Sanders and Bannon Push Data Center Freeze as OpenAI Backs Mandatory AI Audits

Senator Bernie Sanders and former Trump strategist Steve Bannon are now pushing in the same direction on artificial intelligence, demanding hard limits on the technology's expansion. Sanders has called for a freeze on new data center construction, while OpenAI has, for the first time, thrown its support behind the FRONTIER Act, which would require mandatory independent safety audits for advanced AI systems. The alignment of figures from opposite ends of American politics signals that AI regulation is no longer a fringe concern in Washington.

Sanders' proposed moratorium on data centers targets the physical infrastructure underpinning the AI boom, a stance that echoes growing local opposition to the power and water demands of large server farms. His office has not released a detailed legislative text, but the call itself moves the debate beyond voluntary corporate pledges. It also puts him at odds with tech industry groups that argue such a freeze would stall economic growth and cede AI leadership to China.

OpenAI's endorsement of the FRONTIER Act is the more immediate shift. The company had previously resisted binding outside audits, preferring internal safety reviews and voluntary commitments. By backing mandatory third-party evaluations, OpenAI is effectively conceding that self-regulation has not persuaded lawmakers. The bill's exact audit requirements remain under negotiation, but the principle of independent scrutiny now has a major industry player on record supporting it.

President Trump has publicly dismissed the need for sweeping AI rules, yet the bipartisan pressure is mounting from both flanks of his own coalition. Bannon's involvement gives the restrictionist push a nationalist-populist voice, while Sanders supplies progressive muscle. That combination makes it harder for the White House to frame AI oversight as a partisan issue. Congressional aides say the number of co-sponsors for various AI safety bills has quietly grown in recent months.

What changes for the field is the credible prospect of enforceable audits and infrastructure limits, not just ethical guidelines. AI labs that have built their timelines around minimal regulatory friction may now need to budget for compliance reviews and site-permitting delays. Investors, too, will weigh data center freezes against the compute demands of next-generation models. The open question is whether this odd-couple alliance can survive contact with committee markups and industry lobbying, or whether it fractures before any bill reaches the floor.