Argentina Steps Into the Regulatory Vacuum
On June 4, 2026, President Javier Milei and Deregulation Minister Federico Sturzenegger published a co-authored Financial Times op-ed announcing Argentina as Latin America's AI hub. The pitch rested on three pillars: zero AI regulation, a new "non-human corporation" legal category, and the continent's most competitive fiscal environment. It was not purely rhetorical. Two weeks earlier, the government had submitted a draft General Companies Law (INLEG-2026-53661873-APN-PTE) to the Senate that would replace corporate legislation from 1972. And anchoring the bid: a signed letter of intent for a $25 billion, 500-megawatt OpenAI data center in Patagonia — the first Stargate project in Latin America.
The RIGI Makes the Investment Case Real
Argentina's RIGI framework (Law 27,742, enacted July 2024) offers qualifying investors a 25% corporate tax rate against the standard 35%, duty-free capital-goods imports, and a 30-year guarantee of fiscal and regulatory stability. A May 2026 upgrade — the "Super RIGI" — cuts the rate further to 15% for mega-projects, with AI data centers explicitly listed as qualifying. The OpenAI–Sur Energy deal is structured under RIGI precisely for these terms. Patagonia's cheap renewable energy (wind and hydro) reduces both power and cooling costs substantially. Sam Altman described it as "our first Stargate project in Latin America, a region full of talent, creativity and ambition." For infrastructure capital with a multi-decade horizon, a stability guarantee that cannot be legislated away for 30 years is worth more than headline rate reductions alone.
The "Non-Human Corporation" — Overstated but Not Baseless
The legally novel element is the proposed Automated Company: entities operated by AI agents where human shareholders are optional. Milei framed it as the AI era's equivalent of the Dutch East India Company's 1602 invention of limited liability, calling for Buenos Aires to become what Amsterdam was to the age of sail.
Historian Yuval Noah Harari published a direct FT rebuttal, arguing that granting legal status to AI-operated entities gives them "a master key" to financial, economic, and political systems — and that unlike human executives, AI agents cannot be deterred by imprisonment. "Countries that grant legal personhood to AIs risk becoming an AI-state," Harari warned. Microsoft AI CEO Mustafa Suleyman endorsed the concern. These are serious objections: an autonomous legal entity that can own assets, sue in courts, and participate in commerce without any human bearing criminal accountability is a genuinely novel governance challenge.
But legal analysis of the actual bill text shows it is more careful than the political framing suggests. INLEG-2026-53661873-APN-PTE does not constitute AI as an independent legal subject. The company remains the legal person, director liability is preserved, and human anchors remain in the corporate structure. The AI operates the company; it does not own it in any legally autonomous sense. Milei's rhetoric significantly oversells the legislative radicalism. A sensible legal wrapper for algorithmically managed companies is not the same as granting AI personhood — and it is worth separating the two before policy responses are calibrated.
What the Deregulation Bet Omits
The zero-regulation posture creates vulnerabilities the investment pitch deliberately glosses over.
Argentina currently has no dedicated AI law. The country relies on the Personal Data Protection Law No. 25,326 — enacted in 2000, before large language models existed — supplemented by non-binding guidance from the Argentine Data Protection Authority (AAIP) on "transparency, lifecycle accountability, and bias mitigation." Multiple AI governance bills were introduced in Congress in 2024 and 2025; none were enacted. Without data-protection rules aligned to modern AI architectures, multinational companies operating under EU GDPR obligations will face legal uncertainty on cross-border data flows — a significant constraint for any AI ecosystem extending beyond pure compute infrastructure.
The Automated Company framework also leaves liability questions unresolved. When an AI agent makes a decision that causes economic harm, which human principal bears responsibility and under what standard? The 1972 corporate law structure being replaced was not designed to answer this, and the draft bill's liability provisions have drawn scrutiny from Argentine legal academics. Without algorithmic transparency requirements, consumers and regulators have no visibility into how automated decisions are made — a gap that will complicate any future digital trade agreement with the EU or UK.
The Long Game
From a policy-competition standpoint, Argentina's strategy makes a recognisable bet: lower the regulatory floor, offer fiscal stability, and draw investment that would otherwise face compliance burdens in US or EU jurisdictions. The EU AI Act — in full application since 2025 — imposes real conformity-assessment requirements on high-risk systems. For certain enterprise applications, Argentina's framework offers a genuinely lower-friction deployment environment, and that is a real competitive differentiator.
But the countries that have durably attracted technology investment — Singapore, Estonia, Ireland — did so through regulatory clarity, not regulatory absence. They built predictable frameworks specifying exactly what rules applied, what liability investors faced, and what institutions would adjudicate disputes. Tax benefits can initiate capital flows; they cannot by themselves sustain a technology economy.
The $25 billion Stargate data center demonstrates that large-scale infrastructure capital finds the Argentine proposition credible. Whether that translates into a broader AI economy rather than an isolated compute enclave will depend on decisions the Milei government has so far declined to make: a modernised data protection framework, a clear liability regime for automated companies, and baseline transparency standards that let trading partners and investors trust what is built here.