TCS & AMD: Developing Rack-Scale AI Infrastructure for India

Tata Consultancy Services (TCS) and AMD have expanded their strategic collaboration to bring rack-scale AI infrastructure to India. 

Through its subsidiary HyperVault AI Data Center Limited, TCS is partnering with AMD to co-develop a rack-scale AI infrastructure design based on AMD's Helios platform. 

The initiative supports India’s national AI programmes and focuses on sovereign AI capability.

A rack-scale architecture integrates compute, networking and storage within a single rack system to support high-density AI workloads. AI workloads require parallel processing power and fast data movement across servers. 

Helios runs on AMD Instinct MI455X GPUs, which are accelerators designed to train and run AI models. It includes next-generation AMD EPYC Venice CPUs, responsible for general-purpose and high-performance computing tasks inside the data centre. 

AMD Pensando Vulcano NICs – network interface cards that control data traffic between servers – support high-speed connectivity. The system operates within the open ROCm software ecosystem, which enables developers to build and optimise AI applications on AMD hardware.

Dr. Lisa Su, Chair and CEO, AMD, says: "AI adoption is accelerating from pilots to large-scale deployments and that shift requires a new blueprint for compute infrastructure. 

"With Helios, we are delivering an open, rack-scale AI platform designed for performance, efficiency and long-term flexibility. 

"Together with TCS, we are enabling enterprises across India to deploy AI at scale today while building the compute foundation of tomorrow.”

Lisa points to the move from pilot projects to production-scale AI. 

That move places pressure on existing facilities that lack the density, cooling capacity and network architecture required for large AI clusters. 

Rack-scale systems address this by combining hardware and networking in tightly integrated configurations that support higher utilisation and predictable performance.

A 200MW AI-ready blueprint

As part of the expanded collaboration, TCS and AMD are introducing an AI-ready data centre blueprint supporting up to 200MW of capacity. 

A 200MW campus aligns with hyperscale data centre development, where operators deploy tens of thousands of servers to support cloud and AI services.

The companies work with hyperscalers and AI companies to accelerate data centre build-outs in India. 

A standardised rack-scale blueprint supports faster construction and commissioning while maintaining performance and efficiency requirements.

Operational efficiency remains central to the design. High-density AI racks demand advanced cooling systems, resilient power distribution and high-bandwidth networking. 

By combining AMD’s hardware with Tata Consultancy Services’ enterprise integration and engineering capability, the collaboration aims to streamline deployment from initial design through to live operation.

K. Krithivasan, MD and CEO at TCS, says: “This collaboration lays the foundation for AMD’s first Helios powered AI infrastructure in India. 

“By combining our strengths in AI, connectivity, sustainable power and advanced data centre engineering, we are poised to deliver state-of-the-art infrastructure solutions for AI companies and global enterprises. 

“We are thrilled to deepen our longstanding partnership with AMD as we expand our participation in the AI ecosystem – Infrastructure to Intelligence.”

Krithivasan highlights connectivity and sustainable power as core components. AI clusters require resilient fibre connectivity to transfer large volumes of data and stable power to sustain high compute loads. 

HyperVault’s AI data centre strategy

Tata Consultancy Services established HyperVault in 2025 with the vision of delivering gigawatt-scale, secure and reliable AI-ready infrastructure for hyperscalers, AI companies and global enterprises. 

HyperVault focuses on purpose-built facilities designed for AI from inception. By aligning with AMD’s “Helios” architecture, the subsidiary integrates advanced GPUs, CPUs and networking technology into its core infrastructure rather than adapting legacy estates.

The announcement builds on the existing strategic collaboration between TCS and AMD to help enterprises scale AI adoption and modernise hybrid environments. 

For data centres in India, the collaboration centres on rack-scale design, sovereign control and hyperscale power capacity. As enterprises deploy AI at scale, infrastructure architecture becomes central to performance, efficiency and national capability.