AMD, the computer chipset company, anticipates that the rollout of its newest GPU-driven high-performance computing platform, Helios, will commence in the latter half of this year in various countries, including India, according to reports.

The firm seeks to leverage the growing need for AI and suggested expansion of data centre capacity by international tech leaders.

Mahesh Balasubramanian, senior director for data centre GPU product marketing at AMD, informed PTI that one Helios rack combines 72 linked MI455X (AMD GPUs) accelerators, providing a maximum of 2.9 exaflops of FP4 (format utilized in AI) computational power per rack.

The anticipated schedule for deployment pertains to worldwide customer installations of Helios systems, which includes India. Balasubramanian states that initial deployments are anticipated to start in the latter half of 2026.

The need for GPUs (Graphics processing units) has surged due to the immense global demand for AI, with US chip manufacturer Nvidia leading the market, holding more than 80 percent share.

The Helios Platform will rival Nvidia Vera Rubin POD, which asserts it can offer a compute speed of 3.6 exaflops, though at a relatively steeper price.

AMD has teamed up with TCS to jointly create a rack-scale AI infrastructure blueprint utilizing the AMD Helios platform to bolster India's national AI efforts.

Also Read: NVIDIA's Rubin Moment at CES: AI Supercomputing Breakthrough

Through this strategic partnership, the two firms will provide an AI-optimized data center blueprint that accommodates up to 200 MW of capacity and will collaborate with hyperscalers and AI firms to speed up data centre developments in India.

Also Read: AI Autonomy Race: How Advanced are Top Countries' AI Strategies?

Google, Amazon, Microsoft, Reliance Industries, and Bharti Airtel have revealed intentions to establish new data centres in India, with a total investment nearing $ 200 billion.

Also Read: From Silos to Synergy: Building the Future With Agentic AI

AMD is advocating for the use of open architecture in constructing data centers rather than relying on the proprietary systems of its rivals.