AI leaders call for US-China deal as startups build verification tools

By Splitfeed EditorialLast updated Sep 26, 2026, 19:17 UTC

On September 23, 2026, AI industry leaders publicly called for a bilateral agreement between the United States and China on artificial intelligence, coinciding with a summit involving President Trump and Chinese President Xi Jinping. The Washington Post reported the same day that, despite American tech leaders being invited to a dinner with Xi, the White House has shown little interest in pursuing a comprehensive AI deal. Separately, the South China Morning Post covered startups that are already building verification tools designed to confirm compliance with any such agreement, framing this technology as a practical prerequisite for any enforceable arrangement between the two countries.

The timing matters because it places private industry advocacy in direct tension with government reluctance. Verification has historically been one of the hardest problems in arms-control style agreements, and the emergence of startups positioning their tools as a solution signals that at least some in the private sector see a commercial and strategic opportunity in the governance gap. Whether verification technology can actually resolve the trust deficit between Washington and Beijing is a live question, and the answer shapes whether any deal is conceivable at all. The competitiveness dimension adds another layer: restricting or coordinating AI development between rivals risks disadvantaging whichever side moves more cautiously.

The Middle Ground camp, which produced the only quoted reactions to this story, reflects the tension clearly. The Washington Post account noted the White House's lack of interest in a deal even while hosting Chinese leadership. Rep. Adam Smith called for Trump and Xi to discuss controlling AI in what he described as a more intelligent way during the summit. CSIS observed that whatever emerges from the summit will likely serve as leverage in the broader institutional contest over AI governance rather than producing a genuine agreement. Brookings raised the competitiveness question directly, asking whether policymakers can slow advanced AI development without compromising American standing. Tech Policy Press offered a framing that cuts across both concerns, arguing that dominance will belong to whoever achieves the most adopted model, not merely the most powerful one. Pro-AI and Anti-AI camps have not yet reacted publicly; the former would typically argue that voluntary coordination between leading powers accelerates safe deployment, while the latter would likely warn that any deal giving China legitimacy in AI governance carries serious risks.

The clearest near-term signal to watch is whether the Trump-Xi summit produces any joint statement or framework on AI, even a symbolic one, and how the White House characterizes any such outcome in the days that follow.