Hinduja Global Solutions has selected Rackspace Technology to build private AI infrastructure powered by AMD, in a three-year deployment starting December 2026, according to Yahoo! Finance. The agreement, announced October 29, 2026, makes HGS the first announced global system integrator customer of Rackspace's Managed Compute and Inference Platform.
The deal covers private cloud and AI compute for HGS customers in the US and UK, giving the provider direct control over AI economics, security, and compute. HGS is part of the multi-billion-dollar Hinduja Group and operates roughly 16,710 employees across 10 countries.
What the agreement covers
The deployment runs through a multiyear compute agreement between Rackspace and AMD that reserves an initial 30 MW footprint of dedicated AMD silicon across Rackspace's global data centers from late 2026 through 2028. AMD now serves as the strategic technology partner for the silicon layer of Rackspace's governed enterprise AI stack, a design that pairs dedicated compute, private cloud, and managed operations into a single production path.
For HGS, the infrastructure supports a shift from traditional outsourcing toward AI-enabled services. The company plans to use the new capacity to bring secure, real-time customer context into frontline interactions and digital channels, and to offer private, sovereign-ready cloud options to enterprise customers that need tighter control over data and regulatory compliance.
Why enterprises are pulling AI workloads into private infrastructure
The pattern is structural. As inference becomes a recurring cost rather than a one-off training expense, enterprises are repricing where their workloads run. Public cloud economics that worked for bursty web traffic do not always hold for sustained model serving, and governance requirements in regulated industries push toward dedicated environments.
AI data centers differ from general-purpose facilities in that they are optimized for the parallel processing demands of AI workloads, as described by Wikipedia's entry on AI data centers. That difference is why companies like HGS are contracting dedicated capacity rather than renting generic compute.
AMD's position in this market has grown through its Instinct accelerators and EPYC processors, which Agent Jesse's ranking of AI infrastructure companies notes support both cloud and enterprise AI workloads. The company has also pushed an open approach to AI infrastructure, including an open blueprint for agentic AI infrastructure based on Open Compute Project specifications.
Quotes from the announcement
Phil Fersht, CEO and chief analyst at HFS Research, framed the deal as part of a broader industry reset.
“Competitive advantage will increasingly come from providers that can combine deep process expertise with the AI infrastructure, compute and operational capabilities needed to deliver at scale.”
Venkatesh Korla, global chief executive officer of HGS, said production AI forced a rethink of the infrastructure beneath the company's applications.
“Working with Rackspace gives us control over the cost, security and performance of our AI infrastructure, and adds private and sovereign-ready cloud capabilities that broaden what we can deliver to customers.”
Gajen Kandiah, chief executive officer of Rackspace Technology, said the deal shows what production AI actually requires: dedicated compute, private cloud, and managed operations working alongside public cloud.
Why sovereign-ready AI matters
The HGS deployment is a data point in a larger migration of AI workloads toward governed, auditable infrastructure. Enterprises in finance, healthcare, and public services cannot run inference on capacity they cannot inspect. Sovereign-ready private cloud gives them a contractual and technical boundary around their data, their models, and their cost base.
Rackspace positions this as its Enterprise AI growth vector: an Outcomes-as-a-Service model built on secure infrastructure, data foundations, and forward-deployed engineering, aimed at regulated and mission-critical industries where governance, sovereignty, and uptime are non-negotiable.
The HGS deployment begins in December 2026 and runs three years, with the underlying AMD compute agreement extending through 2028.
FAQ
When does the HGS Rackspace deployment start?
The deployment begins in December 2026 and runs for three years under a multi-year infrastructure and services agreement.
What is the Rackspace AMD compute agreement?
It is a multiyear agreement covering an initial 30 MW footprint of dedicated AMD compute across Rackspace's global data centers from late 2026 through 2028, with AMD as the silicon partner in Rackspace's governed enterprise AI stack.
Is HGS the first customer under this AMD agreement?
Yes. HGS is the first announced GSI customer deployment under Rackspace's multiyear AMD compute agreement.