The Most Important Lunch in AI — Altman, Amodei, Hassabis, and Mensch Meet G7 Leaders Today

The CEOs of the four most important AI labs on Earth are sitting down for a 2-hour lunch with G7 political leaders in Evian, France — today. The agenda: AI infrastructure, sovereignty, and regulation. The subtext: the Permission Layer is being negotiated in real time.

The Table — G7 AI Working Lunch

AI Lab CEOs:

Dario Amodei — Anthropic

Fable 5 still dark. Deadlocked with White House.

Sam AltmanOpenAI

$1T IPO. Partner network. S-1 filed.

Demis HassabisGoogle DeepMind

Gemini. Google’s AI strategy.

Arthur Mensch — Mistral

France just deployed Mistral to all civil servants.

Also at the table:

Marc Benioff (Salesforce) · Aidan Gomez (Cohere) · Alex Wang (Meta) · Robin Rombach (Black Forest Labs) · Pratyush Kumar (Sarvam AI) · Victor Riparbelli (Synthesia) · Ren Ito (Sakana AI)

+ G7 political leaders including Macron, Trump, and allies

Why This Lunch Matters More Than Any Product Launch

Five days ago, the US government pulled Amodei’s best model from every user on Earth. Today, Amodei is sitting across from the political leaders who ordered it — at the same table as the competitor (Altman) whose investor (Amazon) triggered it.

AMODEI’S POSITION

His best model is dark. His government pulled it. His biggest investor triggered the action. He needs this lunch to break the deadlock — or Fable 5 stays dead while competitors ship.

MENSCH’S POSITION

Europe’s moment. France deployed Mistral to all civil servants, kicked Palantir out of DGSI, committed €655M. Mensch arrives as the sovereign AI champion — the alternative to the US fence.

ALTMAN’S POSITION

$1T IPO pending. Partner network just launched. Needs the G7 to signal regulatory stability — any uncertainty hurts the IPO pricing.

GOMEZ’S POSITION (COHERE)

“Huge inbound” after Fable 5 recall. On-premise sovereign AI. Now at the G7 table making the case that alternatives to US-controlled AI exist and are ready.

The Geopolitical Fencing Table

This is the Geopolitical Fencing thesis made physical. The people who build the AI and the people who draw the fence — sitting at the same table, eating lunch, negotiating the rules that will govern the most powerful technology ever created.

The agenda — per Politico: Frontier AI risks, infrastructure, sovereignty, regulation, and child safety. But the real conversation is simpler: who gets to build, who gets to deploy, and who gets to say no? The Permission Layer is being drawn at this table.

What comes out of this lunch could reshape:

1

The Fable 5 deadlock — Amodei and US officials in the same room. A resolution path could emerge.

3

European sovereignty — Mensch + Gomez making the case for non-US AI infrastructure at the highest level.

Business Engineer

The Geopolitical Fencing of Frontier AI

The framework that predicted this table — and the analysis of what happens when the fence-builders and the fenced negotiate face to face.

Read the full analysis →

The Bottom Line

The four CEOs who control frontier AI are having lunch with the political leaders who control the fence. Five days after the biggest AI shutdown in history. At a G7 summit in France — the country that just kicked Palantir out of its spy agency and committed €655M to sovereign AI. This isn’t a tech conference photo op. It’s the Permission Layer being drawn over a 2-hour meal.

Sources: CNBC, Politico

Scroll to Top

Discover more from FourWeekMBA

Subscribe now to keep reading and get access to the full archive.

Continue reading

FourWeekMBA