SoftBank seeks $40B loan to deepen massive OpenAI bet

SoftBank is seeking a bridge loan of up to $40 billion to fund most of its planned investment in OpenAI, according to claims from Bloomberg.

If completed, it would be the biggest borrowing in dollars in the company’s history. The proposed loan would run for about 12 months, and four lenders, including JPMorgan, are reportedly going to underwrite it.

The size of the borrowing shows how far Masayoshi Son is willing to go to make OpenAI the core of SoftBank’s next big trade. The company is preparing a $30 billion bet on OpenAI after already putting in more than $30 billion.

That makes the startup the main focus of Masa’s current strategy, which is a far more expensive gamble than his earlier wins in Alibaba and ByteDance.

SoftBank is still interested in building a bigger OpenAI position with debt

At the end of December, SoftBank had like 11% of OpenAI, but to keep adding to that position, the Japanese group has sold assets, including its stake in Nvidia, to raise cash.

OpenAI is now one of SoftBank’s biggest holdings, sitting next to its roughly 90% stake in chip designer Arm.

At the same time, investment activity in other parts of the portfolio has slowed, which has tied SoftBank’s stock more closely to how ChatGPT performs against rival products such as Google’s Gemini and Anthropic PBC’s Claude. If OpenAI wins more users, more customers, and more trust, SoftBank benefits. If the field gets tighter, SoftBank takes that hit too.

This week, S&P cut SoftBank’s credit outlook. The rating agency said the company’s growing exposure to OpenAI could weaken liquidity and hurt the quality of its assets. That warning landed right as Masa was pushing for even more funding.

Altman defends democratic control as OpenAI faces backlash over its Pentagon contract

While SoftBank was trying to raise money, Sam Altman was dealing with the politics around OpenAI’s military work. Speaking on Thursday at the Morgan Stanley Tech, Media and Telecom Conference in San Francisco, Sammy said elected officials, not tech executives, should decide the limits of how AI can be used in national defense.

The OpenAI boss said, “We have to trust in the democratic process” as OpenAI and Anthropic work through separate talks with the Department of Defense.

Sam also said, “This process is messy. This process has some deep flaws, but it is better than all other systems.” He added that abandoning that process because people dislike current leaders would be bad for society.

Sam’s comments came after OpenAI reached a deal with the Pentagon to use its models in classified settings. The deal followed a standoff between the agency and Anthropic, which refused to give up certain red lines tied to control over military uses of its technology.

Earlier in the week, Sam told employees in a memo that he regretted moving too fast to secure the agreement. He said it looked “opportunistic and sloppy.”

During a Tuesday all-hands meeting, he admitted the move created “very negative PR” because it made OpenAI look willing to let its tools be used for activities that could include collecting data on Americans. Sam told staff, “I feel terrible for subjecting you all to this.”

At the conference, Sam said the Defense Department had been “extremely understanding” about the need to “clarify” parts of the contract. He also said a new law is needed to match modern technology and reduce harm.

Sam said one of the country’s most important civil liberties is that the government should not spy on its own citizens without warrants and proper legal process, and he added that the meaning of that standard now needs to change with technology.

Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.