The World Is Not Multipolar. It Is Pluripolar.

There is a particular strangeness in being courted to embrace a product you have just been forbidden to touch. That was the position of the world’s richest democracies at Évian this week: hosts to a summit whose star guests were the men who build frontier artificial intelligence, and whose most powerful systems had been pulled offline in their faces days before. Power, the chancelleries keep telling us, has gone multipolar. What happened in France suggests the truth is stranger, and more useful, than that tired word allows.

When the Group of Seven gathered in Évian-les-Bains on 17 June, the headline was not a head of state but a working lunch. Sam Altman of OpenAI, Dario Amodei of Anthropic, Demis Hassabis of Google DeepMind and Arthur Mensch of Mistral sat among a dozen technology chiefs and the leaders of the G7. Days earlier, Washington had imposed export controls on Anthropic’s most advanced models, Fable 5 and Mythos 5, and the company had switched them off worldwide to comply. The stated concern was security: Mythos has shown an unusual facility for finding and exploiting flaws in software, the sort of capability far more dangerous in the wrong hands than valuable in the right ones. The staging implied a coronation of American technological primacy. The substance implied something stranger.

This is the texture of pluripolarity, and it is worth distinguishing the term from the multipolarity that fills the speeches of Paris, Berlin and Beijing. Multipolarity asks a single question: how many great powers sit at the top table? It produces one ranking and treats that ranking as the structure of the world. Pluripolarity makes a humbler and more accurate observation: the ranking changes with the domain. The United States is close to a sole superpower in frontier artificial intelligence, one power among several in advanced manufacturing, a structural debtor in others, and merely a large customer in rare earth processing, where Beijing sets the terms. There is no master league table. There are many, and they do not align. Power has simply stopped lining up.

Israel lives this every day. It is a near superpower in cyber and artificial intelligence and a small state almost everywhere else, dependent on Washington for some capabilities and fiercely sovereign in others. It has built a strategic identity precisely out of this asymmetry, declining to be ranked on any single axis. And it has learned, at great cost, what the rest of the G7 discovered only last Friday: that access granted by a patron is an instrument, and instruments can be withdrawn.

The lesson has a date. By 1967 France was Israel’s principal arms supplier, the source of the Mirage fighters that formed the backbone of its air force. Then, as the Six Day War broke out that June, Charles de Gaulle imposed an embargo. Fifty newer Mirage jets that Israel had ordered and paid for were impounded in France and never delivered. The embargo did not break Israel. It pushed the country to build its own defence industry, to reverse engineer the aircraft it had been denied, and to deepen the relationship with the United States that has anchored it since. A supply that felt permanent proved revocable on a single statesman’s whim, and Israel responded by ensuring it would never again hold so few of its own keys. Every government locked out of the American technology stack this week is now learning the same arithmetic, several decades late.

That is why Amodei’s plea will struggle to hold. He urged the room, by several accounts, to resist the temptation to splinter, and called for a coalition of democracies, led by Washington, that would share frontier models and trade the necessary components among themselves while shutting out adversaries. The instinct is sound, and for Israel the stakes are immediate: a tier of trusted partners is now being sketched by American officials and European diplomats, and the question of who sits inside that tier and who is left outside is, for a state whose security rests on technological edge, close to existential. Yet a coalition that can lock out a rival today can lock out a partner tomorrow, and prudent states plan for both.

Altman warned the gathering that the gravest danger of the technology is its tendency to concentrate power in very few hands, adding that history is clear this does not work. He meant founders and firms. The warning applies with greater force to states. A world that is genuinely plural in trade, energy, finance and arms will not long tolerate a single dominant layer sitting on top of everything else and governing the substrate on which the others increasingly run.

So pluripolarity is not disorder. It is the misalignment of orders, and the friction at Évian was simply one domain’s master trying to discipline another domain’s coalition. For Israel the counsel that follows is the one it wrote for itself in 1967: welcome every alliance, depend on none of them wholly, and keep your own keys. The states that learned the lesson late spent last week discovering, in real time, what it costs to hold someone else’s. Jerusalem learned it long ago, and has spent half a century making sure it never has to learn it twice.