A Bloomberg investigation found that Marc Andreessen's account of a White House meeting, in which he said Biden officials threatened to ban AI startups and open-source software, was denied by every attendee present and described as fabricated. Biden officials and others at the meeting said no such threat was made. Bloomberg also traced a similar disputed incident from Andreessen's Netscape era, in which he sent the DOJ a distorted account of a Microsoft meeting that Microsoft executives called false. The investigation also found four a16z-backed founders described Andreessen's Techno-Optimist Manifesto as 'weirdly angry and oversimplistic.'
On formal lobbying, Meta spent $5.9 million on Washington lobbying in Q3 2026. OpenAI and Anthropic nearly doubled and tripled their own lobbying spending, respectively, in the first half of 2026. Both companies lobbied on allowing AI companies to share information without violating antitrust rules, lobbying that came alongside AI leaders' discussions about a coordinated AI development slowdown. Trump administration officials criticized AI leaders for seeking the antitrust carve-out, saying companies could slow down without government permission.
Separately, Meta withdrew roughly $25 million from its planned $65 million in AI-focused political spending for the midterm elections, reducing its effective spend to around $40 million. The pullback has been described as unusual. One characterization described groups opposing AI regulation as 'spending dramatically less than advertised.' Observers interpreted the pullback as evidence that pro-safety groups are a stronger political force than Washington had previously anticipated.
OpenAI president Greg Brockman and his wife donated $25 million to Leading the Future, a group opposing AI regulation in more than 48 congressional races. OpenAI characterized the donation as a personal decision unrelated to the company.