What happened in this episode?
At the Masters of Scale Summit 2026, former Google chief executive Eric Schmidt tells host Bob Sapian that the US–China AI gap has narrowed to roughly six months, despite US chip export restrictions he says he helped shape. Schmidt attributes the change to rapid declines in software costs and to Chinese technical progress in efficiency, arguing the contest has become a “two-horse race” focused on increasingly superhuman reasoning systems.
Schmidt says China’s structural advantages include faster build-out of energy and data centres, and he urges US leaders to respond by siting large facilities on available land and building dedicated power generation. He also outlines a strategic divide between open-weight/open-source approaches—more common in China—and closed models prevalent among major US labs, framing it as a defining competition for the next phase of AI.
On safety, Schmidt warns that open-weight releases allow “guard rails” to be removed, leaving raw models that could be exploited for cyber attacks; he predicts a significant cyber event within 12 months could make the risk tangible. He also argues China’s speed of deployment is ahead, and predicts that over time most countries may default to Chinese models because they are cheaper and “good enough”, with geopolitical consequences. He closes by urging business leaders to adopt AI in day-to-day operations and expects regulation to follow as negative outcomes emerge.