AMD and Rackspace Technology have signed a definitive agreement to deploy an initial 30MW footprint of AMD-based compute capacity across Rackspace’s global data center portfolio. The phased rollout is scheduled to begin in late 2026 and continue through 2028, advancing the memorandum of understanding the companies announced in May.
Under the agreement, AMD becomes a strategic silicon partner within Rackspace’s enterprise AI infrastructure strategy. The deployment will combine AMD Instinct accelerators, including the MI355X, MI350P, and future generations, with AMD EPYC processors to support AI training, inference, and enterprise workloads in regulated industries.
At full deployment, the 30MW environment is expected to provide substantial AI compute capacity for enterprise customers, including organizations in healthcare and other regulated sectors. Rackspace said the infrastructure is designed to support large-scale clinical AI initiatives, inference services, and other workloads that require governance, accountability, and operational oversight.
The companies plan to integrate the AMD hardware stack into Rackspace’s Enterprise AI Cloud architecture. The platform is intended to match workloads with the appropriate compute resources while providing centralized management and operational accountability across the infrastructure stack.
Rackspace CEO Gajen Kandiah said regulated industries require AI infrastructure that is governed end-to-end rather than assembled from multiple independent providers. He positioned the collaboration as an effort to combine compute infrastructure and operating services into a single managed framework with accountability extending from the hardware layer through business outcomes.
AMD Senior Vice President and General Manager of Compute and Enterprise AI Dan McNamara said enterprise AI deployments increasingly require a mix of accelerated and general-purpose computing resources optimized for different workload requirements. He noted that the combination of AMD’s AI compute portfolio and Rackspace’s managed cloud operating model is intended to provide enterprises with scalable and accountable infrastructure for production AI environments.
The agreement also includes joint go-to-market activities. Both companies will dedicate sales and marketing resources and commit personnel to jointly develop and pursue customer opportunities across regulated industries built on AMD-powered infrastructure.
The deployment is expected to accelerate the delivery of the four services outlined in the earlier memorandum of understanding:
- Enterprise AI Cloud
- Enterprise Inference Engine
- Inference as a Service
- Bare Metal AMD Instinct
Together, these offerings are designed to provide a managed AI infrastructure stack spanning bare-metal compute through fully operated inference services. The two companies position the initiative as an alternative to traditional self-managed bare-metal AI deployments, targeting enterprises moving beyond pilot projects and into production AI and agentic workflow implementations within core business systems.
Amazon