Partnerships
Fermaca Announces Mexico-Based AI Factory Platform for Latin America
Add Unite.AI to your preferred sources on Google
Fermaca announced on November 4, 2026 that it had been designated the first NVIDIA Cloud Partner based in Mexico and unveiled a phased AI factory platform for Mexico and Latin America, beginning with Fermaca Digital Origins in Mexico City.
The plan combines Fermaca’s infrastructure development and operating experience with NVIDIA accelerated computing, aiming to expand access to computing capacity for businesses and institutions. The platform is designed for enterprises to develop models around their own data and bring AI into operations, while giving developers and research institutions a foundation for new applications on infrastructure located in Mexico.
NVIDIA Partnership and Initial Hardware
NVIDIA’s Cloud Partner documentation defines participants as AI cloud providers within its Partner Network that deliver infrastructure purpose-built for modern AI workloads at production scale. It also describes NVIDIA-validated reference designs and deployment options for regional, regulatory and operational control.
The first deployment is planned around an initial NVIDIA Blackwell Ultra cluster, with NVIDIA Vera Rubin incorporated in a subsequent roadmap. Marcio Aguiar, NVIDIA’s director of enterprise for Latin America, said Fermaca is building the AI factory with NVIDIA B300 NVL72 systems to help enterprises across Mexico and Latin America develop and deploy AI using their own data.
Inventec is providing compute systems, DDN is providing storage, and Fermaca is responsible for infrastructure development and operations. DDN co-founder Paul Bloch said its data infrastructure is designed to keep NVIDIA GPUs supplied with data and turn advanced computing into productive AI capacity.
The Digital Origins product page describes an AI infrastructure platform and labels it Coming Soon. That status places the unveiling ahead of customer availability, so the initial deployment remains a buildout plan.
Power Targets and the Infrastructure Model
Fermaca’s near-term plan targets 60 MW of total facility power supporting operational AI facilities by December 31, 2027. That facility-wide measure includes computing, cooling and supporting systems.
The longer-term Fermaca Digital City development in Durango has an ambition of up to 1 GW of IT capacity, covering computing equipment only. Fermaca presents these as separate development milestones; the figures cover different phases and measurement scopes and do not represent current operating capacity.
Fermaca’s model connects gas infrastructure and energy development with fiber connectivity, AI facilities and accelerated computing. The company presents that integration as a way to address the infrastructure constraints that determine where AI can be built and how far it can scale.
The Fermaca Dreams corporate site gives the group more than 50 years of infrastructure development, construction and operation in Mexico across telecommunications, transportation and energy. It describes Digital City as combining natural-gas supply, electricity generation and advanced digital communications at a single location.
Digital Sovereignty and Workforce Development
Fermaca frames locally located AI development and deployment capacity as a digital-sovereignty objective. The stated goal is to help Mexican businesses, universities and institutions address national priorities and retain more of the knowledge and economic value created by AI.
The workforce component is México Aprende IA con Fermaca y NVIDIA, developed with NVIDIA to equip university teachers with AI knowledge, teaching resources and practical training for classroom use. The Durango government’s launch record places the national launch on March 24, 2026, with nine people preselected from around 50 participants and backgrounds spanning teaching, research, data science, robotics, bioengineering and technology development.
The announcement ties regional compute access to an energy-and-infrastructure buildout and a training program. The central issue is execution: Fermaca has specified the architecture, suppliers and power milestones, but its Coming Soon status shows that the partner designation and regional ambitions have not yet become customer-facing AI capacity.
