Mexico Will Build Latin America’s Largest Public Supercomputer

The Government of Mexico has entered a cooperation agreement with the Barcelona Supercomputing Center, one of the leading public supercomputing institutions in the European Union, to initiate the construction of a National Supercomputer. The project will create the largest public high-performance computing system in Latin America and begin operations through interim capacity hosted in Spain.

The initiative is central to Mexico’s technology sovereignty strategy, says Jose Antonio Peña, Head, National Digital Transformation Agency. The project addresses structural limitations in Mexico’s public computing infrastructure. Government agencies manage large, heterogeneous data systems that require high-performance computing to support fiscal analytics, customs processes, climate modeling, agricultural imaging, and AI development. Existing public infrastructure operates at speeds that extend processing times from weeks to months, which constrains policy execution, scientific research, and risk prevention.

The initiative also aligns with global scientific needs, says Mateo Valero, Director, Barcelona Supercomputing Center. He adds that Mexico has the computing capacity and data to develop capabilities that otherwise would be done for it. 

The Barcelona Supercomputing Center has more than two decades of experience in scientific computing and operates MareNostrum 5, one of the world’s most powerful public supercomputers. The center develops digital twins of physical and biological systems, advanced climate models, and language models designed to reduce dependence on private technology corporations. The agreement grants Mexico immediate access to this infrastructure while the domestic system is built.

The National Supercomputer will be installed over an estimated period of 24 to 36 months. Until the system is completed, Mexico will operate through a dedicated capacity block within the Barcelona Supercomputing Center, which will allow federal institutions to begin priority programs as early as January 2026.

The transition from conventional computing to high-performance processing will reduce workloads that now require weeks or months, says Jorge Luis Pérez, National Coordinator, Digital Infrastructure. 

The computer will allow the SAT, for example, to process large datasets related to fiscal declarations, audits, and customs operations. High-performance computing will also support improvements to the customs traffic light system and will help identify irregular patterns in cross-border transactions. These use cases rely on large data volumes that require parallel processing capabilities.

The system will also allow federal agencies to refine forecasts of extreme weather and strengthen tools for disaster prevention. Climate risk has increased in the last few years, and federal agencies have emphasized the need for faster and more precise modeling capabilities.

The supercomputer is expected to process over 2 million satellite images for agricultural monitoring. These datasets feed machine learning models that detect drought, irrigation deficiencies, and irregular changes in crops. The outputs will support decision-making for water management, food security, and rural planning.

The system will support large-scale training of language models, virtual assistants, and sector-specific AI applications for public agencies. Existing infrastructure limits the scale and speed of these models, which affects development cycles and increases reliance on private international platforms. The government has identified this dependency as a technology sovereignty concern.

The agreement includes technology transfer programs, specialist training, and joint research initiatives with universities, such as the Instituto Politécnico Nacional, says Rosaura Ruiz, Head, National Science And Technology Department. The goal is to develop a domestic scientific community capable of operating and expanding the high-performance computing ecosystem.