México’s Coatlicue Supercomputer Embraces Spectacle Over Sober Scientific Dreams Ahead

México’s promised Coatlicue supercomputer may never live up to its scientific hype. However, the political spectacle around it reveals how grand technological myths still organize power, hope, and regional status in Latin America’s digital age, and may be precisely what México needs.

Spectacle As Engine of México’s Digital Sovereignty

The morning announcement by Coatlicue during the presidential press conference was not just a technical briefing; it was a performance. There, flanked by numbers and acronyms, officials described a machine that most citizens will never see. Still, many will imagine: around 15,000 GPUs, the equivalent of some 375,000 conventional computers working at once, and a projected performance of 314 petaflops, or 314,000 trillion operations per second. According to José Antonio Peña Merino, head of the Agencia de Transformación Digital y Telecomunicaciones, this supercomputer, framed as part of Plan México, would be built in 24 months with some 6 billion pesos in public funds [EFE].

On paper, the comparisons sound glorious. Coatlicue would be seven times faster than Pegaso, Brazil’s biggest private system, and more than 100 times as powerful as Yucca in Sonora, currently México’s most advanced public machine. It would leap over Europe’s Leonardo system at about 250 petaflops, inch toward the 435 petaflops of Alps in Switzerland, and situate the country within shouting distance of the global elite, still dominated by giants like Frontier in the United States, surpassing 1.1 exaflops.

In strict scientific terms, there are obvious questions. No confirmed site. No public tender documents. No clear hardware supplier. At this stage, Coatlicue is more narrative than infrastructure. But that is precisely why it matters, and why the spectacle deserves to be defended.

Research on “technonationalism” in journals like Science and Public Policy has long argued that states do not build massive computing centers solely to crunch equations. They make them tell a story: that the nation is modern, sovereign, able to speak its own language in the code of the twenty‑first century. When Rosaura Ruiz, head of the Secretaría de Ciencia, Humanidades, Tecnología e Innovación, declares that the project is “a huge step” and will lead a national supercomputing network spanning UNAM, IPN, Cinvestav, and state universities, she is not just promising capacity; she is proclaiming status.

And status matters. For a region often treated as a mere “data provider” to platforms headquartered elsewhere, a public supercomputer, however unfinished, signals that México is not resigned to renting its digital future by the hour from foreign clouds. Seen from this Latin American vantage point, the announcement’s political flamboyance is not a bug; it is the motor. Without spectacle, there is no budget, no attention, no sense of urgency.

Latin American Megaprojects and the Politics of Overpromising

Of course, we know the script of the counterargument. In recent years, the Cuarta Transformación has been surrounded by megaprojects that have become synonymous with controversy: the Maya Train, the Dos Bocas Refinery, and the Felipe Ángeles Airport. Overruns, opacity, and underwhelming early results have fed the familiar narrative of a state that builds monuments to itself rather than functional systems for its citizens.

Placed in that lineage, Coatlicue looks like another grand promise walking toward the same trap. The second layer of “what if” practically writes itself. What if this supercomputer is installed, but poorly managed? What if politically favored institutions monopolize access? What if the machine ends up underused, a cold, humming artifact in a guarded data center while researchers in regional universities still beg for bare servers? Comparative studies in the Journal of Latin American Studies have repeatedly shown how regional megaprojects can become precisely that: showpieces of modernity whose daily impact remains modest at best.

There is also the more unsettling question of surveillance. México has a documented history with digital espionage, from the Pegasus spyware scandal onward. A system capable of processing vast volumes of data, tax records, customs flows, telecom metadata, could, in the worst hands, reinforce state control rather than democratic oversight. As scholars in AI and Society point out, the same high‑performance computing that trains language models can also optimize forms of social monitoring.

And yet, even acknowledging those risks, the temptation to retreat into a purely “honest scientific ambition” is its own kind of fantasy. Latin American science has rarely been funded solely based on its technical merit. It has advanced when it could hitch itself to louder political projects: satellites that symbolized independence, observatories that turned remote deserts into “windows to the universe,” and health institutes rebranded as national shields. The pattern is uncomfortable, but it is real.

From that perspective, the overpromising around Coatlicue is not a distortion of the game; it is the entry ticket. By promising to train massive language models in Spanish and indigenous tongues, to run climate simulations that inform water policy, to accelerate genomic analysis for public health, the project dresses its silicon in a mythology that voters, legislators, and foreign observers can recognize. Without that mythology, the same 6 billion pesos would likely evaporate into less visible line items.

When Mythic Machines Matter More Than Measured Results

There is a reason this machine is called Coatlicue, after the Nahua deity of creation and destruction. The name folds together terror and possibility, ancient stone and future code. That choice, highlighted by President Claudia Sheinbaum, who insists it be a “supercomputer of the people of México,” is not accidental. It invites citizens to see themselves reflected in a rack of 200 water‑cooled cabinets and 80 to 100 engineers watching over them. It transforms a data center into a national altar.

For technocrats, this can feel intolerable. The numbers, they insist, do not add up. How can Alps in Switzerland deliver 435 petaflops for around 100 million Swiss francs, roughly 2.27 billion pesos, while Coatlicue projects lower performance for nearly 6 billion pesos? Shouldn’t the priority be to optimize cost per flop, benchmark against international rankings, and follow the rational blueprint popularized in Northern think‑tank reports?

But that misunderstands the region’s political economy. In Latin America, a supercomputer is never just a machine; it is a stage. It exists to show that the state has not surrendered entirely to foreign platforms, that public universities still have a seat at the technological table, that the country can, at least symbolically, stand next to Frontier and Leonardo rather than eternally behind them. Spectacle is not the enemy of science; it is often the only language through which science becomes legible to the society that funds it.

Academic work on science policy in Science and Public Policy underscores that large‑scale research infrastructures thrive when they are embedded in broader narratives of national purpose. The danger is not spectacle itself, but spectacle without friction, without criticism, transparency, or demands from the very scientists and citizens in whose name the machines are built. If Coatlicue ever materializes, the challenge will not be to depoliticize it, but to politicize it better: to debate openly who gets computing time, how data are protected, and how the benefits reach beyond México City or a handful of elite labs.

In the meantime, defending the political theater around Coatlicue is, paradoxically, a way of defending the possibility of any ambitious scientific infrastructure at all. A purely “honest” ambition, modest in its claims and shy in its symbolism, would never survive the budget cycle. The myth of a public machine more powerful than anything else in Latin America creates the pressure, the envy, and the pride that can move concrete, cables, and contracts.

We should absolutely insist on transparency, on clear contracts, on safeguards against abuse. But we should not pretend that a cold, apolitical supercomputer, funded without fanfare, installed without ceremony, would serve México better. In a region where imagination has too often been outsourced along with data, it is no small thing for a government to say, loudly and theatrically: here, on our soil, we will build a machine that makes us dream bigger than our current capacities. If Coatlicue remains a myth, it will still have forced us to argue about the kind of digital future we want. And if it becomes reality, it will be because spectacle, not sober ambition alone, dragged that future into being.

Also Read: Brazilian Strongman Or Bridge Builder: Tarcísio And 2026 Security Politics