(Reuters) -OpenAI has spoken with the U.S. government about the possibility of federal loan guarantees to spur construction of chip factories in the U.S., but has not sought U.S. government guarantees for building its data centers, CEO Sam Altman said on Thursday. 

Altman said the discussions were part of broader government efforts to strengthen the domestic chip supply chain, adding that OpenAI and other companies had responded to that call but had not formally applied for any financing. He said the company believes taxpayers should not backstop private-sector data center projects or bail out firms that make poor business decisions.

The comments come a day after OpenAI's chief financial officer said the company was hoping the federal government would help guarantee the financing of chips behind its data center investments. Those remarks sparked widespread criticism that OpenAI was seeking a government bailout.

Booming demand for artificial intelligence models and products, including OpenAI's popular ChatGPT tool, has driven top tech companies to lay out plans to build more data centers and chips.

Tech officials argue that these investments are tantamount to a national security asset for the U.S. government, given AI's growing role in the U.S. economy. OpenAI has committed to spend $1.4 trillion building computational resources over the next eight years, Altman said Thursday. 

OPENAI EYES WAYS TO SELL COMPUTING CAPACITY DIRECTLY

Unlike rivals Meta and Google, OpenAI doesn't have a wildly profitable advertising business or a thriving cloud services business to help cover the cost of building out these data centers.

OpenAI has worked around that constraint by signing a series of infrastructure deals with tech companies including Nvidia ( NVDA ) and AMD. In all, the infrastructure deals total an estimated $1 trillion.

It's still an open question how the unprofitable company will pay for this build-out. A recent securities filing from Microsoft ( MSFT ) indicates that OpenAI lost more than $12 billion in the third quarter, analysts say. 

In a post on X, Altman outlined how the company is exploring new business models and offerings to fund its expanding infrastructure. The startup expects to finish this year with an annualized revenue run rate above $20 billion, with ambitions to reach hundreds of billions by 2030, according to Altman. OpenAI was last valued at $500 billion. 

"Is OpenAI trying to become too big to fail, and should the government pick winners and losers?" Altman wrote on Thursday. "Our answer on this is an unequivocal no."

OpenAI officials are also exploring a range of financing options.

"If we screw up and can't fix it, we should fail, and other companies will continue on doing good work and servicing customers," Altman said. "The ecosystem and economy would be fine."

He said OpenAI is considering ways to sell computing capacity directly to companies and individuals, a service he described as an "AI cloud."

Such a move would position OpenAI as a potential rival to major cloud providers, including Microsoft ( MSFT ) and Alphabet's Google, as well as emerging "neocloud" firms such as CoreWeave ( CRWV ), which currently supply its GPU infrastructure.

Altman added the effort could require raising additional equity or debt financing. The plan signals OpenAI's intent to compete more directly with its own cloud partners in providing the computational power that underpins AI models.

This week, OpenAI CFO Sarah Friar told a tech conference the government's involvement would help drop the cost of financing chips and data centers and allow companies like OpenAI to take on more debt. 

Friar's remarks come as investors mull over fears of an AI bubble, questioning the returns on hundreds of billions of dollars in investment on AI expansion. 

David Sacks, the White House artificial intelligence and crypto czar, said earlier on Thursday there will be no federal bailout for AI, as the U.S. races to cement its position as a global leader in the booming technology.

"Given our vantage point, we feel good about it. But we of course could be wrong, and the market - not the government - will deal with it if we are," Altman said, addressing talks of the federal government not stepping in if the massive build-out of AI infrastructure does not produce desired results. 

(Reporting by Arsheeya Bajwa in Bengaluru and Krystal Hu in San Francisco; Editing by Alan Barona, Rod Nickel, Kenneth Li and Lincoln Feast.)