- StorageReview reported on August 13, 2026, that RT-One is the first commercial customer of the AMPHIX AI Infrastructure Platform. [1] - RT-One is expected to validate compute, networking, thermal and orchestration configurations at an AMPHIX Center of Excellence near Colusa, California. [1] - The resulting configuration is intended to become a standardized deployment block for RT-One’s federated AI infrastructure model across the Americas. [1] - RT-One previously disclosed a 400MW renewable-powered AI campus under development in Uberlândia, Brazil. [2] - No transaction value, capacity commitment or contract term was disclosed in the available reports. [1][2]
RAVEL and Strata Expanse have named RT-One the first commercial customer of the AMPHIX AI Infrastructure Platform, StorageReview reported on August 13, 2026. AMPHIX is designed to validate AI infrastructure before customers commit to permanent production campuses. [1]
RT-One is expected to use a planned AMPHIX Center of Excellence near Colusa, California, to test and optimize compute, networking, thermal-management and orchestration configurations. RAVEL’s Orchestrate AI software will manage workloads across the environment, while the validated configuration is intended to become a standardized deployment block for RT-One’s infrastructure network across the Americas. [1]
The reported customer designation follows RT-One’s June selection of RAVEL as a strategic technology partner for a federated infrastructure model spanning Brazil and the United States. RT-One’s disclosed pipeline includes a 400MW AI campus under development in Uberlândia, Brazil, which RAVEL described as powered entirely by renewable energy. [2]
The Deal #
AMPHIX combines infrastructure-ready land, power, cooling, secure networking, certified compute and storage systems, and workload-orchestration software. Its Centers of Excellence are intended to provide production-grade validation environments where customers can test workloads and operating policies before making larger infrastructure commitments. [1][3]
For RT-One, the stated objective is to certify one repeatable configuration and carry it into later deployments. Local regulatory adaptations may be made at the affected infrastructure layers while the core technology and operating policies remain standardized. [1]
The Colusa Validation Site #
The Colusa COE is planned as one of the largest planned AMPHIX sites and is expected to use a microgrid developed and managed by Colusa Indian Energy. StorageReview reported that Strata Expanse will supply the physical infrastructure foundation, including land, power, cooling and secure connectivity, while RAVEL will handle technology certification, deployment validation and operator training. [1]
The validation program is expected to cover workload behavior, thermal conditions, energy consumption, cost controls and orchestration policies. The site is also intended to support customer-environment preparation before RT-One’s permanent campuses are available. [1]
RT-One’s Federated Infrastructure Model #
RT-One selected RAVEL in June to help operate geographically distributed and hybrid infrastructure as a single AI platform while preserving data sovereignty and operational control. The model is intended to direct workloads across on-premises, cloud and hybrid environments according to requirements such as sovereignty, resiliency, latency, operating cost and sustainability. [2]
The company’s strategy is built around connecting infrastructure in markets where power availability, renewable resources, connectivity and regulatory conditions support customer requirements. The Uberlândia campus is one disclosed anchor for that strategy, with 400MW of planned capacity. [2]
Platform Background #
RAVEL, Strata Expanse and Available Infrastructure launched AMPHIX in March 2026 as a modular platform combining infrastructure-ready land and power, secure networking, certified compute and storage systems, and orchestration software. [3]
RAVEL says AMPHIX Centers of Excellence provide production-grade environments where customers can test workloads under real-world infrastructure conditions before investing in full-scale deployments. [3]
Further sources #
The stories that matter, in one email. Free — unsubscribe anytime.