Forging a New AI Frontier: How a Federated Model is Reshaping the Americas

SEATTLE, WA – June 23, 2026 – In a move that signals a significant shift in the architecture of global artificial intelligence, AI orchestration software leader RAVEL announced today it has been chosen by RT-One to power a first-of-its-kind federated AI infrastructure across the Americas. The strategic partnership aims to untangle the growing complexities of AI scalability and data sovereignty by creating a unified, intelligent network from a series of geographically distributed data centers, starting with a flagship project in Brazil.

This collaboration is not merely a software deal; it's a foundational play to build the next generation of AI infrastructure. RT-One, a developer of large-scale AI technology parks, will leverage RAVEL's orchestration platform to manage its sprawling network as a single, cohesive entity. The goal is to accelerate a vision where AI workloads can be executed where they are most effective—whether that’s in Brazil or the United States—without sacrificing governance, security, or performance. This initiative directly confronts the central challenge of the AI era: how to meet insatiable demand for computation while navigating a labyrinth of operational, regulatory, and environmental constraints.

The Rise of Sovereign and Federated AI

The concepts of "sovereign AI" and "federated AI infrastructure" are moving from the domain of academic papers to the core of strategic enterprise planning. Sovereign AI refers to the capability of a nation or region to build and control its own AI infrastructure, ensuring that critical data remains within its jurisdiction and compliant with local laws. This has become paramount as governments worldwide grow increasingly wary of data being processed and stored by foreign entities, a concern amplified by geopolitical turbulence.

The federated model is the architectural key that unlocks sovereignty at scale. Instead of relying on a single, monolithic data center, a federated infrastructure links geographically dispersed resources, allowing them to operate as one virtual platform. This approach is a direct response to the limitations of traditional, centralized systems. As AI models grow larger and data sets more massive, the industry faces immense pressure on power grids, resource management, and operational complexity. A federated network allows workloads to be dynamically routed based on a host of priorities—optimizing for data sovereignty, latency, energy cost, or the availability of renewable power.

This trend is underpinned by a rapidly growing market. Analysts project the AI orchestration market, valued at nearly US$10 billion in 2024, to surge to almost US$60 billion by 2033. The fastest-growing segment is customer-managed cloud deployments, reflecting a clear enterprise demand for greater control over data and compliance—the very principles driving the RAVEL and RT-One partnership.

The Engine Room: Orchestration Meets Infrastructure

At the heart of this alliance is the interplay between RAVEL’s intelligent software and RT-One’s physical infrastructure. RAVEL’s Orchestrate AI platform acts as the central nervous system for the entire distributed network. It goes beyond simple workload scheduling by employing a real-time intelligence layer that continuously analyzes telemetry data, cost models, network conditions, and organizational policies. This allows it to dynamically determine the best location and resources for any given AI task, whether it's training a large language model or running a time-sensitive inference job.

“Data centers are juggling demand with navigating operational, power, and regulatory constraints,” said Denise Muyco, CEO at RAVEL. “The relationship with RT-One goes beyond software to support their mission by acting as an AI infrastructure and operational intelligence partner. Working alongside RT-One, we are delivering a unique mix of technology, expertise, and support to speed up go-to-market and optimize operations.”

This intelligent orchestration promises tangible benefits: faster AI execution, improved GPU utilization, reduced energy consumption, and enhanced resilience. For RT-One, it means they can offer customers a platform that isn't just powerful, but also adaptable and compliant.

RT-One, for its part, is providing the muscle. The Latin American technology firm is making an initial investment of over US$1.2 billion in its flagship campus in Uberlândia, Brazil. This site is the first node in a planned network that will eventually span the Americas. By building in markets with reliable energy, favorable regulations, and long-term scalability, RT-One is creating a strategic footprint designed for the future of AI.

“Meeting AI demand will require a more connected infrastructure ecosystem globally,” said Fernando Palamone, CEO of RT-One. “The next generation of AI infrastructure must be federated, sovereign, and sustainable. With RAVEL, RT-One can deliver an intelligent AI infrastructure that preserves data sovereignty as it scales across distributed resources.”

A Green Blueprint for Latin America's AI Ascent

Perhaps the most compelling aspect of this initiative is its deep commitment to sustainability. The 400MW campus in Uberlândia is set to become Latin America’s largest AI data center development and will be powered entirely by renewable energy. Spanning over one million square meters, the project is not just about raw capacity; it's about efficient and responsible power usage. RT-One is targeting a Power Usage Effectiveness (PUE) below 1.2, a benchmark that would place it among the most efficient data centers in the world.

This focus on green energy is a strategic advantage. By leveraging Brazil's abundant renewable resources, RT-One can offer high-density computing at a potentially lower environmental and financial cost. The federated model, orchestrated by RAVEL, further enhances this by enabling workloads to be shifted to locations with the most available green energy at any given time. This provides a powerful blueprint for how the industry can decouple AI’s explosive growth from a corresponding rise in carbon emissions.

Moreover, the project signifies a major step in positioning Latin America as a key player in the global technology landscape. By building critical AI infrastructure locally, RT-One is fostering technological independence and creating a hub for innovation. The company plans to partner with local universities to cultivate a skilled workforce in AI and cybersecurity, ensuring the economic benefits extend beyond construction and into long-term, high-value employment. This investment could catalyze a broader wave of tech development in the region, attracting talent and capital to a market poised for significant growth.

Reshaping the Global AI Landscape

The RAVEL and RT-One partnership is more than a regional development; it's a potential harbinger of a broader restructuring of the global AI infrastructure market. For years, the landscape has been dominated by a few hyperscale cloud providers. While these giants offer immense power, their centralized models can create dependencies and raise concerns about data sovereignty and vendor lock-in.

The federated, sovereign model offers a compelling alternative. It provides enterprises and governments with a pathway to harness world-class AI capabilities while retaining control over their data and destiny. This collaboration demonstrates a viable, large-scale implementation that could inspire similar initiatives in other regions seeking greater technological autonomy, from Europe to Southeast Asia.

As AI becomes more deeply embedded in every facet of commerce and society, the underlying infrastructure that powers it will become a critical competitive differentiator. The ability to intelligently manage distributed resources, optimize for a complex set of variables including cost and sustainability, and ensure data governance is no longer a luxury—it is a necessity. This partnership provides a powerful and timely template for how to build that future, one sovereign, sustainable, and intelligently orchestrated data center at a time.