AMD and TCS Bet on Helios to Shape India’s Rack-Scale AI

Key Highlights:

  • Helios uses AMD Instinct MI455X GPUs, EPYC “Venice” CPUs, and ROCm software.
  • AMD and TCS will co-develop India’s first rack-scale AI system based on the Helios platform.
  • The collaboration targets sovereign AI factories and national AI initiatives.
  • The design supports AI-ready data centers up to 200 MW capacity.

AMD and Tata Consultancy Services have expanded their strategic partnership to bring the Helios rack-scale AI architecture to India. The collaboration focuses on building AI infrastructure designed for national AI initiatives, sovereign AI factories, and large enterprise deployments. It matters because India is moving from AI pilots to full-scale production systems, and that shift needs new data center designs.

What is Helios?

Helios is AMD’s rack-scale AI platform built for large, power-hungry AI workloads. It combines AMD Instinct MI455X GPUs, next-generation EPYC “Venice” CPUs, Pensando Vulcano networking, and the open ROCm software stack. Instead of treating servers as isolated units, Helios designs AI compute at the rack level. This approach improves performance, energy efficiency, and long-term flexibility for AI training and inference.

How TCS fits into the plan

TCS will co-develop the Helios-based infrastructure through its subsidiary HyperVault AI Data Center Limited. HyperVault was set up in 2025 to deliver secure, large-scale AI-ready infrastructure. Together, AMD and TCS will offer an AI data center blueprint that supports up to 200 MW of capacity. They also plan to work with hyperscalers and AI companies to speed up data center build-outs across India.

Why this matters for India’s AI ecosystem

India’s AI demand is growing across government, research, and enterprise use cases. However, many workloads still rely on global infrastructure. Helios changes that equation by supporting sovereign AI factories that keep data, compute, and control within the country. With TCS handling engineering, power design, and operations, the platform aims to reduce deployment complexity and improve reliability at scale.

What comes next?

The collaboration builds on earlier AMD–TCS efforts to modernize hybrid and AI environments. While timelines were not disclosed, both companies position Helios as a foundation for long-term AI infrastructure in India. As AI moves deeper into production systems, Helios could become a reference architecture for how large AI data centers are built and operated in the country.

In short, Helios signals a shift from buying AI hardware to designing AI infrastructure at national scale.

56 Views