Eric Schmidt argues a mutual pause in the AI race is unfeasible

Eric Schmidt does not think the AI race between the United States and China is going to take a breather. In a September 2026 interview with The Economist, the former Google CEO called a mutual pause unworkable.

His reasoning comes down to two problems. US incentives reward building faster, and nobody has a reliable way to confirm the other side actually stopped.

“We’re not going to do a pause…how do you verify it?”

Why Schmidt says a pause breaks down

The first obstacle he identified is structural. American companies, investors, and institutions are built to reward continuous development, and Schmidt sees no mechanism that would make them voluntarily hit the brakes.

The second obstacle is verification. Schmidt’s point is that labs can cheat, and outsiders would have a hard time proving it. A frontier AI model gets trained on computing hardware inside private facilities. From the outside, a lab that has paused can look a lot like a lab that has simply stopped talking about its work.

Progress with a safety net

Schmidt is not arguing for recklessness. He framed the competition as a way to establish a dominant model that can help contain the risks tied to emerging technologies, and he expressed concern about how AI development will shape everyday life.

His alternative to a pause is progress paired with built-in safety mechanisms. One idea he raised: have older AI models supervise newer ones. He suggested that this kind of oversight might use models that are four months older than the system being watched, depending on whether the approach proves effective.

A diplomatic starting point: no surprises

On the international front, Schmidt proposed starting small. He called for initial bilateral treaties between the US and China built around a “no surprises” rule. The goal of that rule would be to head off operational mishaps when AI systems are deployed. After that foundation, Schmidt wants broader cyber norms to follow.

Background: an old position, a new moment

Schmidt’s stance is not new. He has opposed slowdowns that could hand China an advantage since at least 2023, and this interview restates that view at a moment when the stakes feel higher.

The industry remains split on how fast AI should move. One camp backs safety-driven slowdowns, while the other argues any pause would simply let China pull ahead of the US.

An upcoming summit between Chinese President Xi Jinping and Donald Trump could put competitive AI strategy on the agenda, which makes the question of pauses, treaties, and guardrails less theoretical than usual.

What this means

For policymakers, Schmidt’s comments reframe the debate. The question shifts from whether to stop toward how to set rules for a race that keeps running.

For investors watching the sector, the research findings note that a commitment to sustained innovation may encourage more investment by US AI companies and could draw venture capital toward startups and established firms. Firms that position themselves around credible safety measures may also gain partnerships and market confidence, though none of that is guaranteed.

Critics of Schmidt’s position would point out that a race without a pause leans heavily on safety mechanisms that are still unproven, including the very model-on-model oversight he proposed. Supporters would counter that an agreement nobody can verify may offer the appearance of safety while rewarding whichever side is most willing to cheat.