Coatlicue: Supercomputer Without Digital Foundations in Mexico

The unveiling of Coatlicue, the federal government’s new supercomputer, came wrapped in ambitious declarations and promises of national transformation. On November 26, 2025, President Claudia Sheinbaum introduced the system as “the technological brain that will place Mexico among the global elite.” With 15,000 GPUs, 314 petaflops, and an initial cost of 6 billion pesos, it’s an impressive machine on paper.

Coatlicue promises to predict hurricanes, detect tax fraud, improve agriculture, and train “sovereign” AI models. But if you’re familiar with the current state of Mexico’s digital infrastructure, the excitement fades quickly.

Because the problem isn't the supercomputer itself.

The problem is that Mexico lacks the minimum ecosystem needed to support it.

A Familiar Pattern: Grand Megaprojects Without Solid Foundations

Mexico’s recent history is filled with technological showpieces that collapsed under the weight of their own hype:

1. The Mexican Space Agency has yet to launch a single satellite.
   
   
2. The Dos Bocas refinery operates far below expectations, with massive cost overruns.
   
   
3. AIFA airport functions well below its projected capacity.

Now comes Coatlicue, another entry in a familiar pattern: build first, plan later, operate “as best as possible.” These projects come with cutting-edge speeches but lack clear governance models, operational rules, a scientific community, and the basic national tech infrastructure that can’t be replaced by raw petaflops.

Why Build a Supercomputer if the Country Lacks the Basics?

Before discussing high-performance computing, Mexico still faces critical foundational issues:

1. 30% of rural households lack fixed internet.
   
   
2. Most SMEs still don’t have access to fiber.
   
   
3. Half the population lacks basic digital skills.
   
   
4. There's a shortage of professionals in data science, cybersecurity, AI, or HPC.
   
   
5. Public infrastructure remains fragile and frequently targeted by cyberattacks.

In this context, Coatlicue risks becoming like parking a Ferrari on a dirt road—immense power, zero usability.

Read more: 500,000 Passwords and Sensitive Data of Mexicans Leaked on Telegram

The Problem Isn’t Power—It’s the Lack of Talent, Data, and Security

To make a supercomputer truly valuable, you need three things Mexico currently lacks:

1. Reliable, high-quality data
   Weather forecasting, precision agriculture, and fiscal modeling require massive, clean, trustworthy datasets collected systematically. That’s not the current reality.
   
2. Skilled researchers
   Years of budget cuts, lab closures, and forced talent migration have weakened Mexico’s scientific workforce. You can’t build an HPC community in two years.
   
3. Secure infrastructure
   Coatlicue will be a prime target for criminal groups, hacktivists, and state-backed actors. Mexico has no strong track record in protecting critical infrastructure. Without robust cybersecurity, even a top-tier machine can be taken offline in minutes.

The Most Alarming Part: No Rules, No Projects, No Scientific Community

While tech-forward nations assign HPC resources through expert committees and transparent protocols, Coatlicue was presented with:

1. Preselected beneficiaries
   
   
2. No public operational guidelines
   
   
3. No involvement from the scientific community
   
   
4. No assessment of real demand
   
   
5. No plan to develop talent

No concrete project has been announced to justify the scale of the system. Not even one. The government’s flagship use case? Tax fraud detection—something solvable with far simpler and cheaper tools.

An AI-Optimized Machine With No Real AI Projects

The government claims Coatlicue will be the heart of national AI sovereignty. But:

1. No national models have been announced.
   
   
2. No serious research program exists.
   
   
3. No data strategy has been presented.
   
   
4. There’s no academic-industry-government ecosystem to leverage it.

The system is optimized to train large-scale AI—but there are no models, no datasets, and no teams ready to use it. It’s like buying a rocket without any aerospace engineers.

More of a Modernity Simulation Than a Real Tech Strategy

The official narrative talks of a cutting-edge Mexico—but the truth is:

1. The country lacks consistent digital infrastructure.
   
   
2. There’s no national science or AI strategy.
   
   
3. Research centers are underfunded.
   
   
4. Talent leaves the country every year.
   
   
5. Public institutions are frequent targets of cyberattacks.

Coatlicue arrives as a political monument—not a component of a functional scientific ecosystem.

You might also be interested in: Mexico at a Crossroads: Build a Strong Cybersecurity Strategy

Can Mexico Make Coatlicue Work? Yes—but Not Like This

The supercomputer isn’t operational yet. But if the government wants to make it more than a propaganda tool in the next two years, it must:

1. Expand broadband access in rural and underserved urban areas
   
   
2. Set clear rules for operation with scientific involvement
   
   
3. Launch large-scale digital literacy programs
   
   
4. Create a national data strategy and modern AI framework
   
   
5. Strengthen critical infrastructure cybersecurity
   
   
6. Rebuild research centers and recover national talent
   
   
7. Develop an ecosystem where universities, researchers, and industry can use it

If none of that happens, Coatlicue will end up like many past projects: expensive, imposing, and ultimately useless.

Conclusion: Mexico Doesn’t Need Another Promise—It Needs Real Infrastructure

Technological ambition is admirable. So is aiming higher.

But we also need to be honest: Coatlicue won’t fix the country’s structural gaps. It merely puts them on full display.

Mexico didn’t need a 314-petaflop supercomputer. It needed internet in every town, better schools, more scientists, real cyber defense, and a coherent tech strategy.

Without that, Coatlicue is just another symptom of a deeper problem: the government’s obsession with appearing modern instead of building the foundations to be truly modern.

And for a country aspiring to lead in AI, science, and technology—that’s not just a mistake. It’s a setback.