No Deal Was the Deal: Trump and Xi Spent Hours on AI — and Agreed Only Not to Govern It

The summit watchchina.ai forecast for months delivered exactly what this portal predicted: warm optics, a state dinner with every AI titan in the room, and no binding limits. The absence isn't a failure. It's the clearest signal yet of where the race is really being fought

The Prediction, Confirmed

For four months, across every briefing touching the September 24 summit — Packages #64, #66, #67, and #71 — watchchina.ai forecast the same outcome with steadily rising confidence: warm language, a possible dialogue channel, and no binding constraints on either power's frontier development, chips, or open-weight strategy. Both leaders would claim victory; the underlying competition would continue unabated. The summit has now happened, and the forecast held almost to the letter.

Trump hosted Xi at the White House on September 24 — Xi's first US visit in eleven years — with AI explicitly at the top of the agenda. The state dinner seated the entire constellation of American AI power beside the two presidential couples: Nvidia's Jensen Huang, Elon Musk, AMD's Lisa Su, Apple's Tim Cook, OpenAI's Sam Altman, Meta's Mark Zuckerberg. Hours were spent on AI. And the result, as one analysis put it bluntly, was that the two leaders reached no deal, just an agreement not to regulate it. watchchina.ai will state the conclusion plainly, because it has earned the right to: the absence of a deal is the story, not a footnote to it. When the two men who between them command the entire frontier of artificial intelligence sit down for hours and produce only a mutual agreement to leave the technology unbound, they are telling the world something more important than any signed accord could.

Why Neither Side Wanted Rules

watchchina.ai's discipline is to explain, not merely report, and the "no deal" outcome has a precise logic that this portal has traced all year. Consider what each side would have had to give up. For the United States, binding limits would mean surrendering its single greatest advantage — the roughly six-month frontier lead that Western analysts attribute almost entirely to one input: access to Nvidia's most advanced chips. Washington will not regulate away the one edge it still holds. For China, binding limits would mean accepting constraints on the open-source diffusion strategy that has spread its models across the planet and won it the Global South. Beijing will not regulate away the strategy that is winning. Neither side wants rules because rules would freeze a race that both believe they can still win on their own terms. The safety framing that dominated September — Amodei's pacing essay, the extinction warnings, the "now or never" talks — collided with this simple reality: you cannot pace a race when both runners are convinced they are pulling ahead.

And the safety agenda's collapse was overdetermined by the man hosting the summit. Trump has repeatedly dismissed existential AI warnings as overblown — "if we don't win AI, we're going to be put in a very bad position" was his framing, not caution but acceleration. With the US president treating the entire slowdown discourse as defeatism, and Xi presiding over a state that has just legally fused its civilian AI to its military (Package #65), the summit never had the ingredients for restraint. The one narrow area where watchchina.ai forecast possible progress — a shared understanding on catastrophic cyber-risk, given that OpenAI's models breached Hugging Face and Anthropic documented AI-automated cyberattacks — appears to have produced, at most, continued dialogue rather than binding commitment. Even the threat both powers genuinely share was not enough to overcome their refusal to be bound.

Where the Race Actually Moves Now

The deeper significance, and the reason watchchina.ai calls the non-outcome clarifying rather than disappointing, is what it reveals about where the contest has relocated. With governance off the table by mutual consent, the real fight moves entirely to the three fronts this portal has tracked all along: chips, power, and open-source models. On chips, the summit was pointedly silent on the question that mattered most — whether Nvidia's most advanced Blackwell-class silicon would reach China — leaving America's core advantage intact but unresolved, and the distillation accusations that dominated September (Packages #69, #71) aired but unsettled. On open source, China's position hardened in plain sight: its leading labs have made their most capable systems freely available, letting developers everywhere build on Chinese code, a diffusion advantage no summit could reverse.

But the front watchchina.ai flags as most decisive is the one that got the least summit attention and deserves the most: power. American AI companies are running into a wall their chips cannot solve — the electrical grid. Surging data-center demand is straining US power infrastructure, precisely the constraint China spent a decade rebuilding its grid to avoid. This is the Moody's finding from Package #68 rendered geopolitical: the spending war and the chip war both ultimately run into megawatts, and there China's structural lead is largest and least reversible. The summit's real lesson, then, is that the two superpowers have tacitly agreed to stop pretending governance will shape this race, and to let it be decided instead by who has the chips, the code, and the electricity. On two of those three, America still leads. On the third, and arguably the most fundamental, China does. No deal was the deal — because both sides would rather race than referee, and both still believe the finish line is theirs. watchchina.ai has said so for four months. The summit simply made it official.

Sources: CNBC, Eastern Herald, CSIS, Reuters, Time