Javier Milei, the president of Argentina, would like Argentina to be an AI hub, and his strategy is to have no rules. In a Financial Times op-ed co-written with his deregulation minister, Federico Sturzenegger, and headlined “Argentina invites AI to free itself,” Milei built the pitch on three things: a commitment not to regulate AI, a low corporate tax rate, and a new kind of company. He called Argentina the world’s fourth AI hub, which is an ambitious ranking for a country that is not yet any kind of AI hub, but the ranking is the recruitment pitch.
The strategy is regulatory arbitrage, said out loud. Milei is not dressing it up as talent or research. The product is the absence of regulation, and an adviser told the FT that the appeal is “cover” against rising regulatory risk in the US and Europe, which is an unusually frank way to describe an industrial policy: come here because the rules you dislike do not apply. The United States is building a federal testing regime for frontier models. Europe is attaching its gigafactory money to rules. Argentina is selling the lack of both as the feature.
The most striking piece is the new corporate form. Milei wants to create the “non-human corporation,” a company operated entirely by AI agents, with human shareholders optional rather than required. The justification is liability. “Limited liability is not a luxury for such entities; it is a precondition for their existence,” he wrote, which is a way of saying that a company run by software should be a company nobody is quite responsible for. Argentine politicians and technologists have raised the obvious concern, that an unregulated, lightly taxed firm with no humans in it is hard to hold accountable for anything. It is one part of the framework rather than the whole of it, but it captures the spirit, which is to remove friction, including the friction of accountability.
Milei has been courting in person. In May he met Sam Altman and Tim Cook in California and hosted a summit with Marc Andreessen. The framework is a bill before Congress rather than law, and it runs alongside a separate Milei effort to allow unlimited foreign purchases of Argentine land, so the broader project reads less as an AI strategy than as a general invitation to bring money and few questions.
It has produced one prominent arrival. Peter Thiel, last seen chasing offshore sites for data centers, bought a $12 million mansion in Buenos Aires this spring, moved his family, and enrolled his children in school there. He has met Milei at the presidential residence and the same deregulation minister who co-wrote the AI op-ed. Thiel’s reasons, per the New York Times, are taxes, the threat of nuclear war, and the risks of AI, which makes him an AI investor relocating partly to get away from AI, to the country marketing itself as the place with the fewest constraints on it. Argentina has reportedly explored offering him citizenship.
The pitch assumes AI companies want less regulation, which is true of some of them and conspicuously untrue of others. Anthropic has spent the past year asking governments for more oversight rather than less. It argued for tighter controls and harder coordination, it supported the White House plan to test frontier models before release, and it recently published its own warning that AI is improving fast enough to be worth pausing over. A lab that has been lobbying for rules is not an obvious customer for a country selling their absence.
How that plays out is an interesting question. A country with no rules may still attract plenty: the firms that treat regulation as pure cost, the agent startups looking for somewhere friendly to incorporate, the capital that wants low taxes and a government that returns its calls. Whether it also draws the companies building the most capable systems, the ones now asking to be regulated, is far less certain. Milei has turned Argentina into a fairly clean experiment in what the AI industry wants when it is offered the chance to have things entirely its own way.
–
By the Control Plane Editorial Team