Chipmaker Advanced Micro Devices (AMD) and India's largest IT services company Tata Consultancy Services (TCS) on Monday, February 16, said they will co-develop and deploy the next-generation “Helios” rack-scale AI architecture in India, with an eye on large enterprise and sovereign AI demand.
The collaboration, to be executed through TCS subsidiary HyperVault AI Data Center Limited, includes an AI-ready data centre blueprint that can scale up to 200 megawatts of capacity. The companies, in a statement, said the platform is designed to support high-performance AI training and inference workloads for hyperscalers, AI firms, and large enterprises operating in India.
The announcement comes as enterprises globally move from AI pilots to production-scale deployments, sharply increasing demand for high-density compute infrastructure.
Under the expanded partnership, AMD will supply its full-stack AI compute platform, while TCS will bring data centre engineering, integration and enterprise deployment capabilities.
The Helios architecture will be powered by:
• AMD Instinct MI455X GPUs
• Next-generation AMD EPYC “Venice” CPUs
• AMD Pensando Vulcano NICs
• The open ROCm software ecosystem
According to the companies, the rack-scale design is meant to improve performance efficiency and reduce time to deployment for enterprise AI workloads.
AMD Chair and CEO Lisa Su said AI adoption is moving rapidly toward large-scale deployments that require a new compute blueprint. TCS CEO and MD K. Krithivasan said the partnership marks the foundation of AMD’s first Helios-powered AI infrastructure in India.
The move aligns with India’s broader push to build domestic AI capacity.
The government’s IndiaAI Mission has already committed more than ₹10,000 crore toward strengthening national AI compute infrastructure. Policymakers and industry leaders have repeatedly stressed the need for sovereign AI factories, or locally controlled compute and data ecosystems.
At the same time, India’s data centre market is entering a high-growth phase. Industry estimates suggest total capacity could more than double by the end of the decade, driven by data localisation requirements, hyperscaler expansion and AI workloads.
Against that backdrop, a 200 MW AI-focused blueprint signals hyperscale ambition. Large AI clusters increasingly require power capacity far beyond traditional enterprise data centres.
TCS set up HyperVault in 2025 to build secure, AI-ready infrastructure at scale. The Helios collaboration gives the unit an early flagship platform.
For AMD, the partnership strengthens its push to expand the Instinct GPU ecosystem globally as competition in AI accelerators intensifies.
TCS shares were last seen trading largely flat at ₹2,688 apiece, which is 0.16% down for the day.
Also Read: UK government seeks greater powers to regulate online access, rein in AI chatbots
