Mexico and AMD Team Up to Build AI and Supercomputing Talent

Mexico wants to strengthen its own capabilities in artificial intelligence, supercomputing, and semiconductors with AMD’s support. The Secretariat of Science, Humanities, Technology and Innovation (Secihti) and the US chipmaker have signed a memorandum of understanding covering specialized training, knowledge exchange, and collaboration on technology projects, at a time when the country is preparing new high-performance computing infrastructure.

The Mexico-AMD deal in 20 seconds

• Secihti and AMD will collaborate on artificial intelligence, high-performance computing, and semiconductors.
• The agreement covers workshops, seminars, specialized training, and knowledge exchange.
• Mexico is looking to develop its own technological capabilities and connect universities, researchers, and infrastructure.
• The collaboration coincides with projects such as the Coatlicue supercomputer and the National Supercomputing and Artificial Intelligence Cluster.

The agreement was signed by Celina Peña Guzmán, Undersecretary of Technological Development, Liaison and Innovation at Secihti, and Keith Strier, Senior Vice President of Global AI Markets at AMD.

For now, this is only a memorandum of understanding, so it does not involve awarding AMD any contracts to supply processors or accelerators, nor does it mean the company will provide the hardware for Mexico’s future supercomputers. The document sets out a framework for identifying collaboration opportunities and building capabilities in technologies the Mexican government considers strategic.

The stated goal goes beyond teaching people to use AI tools. Secihti aims to connect training, research, infrastructure, and technology transfer to increase the country’s capacity to develop its own solutions.

AI, HPC, and Semiconductors Enter the Same Program

One of the most interesting aspects of the deal is precisely its scope. The collaboration isn’t limited to generative AI — it also covers high-performance computing (HPC), semiconductors, scientific research, and specialist training.

Secihti and AMD will be able to organize academic activities, workshops, and seminars, as well as share knowledge and best practices and identify projects where Mexican institutions could collaborate.

Celina Peña Guzmán explained that the government’s goal is to keep Mexico from simply adopting technology developed elsewhere. According to the Secihti official, the intention is to get the knowledge, talent, and infrastructure available in Mexico working together to generate new domestic capabilities and solutions.

This focus is especially relevant for semiconductors. Mexico has a sizable electronics and manufacturing industry integrated with North America, but the chip value chain covers far more than final device assembly: design, packaging, testing, software, materials research, and engineer training are also strategic areas.

The memorandum doesn’t spell out specific financial investments or announce new semiconductor plants. Its initial focus is mainly on talent, research, and technology collaboration.

AMD, for its part, is well positioned for the chosen focus areas. Besides its EPYC server processors, the company competes in AI accelerators with its Instinct family and takes part in some of the world’s largest HPC systems.

Keith Strier noted that high-performance computing is already used in fields such as energy, biology, climate science, materials research, and other scientific disciplines that need to process large volumes of information.

Coatlicue, Mexico’s Big Bet on Supercomputing

The agreement takes on added significance because it coincides with Mexico’s plans to expand its national computing infrastructure.

Among the projects Secihti explicitly cited is Coatlicue, the supercomputer promoted by Claudia Sheinbaum’s government as part of its scientific and technological sovereignty policy.

Coatlicue by the Numbers

The official Coatlicue announcement hasn’t specified which manufacturer will supply those 14,480 GPUs, which reinforces the need for caution: the memorandum with AMD doesn’t mean the company has been chosen as a supplier.

Also on the list is the National Supercomputing and Artificial Intelligence Cluster, designed to expand access for Mexican researchers and institutions to advanced computing resources.

The relationship between these projects and the AMD agreement should be read with caution. The announcement mentions Coatlicue as an example of the initiatives Mexico is developing, but it doesn’t state that AMD has been selected as the supercomputer’s supplier.

That distinction matters in a market where large HPC systems often combine processors, GPUs, specialized accelerators, high-speed networking, and storage from different vendors.

The challenge isn’t just about buying machines, either.

A supercomputer needs specialists who can program parallel applications, manage clusters, handle high-performance networking and storage, and adapt scientific workloads to CPUs and accelerators. The shortage of this kind of talent has become another constraint on expanding AI infrastructure.

That’s where much of the announced collaboration fits in.

Marco Moreno Ibarra, Secihti’s Director General of Development, Technology Transfer and Innovation, describes exactly this kind of continuity — from knowledge exchange to expert training to the later development of technology projects.

Mexico Wants the Infrastructure to Leave Knowledge Behind in the Country

The strategy reflects a debate playing out in many countries. Building national AI capacity isn’t just about installing GPUs.

It also requires energy, data centers, networks, universities, researchers, software, and specialized professionals capable of maintaining and using that infrastructure.

For Mexico, there’s also an opportunity stemming from its proximity to and economic integration with the United States, plus its sizable manufacturing industry. The expansion of AI and semiconductors is forcing North America to rethink supply chains that, for decades, were concentrated mainly in Asia.

The AMD agreement doesn’t by itself guarantee that Mexico will move toward greater technological autonomy. Nor does it yet set quantitative targets for the number of professionals trained, investment, or computing capacity that will be deployed as a result of the collaboration.

What it does provide is a framework for one of the world’s leading processor and accelerator companies to work directly with Mexico’s scientific administration on training and project design.

Hiram Monroy, AMD’s sales director for Spanish-speaking Latin America, said the collaboration will cover knowledge exchange, technical cooperation, and specialized training.

The global AI race is showing that having chips matters, but having professionals capable of using them and building on top of them can be just as hard to secure. Mexico is now trying to work on both sides of that equation while it prepares a considerably larger national supercomputing infrastructure.

Frequently Asked Questions

What have AMD and the Mexican government agreed to?

AMD and Secihti have signed a memorandum of understanding to collaborate on training, research, and knowledge exchange related to artificial intelligence, supercomputing, semiconductors, and other advanced technologies.

Will AMD supply the chips for the Coatlicue supercomputer?

The announcement doesn’t state that AMD has been chosen as Coatlicue’s supplier. Secihti cites the supercomputer as one of the Mexican projects related to building national capabilities.

What activities does the agreement cover?

It’s expected to coordinate academic activities, workshops, seminars, specialized training, the exchange of best practices, and the identification of possible joint projects.

Why is Mexico betting on supercomputing?

The Mexican government aims to expand the capabilities available for research and artificial intelligence and reduce its reliance solely on technology and knowledge developed outside the country.