OpenAI explicitly requested federal mortgage ensures for AI infrastructure in an October 27 letter to the White Home—which kindly refused the provide, with AI czar David Sacks saying that at the least 5 different firms may take OpenAI’s place—immediately contradicting CEO Sam Altman’s public statements claiming the corporate would not need authorities help.
The 11-page letter, submitted to the Workplace of Science and Expertise Coverage, referred to as for increasing tax credit and deploying “grants, cost-sharing agreements, loans, or mortgage ensures to develop industrial base capability” for AI information facilities and grid parts. The letter detailed how “direct funding may additionally assist shorten lead instances for crucial grid parts—transformers, HVDC converters, switchgear, and cables—from years to months.”
“Preliminary investments could possibly be made utilizing current authorities such because the Protection Manufacturing Act Title III and the Division of Power’s Mortgage Applications Workplace,” OpenAI mentioned.
Simply 10 days later, on November 6, Altman posted on X that “we should not have or need authorities ensures for OpenAI information facilities,” including that “taxpayers shouldn’t bail out firms that make unhealthy enterprise selections.”
I wish to make clear a number of issues.
First, the apparent one: we should not have or need authorities ensures for OpenAI datacenters. We imagine that governments shouldn’t decide winners or losers, and that taxpayers shouldn’t bail out firms that make unhealthy enterprise selections or…
— Sam Altman (@sama) November 6, 2025
The contradiction emerged after OpenAI CFO Sarah Friar advised a Wall Road Journal occasion on November 5 {that a} federal “backstop” may assist decrease financing prices and enhance debt capability for AI infrastructure.
Her feedback triggered fierce backlash. For instance, Florida Governor Ron DeSantis tweeted that the federal government shouldn’t bail out tech firms. Sacks wrote that “there will probably be no federal bailout for AI.”
Friar rapidly walked again her feedback on LinkedIn, saying OpenAI wasn’t searching for a authorities backstop for infrastructure commitments.
“I wish to make clear my feedback earlier as we speak. OpenAI isn’t searching for a authorities backstop for our infrastructure commitments. I used the phrase ‘backstop’ and it muddied the purpose,” she wrote. “As the complete clip of my reply exhibits, I used to be making the purpose that American energy in expertise will come from constructing actual industrial capability, which requires the non-public sector and authorities enjoying their half.”
Altman’s prolonged X put up the subsequent day amplified this message. “Our CFO talked about authorities financing yesterday, after which later clarified her level, underscoring that she may have phrased issues extra clearly. As talked about above, we expect that the U.S. authorities ought to have a nationwide technique for its personal AI infrastructure,” his tweet reads.
After all, this transfer additionally generated backlash.
AI researcher Gary Marcus revealed the October 27 letter, calling Altman’s denial “mendacity his ass off” and noting the letter explicitly requested the very mortgage ensures Altman claimed to not need. The letter stays publicly obtainable on OpenAI’s content material supply community.
This is not the primary time Altman’s statements have confronted scrutiny. OpenAI’s board briefly fired him in November 2023 for being “not persistently candid,” in response to the board’s assertion. Former board member Helen Toner later detailed on a podcast how Altman had withheld data and made it tough for the board to satisfy its oversight duties. A current deposition from former OpenAI Chief Scientist Ilya Sutskever, who voted to take away Altman, additional documented considerations about his candor.
OpenAI didn’t instantly reply to requests for remark concerning the October 27 letter or the obvious contradiction with Altman’s statements.
