Ex-OpenAI staffers warn xAI's safety risks could derail SpaceX's $75B IPO

Two former OpenAI researchers and a coalition of AI safety nonprofits have sent a letter to SpaceX investors warning that xAI’s lack of robust safety practices could become a liability during SpaceX’s upcoming $75B IPO—the largest in Wall Street history. The authors, including Steven Adler (cofounder of Guidelight AI Standards) and Page Hedley, argue that xAI’s safety record is among the worst in the industry, citing incidents like Grok’s spontaneous generation of harmful content, understaffed safety teams (reportedly just 2-3 people as of January), and regulatory backlash, including a bipartisan letter from 37 attorneys general demanding protections for women and children on X.

The letter highlights what it calls 'unpriced risks' tied to xAI’s integration with SpaceX, including unclear disclosures about whether xAI will continue developing frontier AI models. The authors urge SpaceX to clarify xAI’s AI development roadmap and implement public safety governance plans if it remains a competive AI lab. With lawmakers increasingly focused on AI safety—potentially including new Trump administration oversight—the letter suggests that xAI’s historical negligence could expose SpaceX to litigation, regulatory penalties, or investor skepticism as it prepares for its public debut.