OpenAI CEO Sam Altman took the stage at Stanford's CS 153 class to map out the next decade — and his prognosis is both bold and unsettling. The biggest takeaway? Global compute shortages might not be a temporary bottleneck but a permanent condition.
Altman sees compute as the most critical public utility of the future. The smarter and cheaper models get, the more insatiable the demand becomes, he argued. As personal AI agents become ubiquitous, users will want dozens or even hundreds of them running 24/7, creating an endless appetite for inference compute. Once supply and demand get that out of whack, the question of how to fairly allocate compute becomes an urgent social problem.
On the distribution of wealth and control, Altman estimated there's an 80 percent chance the world will go the democratic path of widespread technology distribution. But he warned that forces invoking safety and stability will try to concentrate power in a few giant hands. Even as the CEO of OpenAI, he said, the world should not tolerate a handful of companies monopolizing the immense wealth AI creates. As the economic lever shifts from labor to capital, Altman prefers a 'Citizens' Wealth Fund' model — where the public collectively holds equity in companies — to let everyone share in capitalism's long-term rewards.