Why Silicon Valley's Bosses Now Sit at the Top Table for US-China Talks

The Head Table Isn’t About Etiquette

It never was, not when you look closer.

The White House state dinner on September 24 placed Donald Trump and Xi Jinping at the center of the head table — but flanking them were people who don’t typically appear on diplomatic guest lists: Jensen Huang, CEO of NVIDIA; Lisa Su, CEO of AMD; Elon Musk; and Tim Cook.

What you’re seeing isn’t a coincidence of protocol. It’s a seating chart that reads like a structural reassignment of power.

The Chips That Matter

Huang and Su are the most consequential names in that arrangement.

NVIDIA and AMD control the vast majority of the AI accelerator market — the GPUs that power large-language models, training clusters, inference engines. They are also the companies most directly impacted by US export controls on advanced semiconductor sales to China. Every restriction written in Washington echoes through their revenue projections, their inventory pipelines, their design schedules.

That dual position — technology supplier and compliance subject — is exactly why they deserve the seat. The alternative would be talking about the consequences without the people who can quantify them.

Lisa Su reinforced this before the dinner. She told reporters that “technology is a field where the global ecosystem must move together,” a statement that sounds diplomatic but is really a plea for access to the Chinese market and the supply chains embedded there.

Huang, for his part, was even more direct in pre-dinner remarks, describing the current regulatory environment as “unnecessary friction” that slows deployment without meaningfully reducing risk. His comments drew immediate attention from trade policy circles in both capitals.

What the Talk Produced

The outcome was narrower than many expected.

There was no easing of the AI chip export controls. Reuters reported that the US is likely to maintain the restrictions China has been pressing to lift. Instead, the two sides created a formal dialogue on “super intelligence” — a term both governments agreed to use rather than “AI” — and committed to a follow-up meeting before November. They also established a direct government-to-government communication channel for incidents involving super intelligence systems.

Notably absent from the final statement: any language about AI speed limits or regulatory coordination. The “AI deceleration” argument pushed by some in the industry did not make it into the text.

For NVIDIA, that silence carries weight. The company’s AI chip sales scale with infrastructure buildout velocity. Slower global deployment means slower demand growth. Huang made his position clear in a CNN interview before the dinner: regulation isn’t what the industry needs now, he said. What’s required is technology to monitor AI itself.

When asked directly whether regulation was necessary for the current situation to proceed correctly, Huang answered simply: no.

The failure to include AI safety language in the joint statement suggests that the Chinese side, which has been more aggressive in its domestic regulatory posture, did not push hard for it — or that the Americans declined to concede. Either reading favors the industry’s immediate commercial interests, if not its long-term strategic positioning.

Who Wasn’t Invited

The absence of Chinese business leaders is as telling as the presence of American ones.

Chinese companies were reportedly considered for the delegation but ultimately excluded, according to Reuters, citing US concerns over some Chinese firms’ military-related activities. That leaves a gap in representation — the very companies affected by these export controls had no voice at the table.

Huawei, SMIC, and Baidu’s AI division were among the notable omissions. Their exclusion underscores a fundamental tension: the US wants to constrain China’s AI advancement while simultaneously engaging with Chinese political leadership. Leaving out the private sector actors who are the actual conduit of that technology transfer weakens the dialogue’s credibility.

On the American side, the list stretched beyond the head table: Sam Altman, Greg Brockman, Mark Zuckerberg, Sundar Pichai, Satya Nadella, Jeff Bezos, Qualcomm’s Cristiano Amon, Micron’s Sanjay Mehrotra. Anthropic’s Dario Amodei — a vocal advocate for AI pause and regulation — was notably absent.

Financial executives from JPMorgan and BlackRock were present. This wasn’t just a tech dinner. It was a full-spectrum briefing on the economic stakes.

The Supply Chain Reality

Musk and Cook sit at the head table not only for their companies’ technological relevance, but for their operational dependence on China.

Tesla operates a major factory in Shanghai that Musk called a “jewel” in a CCTV interview the day before Xi’s arrival. Apple generated over $64 billion in China-related revenue last year, with significant portions of its manufacturing supply chain still rooted in Chinese production facilities.

Their seats reflect a simple fact: Silicon Valley’s greatest companies cannot decouple from China overnight, and the people who understand that constraint best belong in the room where those constraints are discussed.

The second-order effect here is significant. When the CEOs of the world’s most powerful technology companies are personally invested in maintaining access to the Chinese market, they become a lobbying bloc with incentives that may diverge from US national security objectives. Export controls that make strategic sense from a defensive standpoint can appear as commercial headwinds to the executives living with them daily. That tension — between security logic and market logic — will now have a direct line to the negotiating table.

What This Means Going Forward

The seating arrangement signals a shift in how technopoly power is calibrated. Diplomatic protocol once reserved the head table for foreign ministers, trade negotiators, national security advisors. Now it goes to the people who control the hardware the competition runs on.

This doesn’t mean export controls will ease. The consensus in Washington remains firm. But it does mean the calculus behind those controls will now include the direct input of the companies most affected — and those companies will have a platform to argue their case.

The next meeting is set for November. The communication channel between governments is active. The question isn’t whether this structure persists, but whether it becomes the default model for how major tech economies negotiate around strategic technologies.

If the answer is yes — and the evidence from this dinner suggests it will be — then the next time two superpowers sit down to discuss something that reshapes the world, the guest list will look very different from the one that signed the treaties of the twentieth century.

The chips are on the table now. So are the people who make them. And their presence changes what the conversation is actually about.