RT-One picks RAVEL to orchestrate sovereign AI across Americas
RT-One, the developer of what it describes as Latin America's largest AI technology parks and data centres, has selected RAVEL as its strategic technology partner for building a federated sovereign AI infrastructure platform spanning Brazil and the United States. The partnership centres on RAVEL's Orchestrate AI platform, which the company says can manage geographically distributed infrastructure as a single logical entity.
The deal's centrepiece is RT-One's 400MW campus currently under development in Uberlândia, Brazil. The facility, which the company positions as Latin America's largest AI data centre development, is planned to run entirely on renewable energy. RT-One intends to connect this site with infrastructure in the US, creating a cross-border AI platform that allows workloads to be routed according to sovereignty requirements, latency targets, operating cost or sustainability priorities.
The partnership
RAVEL's Orchestrate AI platform performs real-time workload placement across on-premises, cloud and hybrid environments, dynamically selecting the most appropriate compute resources for each job. The vendor says this improves GPU utilisation, reduces energy waste and maintains governance and compliance controls across the full distributed estate.
Denise Muyco, chief executive at RAVEL, said the relationship goes beyond a conventional software licence. "Working alongside RT-One, we are delivering a unique mix of technology, expertise, and support to speed up go-to-market and optimise operations," she said. "RT-One can maximise capacity for their customers while managing critical regulatory requirements."
Fernando Palamone, chief executive of RT-One, framed the deal in terms of infrastructure philosophy rather than product. "The next generation of AI infrastructure must be federated, sovereign, and sustainable," he said. RT-One's stated customer base spans enterprises, governments and cloud providers across the Americas.
Market context
Digital sovereignty has become a defining pressure on hyperscale and colocation operators, particularly in Latin America where data-localisation rules vary by jurisdiction and governments are increasingly mandating that sensitive workloads remain within national borders. Brazil's Lei Geral de Proteção de Dados (LGPD) imposes transfer restrictions broadly comparable to GDPR, while several regional governments are piloting sovereign-cloud procurement frameworks that require domestic infrastructure.
Federation-based orchestration sits at the intersection of two competitive markets: colocation and managed AI infrastructure on one side, and workload-scheduling software on the other. Hyperscalers including AWS, Azure and Google Cloud each offer multi-region AI inference and training services, but the sovereign-cloud argument typically favours independent operators who can credibly claim data will not traverse US-governed infrastructure without explicit consent. RAVEL is positioned by the company as an independent layer that can sit atop any substrate, which is a differentiator in regulated and government markets, though the durability of that advantage will depend on how quickly hyperscalers extend dedicated-region and sovereign-zone product lines.
The 400MW Uberlândia campus is a substantial bet on Latin American AI demand. Regional data centre investment has accelerated sharply since 2024, driven by nearshoring trends, expanding cloud adoption among Brazilian enterprises and growing interest from US operators seeking lower-latency connectivity to South American markets. Renewable energy availability in Brazil's interior, particularly hydroelectric and solar, gives RT-One a cost and sustainability argument that fossil-heavy markets cannot easily replicate.
Neither company disclosed commercial terms, contract value, or a timeline for the Uberlândia campus going live. The next milestones to watch are a confirmed operational date for the Brazilian site, the first named enterprise or government customer, and any RAVEL platform certifications relevant to regulated workloads in either jurisdiction.