Rackspace Shifts Toward AMD for AI Expansion

Rackspace is diversifying its compute architecture by integrating AMD hardware into its infrastructure. The company plans to deploy 30mw of AMD AI chips across its data centers.

This move signals a strategic pivot toward alternative silicon providers as the demand for AI-ready capacity intensifies. While the industry often focuses on a single dominant chipmaker, Rackspace's deployment of 30mw of capacity suggests that managed service providers are actively seeking to balance their hardware ecosystems.

The deployment of these chips is not merely a hardware refresh; it is a capacity play. By expanding its AI-capable footprint, Rackspace is positioning itself to capture workloads that require specialized compute power. This expansion comes as the broader market sees massive capital flows into data center space, evidenced by reports that Anthropic is set to pay $1.25bn a month to rent xAI data center space.

For the managed services sector, the ability to offer diverse silicon options is a competitive necessity. As AI workloads scale, the bottleneck is no longer just software, but the physical availability of power and specialized chips. Rackspace is addressing this by scaling its power-dense deployments.

Watch the deployment velocity of these 30mw clusters. The success of this expansion will depend on how effectively Rackspace can integrate these AMD chips into its existing managed cloud services to meet the rising demand for inference and training.