Pranav Hotkar
Mumbai, India - February 15, 2026 - AMD and Tata Consultancy Services (TCS) have expanded their partnership to jointly develop a large-scale AI infrastructure platform and deliver a standardized data center blueprint capable of supporting up to 200 megawatts of capacity, as enterprises shift from experimental artificial intelligence deployments to production-scale systems.
The collaboration centers on AMD’s Helios rack-scale architecture, designed to integrate compute, networking, and software into a unified platform for high-density AI workloads. The companies said the design will be offered to hyperscalers, enterprises, and sovereign AI initiatives seeking faster deployment of AI-ready facilities.
TCS said the initiative combines its data center engineering and integration capabilities with AMD’s compute platforms to help customers move from pilot projects to large operational environments. The blueprint is intended to shorten build timelines and reduce design complexity by providing a repeatable architecture that operators can adapt to regional power, cooling, and regulatory requirements.
The Helios platform incorporates AMD’s data center processors, accelerators, networking technology, and open software stack in a single rack-level system. The companies said this approach allows infrastructure to scale efficiently while supporting high-performance AI training and inference workloads.
The 200 MW design target reflects the growing scale of AI campuses, where facilities increasingly resemble industrial energy infrastructure rather than traditional enterprise data centers. Operators are seeking standardized designs that can be replicated across multiple sites as demand for compute capacity accelerates.
Both companies positioned the project as part of a broader push to expand advanced computing infrastructure in India and other regions adopting local AI capacity. Standardized architecture is expected to help organizations deploy domestic compute resources while maintaining flexibility in hardware configuration and energy sourcing.
Industry analysts note that vendors are increasingly moving beyond selling individual components toward delivering full reference designs for AI infrastructure. As clusters scale, integration challenges across servers, networking, and power systems have become as critical as processor performance.
The collaboration marks a shift toward platform-level partnerships in the data center sector, where chipmakers and system integrators jointly define facility architecture rather than leaving design entirely to operators. For customers, pre-engineered designs may reduce deployment risk while enabling faster expansion.
The companies said the Helios blueprint will support next-generation AI workloads and long operating cycles, reflecting expectations that AI infrastructure will run continuously at high utilization.
As demand for AI computing continues to rise, the partnership underscores a broader industry move toward standardized, large-scale facility designs intended to support hundreds of megawatts of capacity rather than isolated high-performance clusters.
