Political risk / Jul 6, 2026 / 4 min
Washington Skipped Geneva. Silicon Valley Didn't.
On July 6, 193 UN member states opened the inaugural Global Dialogue on AI Governance in Geneva while the Trump administration publicly rejects multilateral rulemaking — yet Jensen Huang, Brad Smith, and Jack Clark are formal members of a UN commission meeting there Wednesday, proving frontier AI's real governors showed up without Washington.
On July 6, Geneva opened the UN's inaugural Global Dialogue on AI Governance — 193 member states at one table — while the Trump administration has spent months saying it "totally rejects" global AI rulemaking. The CEOs who actually build frontier models didn't get the memo: NVIDIA, Amazon, Microsoft, and Anthropic are formal members of a UN commission convening in the same building Wednesday.
What's happening in Geneva:
- July 6–7: Global Dialogue on AI Governance — all 193 UN member states, non-binding, co-chaired by El Salvador and Estonia
- July 7–10: AI for Good Global Summit at Palexpo
- July 8: Inaugural session of the AI for Good Global Commission — 44 members including heads of state and frontier-lab CEOs
What Washington said no to:
- At the India AI Impact Summit, White House science adviser Michael Kratsios declared: "We totally reject global governance of AI."
- He named the UN Dialogue specifically, calling such forums an "atmosphere of fear" and warning that safety-focused policy "inhibits a competitive ecosystem, entrenches incumbents, and isolates developing countries."
- State Department Under Secretary Jacob Helberg told the Atlantic Council the administration is "highly skeptical of supranational bodies in the business of governance" — while pitching Pax Silica, a U.S.-led alliance where 35 countries signed a declaration on AI opportunity in June.
- At a UN Security Council debate, Kratsios rejected "all efforts by international bodies to assert centralized control and global governance of AI."
Who showed up anyway:
- The ITU launched the AI for Good Global Commission on July 2 — the first UN-mandated body where frontier CEOs are full members, not observers.
- Jensen Huang (NVIDIA), Andy Jassy (Amazon), Brad Smith (Microsoft), Jack Clark (Anthropic), and James Manyika (Google) sit alongside Rwandan President Paul Kagame, Estonian President Alar Karis, and UN agency chiefs.
- Co-chaired by Marc Benioff and Kagame. Benioff told Axios the body brings together "the people who build AI, deploy it, shape policy, and represent communities."
The science on the table:
- On July 1, the UN's 40-scientist Independent Panel warned AI capabilities are "outpacing both scientific understanding and governments' ability to adapt."
- Co-chair Yoshua Bengio: "Science currently cannot guarantee that as capabilities continue to increase, AI will not cause catastrophic harm."
- Co-chair Maria Ressa: "If you can't tell fact from fiction, you cannot have a democracy."
- The panel found the U.S. controls 75% of computing power among the world's top 500 AI supercomputers; China holds 15%. Virtually all frontier models come from those two countries.
The developing-world pitch:
- Estonian co-chair Rein Tammsaar told UN News AI could be "a great equalizer" — or, in the wrong hands, a tool to "erode trust in governments."
- El Salvador's Egriselda López called the AI divide "real": some nations have infrastructure and research capacity; others still lack connectivity.
- China backs multilateral governance. Washington backs the American AI stack. Both are recruiting the Global South.
The parallel track in D.C.:
- While Geneva talks inclusion, the White House is finalizing voluntary frontier-model standards with OpenAI, Google, and Anthropic — a guest-list regime with 30-day review holds, not a treaty.
- OpenAI is separately negotiating a 5% equity donation to a national wealth fund. That's political armor before IPO roadshows, not public ownership.
Why CEO seats matter — and don't:
- Bletchley Park in 2023 was government-dominated; industry attended as guests. This commission seats builders as members.
- But multistakeholder bodies produce personal endorsements, not binding corporate commitments. Huang at the table doesn't obligate NVIDIA to implement anything.
- CSIS analysts say Washington's rejection leaves a leadership vacuum China is positioned to fill — while U.S. tech giants keep seats at the UN table anyway.
Convina's view: Washington wants AI governance to look like Pax Silica and voluntary IPO-season equity deals — bilateral, non-binding, and written by the labs that benefit. Geneva wants 193 nations to agree on guardrails before Bengio's catastrophic-harm scenarios stop being speculative. The absurdity is that both conversations are happening in the same convention center, and only one includes the people who can actually flip the kill switch. D.C. rejected the table. Silicon Valley took the chairs.