Fermaca Partners With Nvidia for 60MW AI Infrastructure in Mexico
Mexican infrastructure firm targets 2027 completion for AI-ready data centers using Nvidia technology.
1Strategic AI Infrastructure Expansion
Mexican firm Fermaca has joined Nvidia’s cloud partner program to develop AI-focused data centers. The company aims to deploy 60 megawatts of total power capacity by December 31, 2027. This initiative marks Fermaca’s entry into the high-performance computing market, providing businesses with external processing power for complex model development and deployment.
2Initial Deployment and Technology
The rollout begins with the Digital Origins facility located in Mexico City. This site will utilize Nvidia’s Blackwell Ultra technology to support intensive AI workloads. Fermaca plans to integrate future hardware generations, such as the Vera Rubin architecture, as they become available, ensuring their infrastructure remains competitive for evolving enterprise AI requirements.
3Scaling Through Digital City
Beyond the initial 60-megawatt goal, Fermaca is developing a larger-scale project known as Fermaca Digital City in Durango. Long-term projections for this site contemplate up to 1 gigawatt of computing capacity. As of April 2026, state authorities estimated the total investment for this massive data center project at 2.7 billion dollars.
4Nvidia’s Market Dominance
Nvidia remains a central driver in the global AI hardware market. During the second fiscal quarter of 2026, the company’s data center division generated approximately 89 billion dollars in revenue. This segment accounted for 92.5 percent of Nvidia’s total quarterly income, which reached 96.221 billion dollars, highlighting the massive scale of current demand.
5Energy Infrastructure Integration
Fermaca is leveraging its existing expertise in energy and gas infrastructure to support these data centers. The 60-megawatt target includes power requirements for both computing hardware and essential cooling systems. By combining energy assets with fiber optic connectivity, the company aims to provide a comprehensive platform for specialized cloud-based AI processing.
6Grid Connection Challenges
The expansion of data centers in Mexico faces significant hurdles regarding electrical grid connectivity. As of April 2026, the country held 279 megawatts of operational capacity, with another 205 megawatts under construction. Private firms have invested roughly 340 million dollars in electrical infrastructure to reinforce local subestations and transmission lines, particularly in the Bajío region.
7Engineering and Design Adaptations
Rapid advancements in server technology force companies to modify building designs during construction. Experts note that engineers must adapt electrical supply and cooling systems to accommodate the latest hardware specifications. Consequently, firms are shifting design work directly to construction sites to ensure facilities can support the high-density power needs of modern AI.
8Industry Outlook and Delays
Infrastructure projects in Mexico are experiencing extended timelines due to complex electrical works and land permitting processes. For instance, CloudHQ has faced delays in Querétaro while constructing high-voltage subestations. These logistical challenges underscore the difficulty of scaling AI infrastructure, requiring firms to balance long-term capacity goals with immediate grid limitations.