Advanced Micro Devices (AMD) and Tata Consultancy Services (TCS) have announced a significant expansion of their strategic partnership. This collaboration will see the two technology powerhouses co-developing a state-of-the-art, rack-scale Artificial Intelligence (AI) platform, built upon AMD’s innovative Helios architecture. The primary goal is to accelerate research and foster innovation across diverse industries. AMD states that this initiative will seamlessly merge its advanced data center processors with TCS’s extensive integration and service capabilities, ultimately providing robust computing infrastructure capable of handling the most complex workloads.
AMD and TCS: Pushing the Boundaries of Rack-Scale AI
According to information released by both companies, this joint effort is designed to empower clients with the ability to construct highly scalable and optimized High-Performance Computing (HPC) infrastructure. This infrastructure will be perfectly suited to manage the demands of future-generation workloads. TCS will play a crucial role by delivering comprehensive system integration, deployment, and ongoing management services, helping businesses streamline their IT operations and boost overall efficiency.
The Helios system, as described by AMD, is specifically engineered to cater to highly demanding applications including Artificial Intelligence (AI), Machine Learning (ML), intricate data analytics, and cutting-edge scientific research. This powerful platform integrates AMD Instinct MI455X GPUs, the upcoming AMD EPYC “Venice” CPUs, AMD Pensando Vulcano NICs, and is further enhanced by the versatile open ROCm software ecosystem.
Collectively, these sophisticated technologies create a formidable rack-scale AI platform, meticulously engineered to power sovereign AI factories and state-of-the-art data center environments.
AMD states that the architecture is optimized to provide exceptional compute density and superior networking performance, crucial for managing extensive AI and machine learning workloads. Dr. Lisa Su, Chair and CEO of AMD, commented, “With ‘Helios,’ we are delivering an open, rack-scale AI platform designed for performance, efficiency, and long-term flexibility.”
The Helios platform boasts a highly flexible architecture, allowing it to be customized to perfectly align with the unique requirements of individual customers. This adaptability encompasses support for hybrid cloud deployments, advanced networking capabilities, and finely tuned software stacks specifically for AI and analytical tasks.
A key outcome of this collaboration will be the introduction of an AI-ready data center blueprint, engineered to support an impressive capacity of up to 200MW. This blueprint will serve as a scalable foundation for hyperscalers, large enterprises, and AI-focused companies aiming to construct or expand their next-generation data centers. Furthermore, AMD and TCS will actively partner with these hyperscalers and AI firms to expedite the development of new data centers within India.