Sam Altman says he doesn’t want the government to bail out OpenAI if it fails

OpenAI executives have answered a lot of questions about how they expect to pay for the $1.4 trillion worth of data center buildouts and usage commitments they have accumulated this year, given that their revenue — while rising quickly — is at an annual run rate of $20 billion, CEO Sam Altman said Thursday in a post on X.

Altman’s comments came in response to comments made by OpenAI CFO Sarah Friar, which she quickly walked back. Speaking at a Wall Street Journal event on Wednesday, Friar said she wants the US government to “back” her company’s infrastructure loans. This would make the company’s loans cheaper and help ensure it can always use the latest and greatest tranche, she explained.

A subsidized loan is when the government guarantees it, so if the company defaults, taxpayers pick up the bill. Lenders tend to reward low-risk loans like these with better terms.

Using older chips, which is what compute-constrained OpenAI must do, makes financing options affordable, but the company’s goal is to always base its modern models on the latest and greatest chips, Friar said.

So how do we pay for this revolving door of chips? She said the company is looking for an “ecosystem” to help, including banks and private equity firms, the government hopes.

When asked what she wanted the government to do, she said: “…backstop, the guarantee that allows financing to happen. This would lower financing costs but also increase the loan-to-value, so the amount of debt you can take on top of the equity.”

She also implied that such talks, especially in the US, were already in the works, saying: “I think we’re seeing that. The US government, in particular, has been incredibly forward-leaning and has really realized that AI is almost a national strategic asset.”

After the Wall Street Journal published a clip of her discussing this desire for federal support, and many X users with large followings mocked the idea, Friar quickly retracted her comments.

“I want to clarify my comments earlier today. OpenAI is not seeking government support for our infrastructure commitments. I used the word ‘backstop’ and distorted the point,” she wrote on LinkedIn.

On Thursday, David Sachs, Trump’s AI czar, commented on the matter. Sachs (himself a major Silicon Valley investor), wrote on X that the United States has no plans to bail out any AI companies.

“There will be no federal bailout for AI. The United States has at least 5 leading model companies. If one company fails, others will take its place,” he said, adding that what the government wants to do is make it easier to “issue permits and generate energy.” While he did not mention her by name, he also forgave the monk for “clarifying” her position.

In the aftermath, Altman wrote a lengthy post on X echoing Sacks’ sentiments.

“We do not have and do not want government guarantees for OpenAI data centers. We believe that governments should not pick winners or losers, and that taxpayers should not bail out companies that make poor business decisions or lose in the market,” he wrote.

He also explained that subsidized loans had been discussed – but not for his company.

“The only area in which we have discussed loan guarantees is in part to support the construction of semiconductor fabs in the US, where we and other companies have responded to the government’s call and where we would be happy to assist (although we have not made a formal request).”

It’s hard to blame the monk for coming up with the idea. She is right that such a guarantee would make her financing task easier, even if the idea of ​​requesting a taxpayer-funded bailout is “ridiculous,” as Sachs wrote in his article.

Since she’s now heard a resounding public “no” from someone she needs on her side to the idea, she and OpenAI CEO Sam Altman can expect more questions about how they expect to pay for the trillion-dollar construction costs.

In fact, Altman seems prepared for such a thing.

“We expect to end this year with an annual revenue run rate of more than $20 billion and grow into the hundreds of billions by 2030. We are looking at commitments of approximately $1.4 trillion over the next eight years,” he wrote, adding that the company feels good about its “prospects,” particularly its enterprise offerings, new consumer devices, and robotics.