IndiaAI Mission appears to be planning to provide common compute capacity through about 14,000 graphics processing units (GPUs).
IndiaAI Portal Launch
For doing so, they have shortlisted ten companies which have met the lowest bids for different GPU types as informed by a senior government official on the condition of anonymity.
Out of these,10,000 GPUs are already available with the companies such as Yotta Data Services, E2E Networks, Tata Communications, and AWS’s managed service providers.
They plan to purchase the remaining 4,000 GPUs by companies including Jio Platforms and CtrlS Datacenters.
All in all, ten companies had proposed that they would provide 18,693 GPUs for the mission’s AI Compute pillar.
The bidders had until February 14 to meet the lowest bids which were earlier announced by the IT ministry on January 30 for different types of GPUs.
It appears that they have finally met the L1 bids for about 14,000 GPUs.
Reportedly, ₹10,371.92 crore ($1.25 billion) were approved for the IndiaAI Mission in March.
Out of these, 44% ( ₹4,563.36 crore) is earmarked for compute capacity through more than 10,000 GPUs over five years.
They are planning to empanel the companies for 36 months which will further be extendable by 12 months.
It seems that the seventy percent of the promised GPUs are high-end, such as Nvidia’s H100.
While, the remaining thirty percent are low-end GPUs as they come with lower capacity or older generations.
Out of these, Yotta Data Services accounts for the bulk of the compute capacity having 9,216 GPUs; it further includes 8,192 Nvidia H100s.
For this, AWS has done its bidding through four managed service providers including CMS Computers, Locuz Enterprise Solutions, Orient Technologies, and Vensysco Technologies.
The company will provide 1,200 low-end GPUs i.e. 800 AWS Inferentia 2 and 400 Trainium 1.
How Did This Happen?
Earlier, IndiaAI Compute Pillar sent a memo to all union ministries, departments, and chief secretaries.
The memo was regarding the discounted compute capacity, network, and storage services.
Basically, with this empanelment, they aim to serve Indian users including academia, researchers, MSMEs, startups, students, industry, and government departments and agencies.
IndiaAI Mission has plans to cover up to 40% of the total computer costs for “eligible users” as per the information provided by the February 7 memo signed by Abhishek Das, general manager of IndiaAI Compute Pillar.
So far, the lowest bids averaged 42% below market price. Hence, costing companies ₹115.85 and ₹150 per GPU hour for low-end and high-end compute capacity, respectively.
This will further reduce to ₹100 per GPU hour for certain users as the additional government subsidy will be provided.
Further, they plan to form a committee to evaluate subsidy proposals, with the amount of subsidy varying by user category, as per the sources.
This comes under the projects of “national importance” as its main aim is to benefit citizens who could receive the maximum 40% subsidy.
