Rackspace, the operator of the full enterprise AI stack, and Together AI, the AI Native Cloud, has announced an initial multiyear strategic AI infrastructure agreement under which Rackspace expects to provide and operate dedicated NVIDIA GB300 NVL72 infrastructure connected with NVIDIA Spectrum-X Ethernet networking to support Together AI inference workloads.
Deployment is planned for 2027, subject to the satisfaction of certain conditions. This deployment reflects Rackspace’s strategy to become the operator of the full enterprise AI stack, leveraging its decades of experience operating mission-critical environments to give customers compute solutions that meet the security, trust, and governance needs in regulated, production environments.
The companies also entered into a separate, non-binding letter of intent (LOI) to explore a joint go-to-market approach for enterprise-focused AI solutions, with Together AI providing inference capabilities and Rackspace providing the supporting AI infrastructure, platforms, solutions and forward-deployed engineering expertise.
Dedicated AI infrastructure backed by operational accountability
Together AI has partnered with Rackspace based on its deep infrastructure and engineering expertise, operational capabilities and ability to support the demanding requirements of enterprise AI workloads. Under the initial agreement, Rackspace expects to provide and operate the dedicated NVIDIA Blackwell Ultra infrastructure, including associated power, liquid cooling, networking and storage. The parties expect to evaluate additional security, governance and compliance controls for future offerings, depending on end-customer requirements and agreed configurations. Together AI expects to provide the managed cloud platform operating on that infrastructure, including access to open source models, inference optimizations and developer tools.
Rackspace intends to build a repeatable model to support AI platform companies, systems integrators and enterprises as they scale AI infrastructure while retaining control of their technology and customer relationships. Rackspace’s approach to scaling AI infrastructure is to seek contractual commitments for dedicated capacity and manage related financing and capital commitments with the goal of aligning to expected customer demand and the useful life of the underlying hardware.
“Enterprise AI requires more than access to compute. Success depends on bringing compute into production, operating it reliably and delivering sustained value. That requires deep infrastructure and engineering expertise combined with operational accountability,” said Gajen Kandiah, Chief Executive Officer of Rackspace. “Rackspace brings more than 25 years of experience operating complex enterprise environments. Today’s announcement provides an opportunity to apply that experience to the dedicated infrastructure supporting Together AI’s inference platform. We believe this relationship advances our strategy to be the operator of the full enterprise AI stack.”
Exploring additional enterprise AI opportunities
Beyond the infrastructure agreement, Rackspace and Together AI intend to explore potential opportunities to jointly serve enterprise customers requiring greater control, compliance, and data sovereignty. The non-binding LOI contemplates evaluating how Together AI’s inference, post-training and reinforcement learning capabilities could be combined with Rackspace’s infrastructure, platforms and engineering services to provide additional options for customers deploying AI in production. Any joint offerings would depend on further agreement between the parties and applicable customer requirements.
Rackspace’s experience supporting enterprises in regulated industries, combined with its expertise across private cloud, hyperscaler and edge environments, may create opportunities to evaluate deployment configurations that bring Together AI’s cloud closer to enterprise data and workloads. The availability and characteristics of any such configurations, including data residency, security and compliance features, would be determined by the specific offering and customer arrangement, potentially providing greater flexibility in where and how customers deploy AI.
“Scaling in the enterprise requires infrastructure that can grow with our ambitions while delivering the performance and reliability customers expect,” said Vipul Ved Prakash, founder and CEO of Together AI. “Rackspace brings the infrastructure engineering and operational expertise to support that growth and we look forward to exploring additional opportunities together.”
Powered by next-generation NVIDIA accelerated computing
Rackspace, an NVIDIA Cloud Partner, plans to operate the deployment using dedicated NVIDIA GB300 NVL72 systems in a purpose-built, liquid-cooled facility, supported by NVIDIA networking. The planned infrastructure is designed to provide the performance, reliability and operational foundation needed to support Together AI’s enterprise inference workloads at scale.
“Enterprises need infrastructure that turns AI advances into services they can deploy at scale,” said Raj Mirpuri, Vice President of Global AI Clouds and Infrastructure Ecosystem at NVIDIA. “With NVIDIA Blackwell Ultra, Rackspace and Together AI can run inference workloads efficiently and bring the next wave of enterprise AI services to market.”
