Rackspace & Palantir launch sovereign AI framework

Rackspace Technology and Palantir have launched an operating framework for regulated and sovereign enterprises using artificial intelligence in production. It combines Rackspace infrastructure with Palantir's Foundry and AIP software.

Rackspace will serve as a preferred operator for on-premises, private cloud and sovereign deployments in sectors including healthcare, financial services and energy. The companies are targeting organisations that need tighter control over where data sits, how systems are governed and who manages day-to-day operations.

The arrangement extends an existing partnership and puts Rackspace in charge of infrastructure, certified engineering staff and managed operations around Palantir's software layer. Palantir's Foundry and AIP products sit at the data and AI platform level, while Rackspace provides the environment in which customers run them.

Regulated industries have been more cautious in adopting AI because of concerns over compliance, auditability and data sovereignty. Hospitals, banks, energy operators and public sector bodies often face restrictions on cross-border data transfers, model governance and the use of third-party systems in sensitive environments.

That has created an opening for suppliers offering tighter operational control rather than open-ended public cloud use. Rackspace and Palantir are positioning the framework for customers that want AI deployed inside private or sovereign environments instead of placing data and operations on broader shared platforms.

Gajen Kandiah, Chief Executive Officer of Rackspace Technology, said many large organisations had moved beyond AI planning but were still struggling to operationalise systems safely.

"While most regulated enterprises have an AI strategy, they often lack the operating model to put AI into production safely and at scale. This effort by Rackspace closes that gap," said Gajen Kandiah, Chief Executive Officer of Rackspace Technology.

He added: "Rackspace brings the governed infrastructure, the Palantir-certified engineers, the managed operations, and the accountability for outcomes in the environments where our customers actually live. This is deploy and operate, not deploy and leave. This is how organizations with the most demanding requirements move AI into production at scale."

Palantir described the tie-up as a response to demand from customers that want AI tools deployed closer to sensitive operations and under stricter oversight. The company has built much of its business in government and heavily regulated settings, where permissions, audit trails and deployment location are central concerns.

"Sovereign AI requires more than access to a model. It requires an operating layer that lets enterprises govern data, enforce permissions, route models, audit actions, and deploy capability where the mission lives," said Alex Karp, Co-Founder and Chief Executive Officer of Palantir Technologies.

He added: "This framework brings Palantir Foundry and AIP together with Rackspace's infrastructure and delivery capabilities for mission-critical environments."

Early traction

The partnership has already expanded since its initial agreement earlier this year. Rackspace has built roughly 400 Palantir certifications across sales, engineering, delivery and operations, including a sizeable group of Palantir-certified Forward Deployed Engineers.

According to Rackspace, those engineers are expected to support projects across healthcare, financial services, energy and the mid-market. The model relies on specialist staff working inside or closely alongside customer environments, particularly where systems are isolated or subject to strict governance rules.

One early joint deployment was completed in less than two months at a US-based solar tracking manufacturer. In that project, Rackspace engineers deployed AI-enabled workflows on Palantir Foundry, and the customer recorded a 94% reduction in quote cycle time.

That example gives the companies a reference point beyond theory, though the broader commercial test will depend on how many regulated enterprises are prepared to commit significant workloads to combined software and managed infrastructure offerings. Buyers in these sectors tend to move slowly, and sales cycles can be lengthy because procurement, compliance and security reviews are extensive.

Internal use

Rackspace is also adopting the software internally. It plans to deploy Foundry and AIP across more than 70% of its own back-office operations under a programme called Rackspace OneOS.

By using the same stack itself, Rackspace is seeking to show prospective customers that the model works in live operations rather than being limited to pilot schemes. It also reflects a wider trend among infrastructure and software providers to use internal deployments as evidence when selling AI systems into cautious industries.

The companies said they will work together to win customers in healthcare, financial services, energy, private equity and the mid-market. They also plan larger private cloud and sovereign installations in which Rackspace and Palantir engineering teams work side by side within customer environments.

For Rackspace, the framework also supports a broader effort to define its role in AI around operation and governance rather than model development.