Despite the global shortage of NVIDIA’s much-sought-after graphics processing units (GPUs), Darshan Hiranandani-led Yotta Infrastructure, a data centre and services firm owned by Mumbai-based Hiranandani Group, has come up with a new trick to counteract the challenge.
For less than 10% off the global average price, Yotta offers companies using its NVIDIA processing chips as opposed to what other companies pay for accessing NVIDIA GPUs that are used in their artificial intelligence (AI) needs.
“We have recently dispatched the first batch of 4,000 NVIDIA GPUs at our Navi Mumbai data centre; consequently, most, if not all, are already exhausted by enterprises. We have requested NVIDIA to expedite delivery of the remaining 12 000 GPUs so that we can complete placing an order based on anticipated demand for the AI cloud platform,” Sunil Gupta, co-founder and CEO of Yotta, said.
With Darshan Hiranandani at the chair, in December, Yotta announced that it had placed an order for 16,000 NVIDIA H100 GPUs. The company’scompany’s order and project to build cloud infrastructure that supports training of AI models were worth “”almost $1 billion.””.
On this issue, Gupta stands out from industry veterans who argue that India’s generative AI market has yet to mature adequately regarding conservative tech spending. Nevertheless, with nearly a 90% undercutting strategy against competition, Yotta aims to attract customers from North America and Europe, where 70% of its AI cloud clients currently come from global markets.
NVIDIA H100 Tensor Core GPUs—World’sWorld’s Fastest—were recently added into Yotta’sYotta’s NM1 data center, thereby setting a new yardstick for AI development capability across India. By 2025, they plan to upscale their inventory till exceeding over thirty-two thousand gpus towards enhancing AI operations.
Strategically deploying H100 tensor core GPUs at Yotta’s NM1 not only upgrades India’s AI development setup but also makes the country a significant contender in the global technological race for AI superiority.
By increasing access to GPU resources and enabling innovation and competitiveness across sectors with its vision, Yotta Infrastructure wants to double down. The Shakti Cloud platform will have foundational AI models and applications that empower Indian companies to build robust AI tools and products quickly, reducing the time-to-value of their AI-powered services.
NVIDIA’s H100 GPU has been designed specifically for AI applications in the Hopper architecture. It incorporates 80 billion transistors, enabling it to process massive amounts of data instantly compared to its predecessor, the A100, which had six times fewer transistors.
Stewarded by Darshan Hiranandani, Shakti Cloud is India’sIndia’s largest and fastest AI-HPC supercomputer, fueled by the world’sworld’s fastest NVIDIA H100 Tensor Core GPUs. These GPUs are poised to offer state-of-the-art GPU computing infrastructure, platforms, and services, including IaaS (infrastructure-as-a-service), PaaS (platform-as-a-service), and SaaS (software-as-a-service).
Yotta’sYotta’s initiative would enable Indian businesses, government entities, start-ups, and researchers to accelerate innovation and productivity, thereby contributing heavily to India’sIndia’s position within the global landscape of artificial intelligence (AI). The Indian Cabinet has allocated Rs 10,372 crore over five years under the India AI Mission, indicating a commitment towards promoting AI.
Yotta Infrastructure’s unparalleled approach of selling NVIDIA GPUs at a fraction of the market cost, however, is coupled with an ambitious plan for expanding its inventory and positioning it as a significant driver of AI in India. The deployment of NVIDIA H100 GPUs at its data centres is designed to promote access to GPU resources for all and speed up India’s global AI race.