Pátria plans to deploy infrastructure to assist the growing demand from AI, cloud computing and high-performance workloads
In sum – what you might want to know:
Pátria returns to information facilities – The Brazilian asset supervisor has launched Omnia, a brand new platform focusing on AI-ready hyperscale information facilities in Brazil, Mexico and Chile, backed by a $1B funding.
Renewable-first infrastructure – All Omnia services will probably be powered by devoted wind tasks, minimizing water use and environmental impression.
LatAm positioned for AI development – CEO Rodrigo Abreu says Latin America presents the right combination of regulation and neutrality to draw world AI infrastructure investments.
Brazilian various asset supervisor Pátria has launched a brand new information middle enterprise named Omnia, aiming to increase throughout Brazil, Mexico, and Chile.
The agency plans to deploy $1 billion in infrastructure to assist the growing demand from synthetic intelligence, cloud computing, and high-performance workloads.
This transfer marks Pátria’s reentry into the info middle sector following its 2023 divestment of Odata to Aligned Knowledge Facilities for almost $2 billion.
Omnia’s focus will probably be on hyperscale information facilities constructed to deal with intensive processing necessities from massive expertise firms. The preliminary facility is anticipated to interrupt floor within the second half of 2025, with operations slated for late 2027.
The brand new information facilities will probably be absolutely powered by renewable power and designed to restrict water utilization and environmental impression. Pátria intends to supply power from a devoted wind energy challenge developed particularly for Omnia, fairly than counting on conventional grid infrastructure.
Felipe Pinto, Pátria’s associate overseeing infrastructure in Latin America, stated: “We’d like a devoted challenge and a devoted power supply for this initiative. Subsequently, we’re not specializing in short-term power imbalances however on large-scale tasks that require built-in options.”
In keeping with Omnia CEO Rodrigo Abreu, the primary tasks will include campus-style information facilities exceeding 100 megawatts in capability to assist high-density computing. “Knowledge middle capability is measured in power capability. Power is an asset for information middle builders. Up to now, a big information middle had 10 MW. As we speak, to fulfill AI calls for, we begin at 100 MW,” Abreu stated.
Whereas particular websites have but to be introduced, Abreu famous the problem of discovering places with direct entry to high-voltage transmission networks—a problem affecting main areas just like the São Paulo–Campinas hall and the Pecém space in Ceará. He expressed optimism that transmission constraints will probably be addressed quickly.
Abreu additionally identified that AI is shifting the trade’s geographic priorities. “For coaching AI fashions, we don’t want information facilities positioned close to end-users,” he defined. “The regulatory circumstances and geopolitical neutrality of this area make it appropriate for attracting investments from the U.S., Asia, or Europe. We anticipate Latin America to obtain important investments on this new class of knowledge facilities.”
As synthetic intelligence (AI) continues to quickly evolve, the demand for high-performance computing environments is rising at a quick tempo. AI purposes like generative fashions, real-time language translation and self-driving expertise require an enormous quantity of knowledge processing energy. That is the place AI-optimized information facilities come into play.
In contrast to conventional information facilities, synthetic intelligence information facilities are specifically designed with the intention of dealing with the distinctive challenges of AI workloads: huge quantities of knowledge, heavy computing duties and the necessity for speedy coaching and inference of machine studying fashions. To ship the required efficiency, 5 parts are important: GPUs, TPUs, high-speed networking, superior cooling methods and quick, scalable storage.
AI-optimized information facilities are the powerhouses making this transformation doable. By combining GPUs and TPUs for compute, high-speed networking, cutting-edge cooling methods and superior storage options, these services are geared up to fulfill the calls for of AI.
