Rackspace (RXT.US) Joins Nvidia's AI Cloud Computing Ecosystem and Partners with Palantir to Advance “Sovereign AI” and AI Cloud Growth Engine

Zhitong Finance App learned that after Rackspace Technology (RXT.US) announced that it joined the artificial intelligence cloud computing partner program led by NVDA.US (NVDA.US), the company's stock price rose more than 14% before the US stock market on Thursday. The core of this cooperation is to integrate Nvidia Blackwell AI accelerated computing resource infrastructure, Palantir enterprise data and AI superplatform, and Rackspace's infrastructure and continuous operation and maintenance ecosystem, and service Businesses that value data sovereignty, governance, and rapid availability of computing systems. For Rackspace, this provides a new product portfolio for expanding existing cloud service customers to enterprise AI customers; the announcement has yet to disclose additional order amounts or revenue contributions. The current benefits are mainly reflected in the latest strengthening of business expansion capabilities and the shift of the cloud computing business to AI cloud service provider positioning.

Rackspace Technology is not accurately positioned as a new cloud force such as CoreWeave (NeoCloud) NeoCloud, but rather an established multi-cloud and hosting service provider for enterprise AI transformation. Its business covers consulting, migration, deployment, management and cost optimization for private clouds, public clouds, and hybrid clouds, as well as dedicated hosting, data center operation hosting, and ongoing engineering services, and supports hyperscale cloud computing platforms such as Amazon AWS, Microsoft Azure, and Google Cloud.

RackSpace and the new cloud forces intersect in terms of AI infrastructure and hosting operations, but RackSpace's main entry point is the integration, governance, and continuous operation of the enterprise's existing IT environment, rather than focusing on the sale of AI computing power resources or long-term computing power leasing like the new cloud forces with huge Nvidia AI GPU computing power clusters.

From an engineering perspective, after purchasing GPUs, enterprises also need to solve data access, model deployment, permission control, network and storage configuration, fault recovery, and service quality management. The institutional-level sovereign computing unit launched by Rackspace combines Blackwell with Palantir Foundry and AIP, and the engineering team is responsible for integration and continuous operation. The aim is to shorten the distance of AI from deployment to actual business use. The value of Rackspace's joint construction of “sovereign AI” with Nvidia and Palantir is also reflected in whether customers can control the location of data, access rights, model operation environment, and operation and maintenance responsibilities. Such capabilities are particularly important for financial, healthcare, and other regulated industries, providing a strong foundation for RackSpace to seek long-term AI computing power resource operation service revenue and orders.

This partnership will enhance RackSpace's ability to provide and manage AI computing services based on Nvidia's AI GPU computing power cluster infrastructure for regulated enterprises and enterprises with specific requirements for sovereign AI.

Gajan Kandia, CEO of Rackspace Technology, said, “AI sovereignty comes down to control: control over data, models, and the intelligent results created by the organization. Nvidia provides the foundation for accelerated computing, and Rackspace provides the infrastructure, engineering, and operations services needed to run in a production environment. For organizations with uncompromising requirements in terms of governance, sovereignty, and system availability, our responsibility is to keep the entire system running properly and continue to be responsible for the results.”

Rackspace said its new institutional sovereign computing unit (Institutional Sovereign Pod) is an AI infrastructure architecture with a governance mechanism. It will combine Nvidia Blackwell technology with Foundry, an AI superplatform created exclusively by Palantir (PLTR.US), and the AIP application platform to enable enterprises to control their own data, models, and AI environments.

Rackspace can provide pre-deployment engineering services and managed operation and maintenance for this environment, helping customers move from deployment to production operations faster and realize value earlier.

Blackwell's AI computing power solution is 100% dependent on the supply and delivery of Nvidia GPUs and supporting technology, but RackSpace's overall business does not only rely on purchasing Nvidia GPUs and then renting out computing power. The company has previously announced a partnership with AMD to integrate Instinct GPUs, EPYC CPUs, and ROCM software ecosystems, and allocate computing resources according to the workload. Customers can purchase a complete service including dedicated computing power, platform integration, and continuous operation and maintenance; the announcement did not define the cooperation as a simple hourly GPU rental, nor did it disclose the guaranteed GPU supply quota. From an investment perspective, the key is whether the new solution can generate ongoing contract revenue and generate reasonable profits after hardware, software licensing, and engineering investments.