Yotta Data Services is preparing to provide artificial intelligence (AI) computing capacity to startups, large IT firms, the Indian government, and even foreign enterprises at competitive rates, which it asserts could be the most economical globally.
“Current GPU (graphics processing unit) usage rates among global hyperscalers range from $9 to $12 per GPU hour,” stated Sunil Gupta, CEO of Yotta Data Services. “However, with our distinctive market position, we could offer rates as low as $1.8 per GPU hour for multi-year contracts, compared to $5 per GPU hour for short-term use.”
Gupta emphasized that as one of the select priority partners of global AI superchip-maker NVIDIA, Yotta is uniquely positioned to provide GPU access to global enterprises, particularly at a time when countries are facing challenges in acquiring GPUs due to supply constraints.
“High-compute processor chips, or GPUs, are scarce resources globally. Acknowledging the potential AI presents, Yotta took proactive steps early on by establishing a partnership with NVIDIA,” Gupta informed the media on the sidelines of the NVIDIA GPU technology conference in San Jose, California.
He added that Yotta operates within special economic zones, enjoying zero taxation and import duties on AI hardware, which further reduces costs. “Our group company, H-Energy, supplies power to our data centers, contributing to lower energy sourcing expenses.”
With its AI compute infrastructure, comprising both GPU hardware and the software stack, Yotta aims to rapidly scale up by offering affordable subscription rates. “We aim to generate revenue through high volumes and rapid adoption, even if it means extending our payback period,” Gupta remarked, anticipating Yotta to break even within 2-3 years on its $1 billion investment in the NVIDIA GPU deal.
More on Yotta Data Services
Yotta is affiliated with the Nvidia Partner Network and offers its sovereign AI cloud platform called the Shakti Cloud platform, catering to the end-to-end data storage needs of governments and enterprises. Last year, Yotta announced plans to import 24,000 GPUs, including the popular NVIDIA H100s and L40S, in phases. The first batch arrived in March, with subsequent shipments expected every three days at the Navi Mumbai campus. Gupta disclosed plans to deploy 16,384 H100s and 1000 L40S by May 15th, with future imports potentially including NVIDIA’s latest Blackwell chips.
The company is exploring equity or debt financing to fulfill the order and is in advanced discussions with private equity investors and banks. Yotta has attracted interest from AI startups like Ola’s Krutrim and Sarvam AI, as well as customers in Europe and the Middle East.
Furthermore, Yotta will engage in the Center’s AI Mission, wherein it will supply compute capacity to the government for distribution to R&D facilities, startups, and academic institutions through a marketplace. The government has allocated Rs 10,000 crore for this initiative.
See also: First Global Artificial Intelligence Resolution Adopted by the UN
