Founders lock in AI control as SpaceX targets IPO

SpaceX has finally released its long-awaited prospectus for an initial public offering, with OpenAI and Anthropic expected to follow later this year, raising critical questions about the unprecedented corporate governance structures that will govern some of humanity's most consequential technologies. An analysis by Gill Whitehead published by the Oxford Internet Institute reveals that all three companies have adopted radical departures from traditional corporate oversight, including dual-class share structures that grant founders like Elon Musk ten times the voting power of ordinary shareholders, enabling minority ownership while maintaining absolute control. These governance experiments, which include regulatory arbitrage across U.S. states and novel hybrid structures such as Anthropic's Long-Term Benefit Trust, have emerged partly as a backlash against perceived compliance costs and short-term market pressures on innovation. The implications extend beyond corporate law into fundamental questions about human agency and meaning. Writing in The Times of Israel, Lemonade CEO Daniel Schreiber argues that AI revives ancient dilemmas about who should benefit from abundance, what learning is for, and whether superior intelligence can safely replace human judgment. Schreiber draws on Talmudic tradition to suggest that even when AI produces better outcomes than human deliberation, something irreplaceable is lost by outsourcing moral reasoning and self-governance. Similarly, tech writer Wendy Liu describes her deliberate choice to avoid AI tools despite personal efficiency costs, arguing that preserving cognitive sovereignty and the hard work of thinking may be essential to maintaining humanity itself. Institutional investors led by public pension funds are already lobbying for governance reforms, including time limits on dual-class share structures. Whitehead warns that as companies engineer digital AI guardrails, many corporate governance backstops have been simultaneously dismantled, leaving decisions about humanity-impacting technology dependent on the good intentions of a few founders. Future AI regulation, she argues, must account for these concentrated power structures and resist the temptation to trust individual judgment as the final line of defense.