SoftBank cuts OpenAI-backed loan target to $6B as lenders balk at valuation

SoftBank has cut a planned OpenAI‑backed margin loan from around $10 billion to roughly $6 billion after banks and private credit funds pushed back on the deal’s structure and the difficulty of valuing OpenAI, an unlisted AI unicorn.

SoftBank Group is scaling back an ambitious financing plan that would have raised around $10 billion using its OpenAI shares as collateral, after lenders balked at both the structure of the margin loan and the private valuation underpinning the deal.

According to Bloomberg, cited by Reuters and other outlets, SoftBank and its arranging banks have floated a reduced target “as low as $6 billion” in recent discussions with prospective lenders, implying a 40% cut from the original size. U.S. News reported that the initial pitch “had investors concerned about the difficulty of reaching a valuation for an unlisted company like ChatGPT maker OpenAI.”

Lenders question private OpenAI valuation and structure

Bloomberg’s April scoop on the deal said SoftBank was seeking “a $10 billion loan secured by its shares in US artificial intelligence giant OpenAI,” structured as a two‑year margin loan with an option to extend by another year. Bloomberg said the facility would let the Japanese conglomerate “take on more debt for its push into AI” without selling down its OpenAI stake.

In practice, the mechanics are straightforward but risky: SoftBank borrows against its OpenAI equity, and if the value of that collateral falls, lenders can demand more margin or seize shares. The sticking point, lenders now say, is how to price that collateral. As the Economic Times summarized from Bloomberg’s reporting, “some creditors expressed concerns over how to value OpenAI, a privately held company,” which has missed some internal sales and user milestones in recent quarters. A separate breakdown on Chinese platform Futu noted that “the crux of the issue lies in lenders’ inability to determine a reasonable valuation” for OpenAI, calling the loan talks “a major setback” for SoftBank’s AI leverage strategy.

Those worries are layered on top of SoftBank’s already sizable AI financing stack. In March, Bloomberg reported that SoftBank had secured a $40 billion bridge loan to fund its OpenAI investment and general corporate needs, a deal backed by a syndicate of global banks and now being syndicated out to more lenders. Bloomberg said institutions such as HSBC, BNP Paribas and Intesa Sanpaolo were joining as sub‑underwriters, each asked to commit around $5 billion. Reuters, summarizing the same reporting, added that SoftBank “is pursuing a loan of as much as $40 billion” to support its OpenAI bet. Yahoo Finance echoed that figure, underscoring just how much leverage is already tied to the AI trade.

What the scaled‑back loan means for SoftBank’s AI push

The margin loan was meant to be another pillar in that structure: by borrowing against OpenAI shares rather than selling them, SoftBank can raise cash to expand its AI investments — potentially into infrastructure projects like the “Stargate” data‑center initiative — while preserving upside if OpenAI’s valuation continues to rise. TechFundingNews noted that the $10 billion margin loan was “just one part of SoftBank’s larger AI financing plan,” which reportedly includes commitments of more than $60 billion to OpenAI and related ventures through Vision Fund 2, the $40 billion bridge loan and other facilities.

Cutting the margin‑loan target to around $6 billion does not end that strategy, but it signals that creditor appetite for concentrated, private‑equity‑collateral risk is not unlimited, even in a market obsessed with AI. As Bloomberg put it in an earlier piece on the $40 billion bridge, the OpenAI exposure is already “one of the biggest tests yet of creditor sentiment toward the Japanese conglomerate’s debt-fueled push further into artificial intelligence.”

The latest downsizing suggests that, for now, banks and funds are willing to back SoftBank’s AI ambitions — but only up to the point where they can still convince their own risk committees that the collateral they are lending against can be valued with something more than vibes and secondary‑market whispers.