assets, rather than simply treating the largest number as a verdict.

Nor would a hypothetical future trillion-dollar valuation automatically prove an extraordinary investment return. Purchase prices, ownership changes and financing costs all matter. Multiplying a headline valuation by an old ownership percentage is not the same as calculating profit.

The Pudgy Cat Take

This is a story about the scale of the financial commitments around AI as much as the technology itself. SoftBank is willing to commit heavily to OpenAI’s growth; the reported borrowing would help fund that commitment.

That does not settle whether models, distribution, infrastructure or financing will matter most in the next stage of the industry. They affect one another. An impressive balance sheet cannot make an unhelpful product useful, and a useful product still needs a sustainable way to operate.

For now, the confirmed investment agreement and the reported loan should remain separate in our minds. Son is making a large bet, and borrowed money adds another deadline to it. Nobody knows how this ends. We do know the spreadsheet needs room for quite a few zeroes.