Mexico IA+ and the Plan Behind Its AI Factories

Earlier this week, I had the privilege of attending the inauguration of Mexico IA+ in Mexico City. Key figures from government and industry were present, including Rosaura Ruiz Gutiérrez, Mexico’s Secretary of Science, Humanities, Technology, and Innovation; Calista Redmond, NVIDIA’s Vice President of Global AI Initiatives; and Rodrigo Amodio, a member of the Mexico IA+ board. Together, they outlined some of the ideas shaping the development of Mexico’s artificial intelligence (AI) infrastructure.

The conversation made one thing clear: Mexico is entering a new phase in its AI infrastructure journey. I’ve written before about different aspects of AI sovereignty — particularly how countries in the Global South can finance it. But today there’s a concept worth highlighting: AI factories.

So, will these factories be producing robots and letting them stroll across Mexico as they come off the assembly line? Fortunately, no. Anyone following the evolution of robotics knows they’ve come a long way, though they’re still not anywhere near that level of production. Just look at what we saw this week in Moscow with AIdol. This advanced robot walked onstage trembling, tried to raise its hand, and ended up flat on its face in front of the audience. The most memorable part? It couldn’t stand back up.

The Factory Where the Product Is Intelligence

So then, what exactly is an AI factory? Just as a traditional factory takes in raw materials and outputs a range of finished products, an AI factory does the same — only with data. These are specialized infrastructures that transform raw data into scalable, robust intelligence aligned with a country’s norms, laws, contexts, and languages.

Why does this matter? Because as AI becomes embedded in more tasks, products, services, and parts of society, people need access to local information that reflects their lived reality.

AI factories make this possible by managing nearly the entire lifecycle of the AI process within a country’s borders — from the advanced hardware powering the system, to the collection and processing of data, the training of models, and the monitoring of what those models produce over time. This gives the country developing the factory — in this case, Mexico — greater control over the AI backbone (its models and outputs) that supports its institutions, economy, and society.

It’s important to clarify that these factories do not produce semiconductors or hardware for AI infrastructure. Their role is to refine data and generate the models and outputs I’ve mentioned. The word “factory” often gives the impression that something physical is being manufactured, but in reality, everything here is digital. The semiconductors will continue to be imported, since they’re designed by NVIDIA and manufactured by Taiwan Semiconductor Manufacturing Company (TSMC). Which brings us to another key point.

The Next Chapter in Mexico’s AI Story

There’s no doubt that what Mexico is building is a hybrid form of AI sovereignty — and that’s entirely reasonable and strategically sound. Very few countries outside the United States and China can aspire to full AI sovereignty, meaning control over virtually the entire supply chain from chips to algorithms. And even those two nations rely on each other for essential materials and frontier research, to name just a couple of examples.

Mexico made a smart decision by taking this first step into its AI future. By investing in AI factories that will produce high-quality models, the country is creating the foundation for what comes next. And let’s be honest: it’s also comforting to know that, at least for now, we won’t have AI robots wandering the streets or lying face-down in the office.

Originally published in Spanish for Fast Company Mexico:
https://fastcompany.mx/2025/11/18/mexico-ia-plan-fabricas-inteligencia-artificial/