Govt’s big AI push

reportedaboutthecabinet’slikely approval to the project and that the government was looking to fundthedevelopmentofcomputing capacity in the country by setting up 10,000-30,000 graphics processing units (GPUS) through viability gap funding.

Last year, Prime Minister Narendra Modi had announced the Mission and said its aim was to establish the computing powers of AI within the country. This, he had said, will provide better services to startups and entrepreneurs and also promote AI applications in the sectors of agriculture, healthcare and education.

Undertheindiaaimission,the government will look to establish a computing capacity of more than 10,000 GPUS and also help develop foundational models with a capacity of more than 100 billion parameters trained on datasets covering major Indian languagesforprioritysectorssuch ashealthcare,agricultureandgovernance.aicurationunits(acus) will also be developed in 50 government departments. The proposal also includes the establishment of an AI marketplace designed to offer AI as a service and pre-trained models to those working on such applications.

The implementation of this AI compute infrastructure will be done through a public-private partnership model with 50 per cent viability gap funding. If the compute prices come down, the private entity will have to add more compute capacity within the same budgeted amount to meet increased demand. Of the total outlay, Rs 4,564 crore has been earmarked for building computing infrastructure.

“Basically, there will be a tender inviting companies to set up data centres. When a company applies for, let us say a centre which may cost Rs 10,000 crore, theycanseekaviabilitygapfundingfromthegovernmentforacertainamountofthat,”aseniorgovernment official said.

As part of its funding, the government will also introduce conditionsthatthecomputingaccess begiventostartupsandacademic institutions at concessional rates.

During the setting up of infrastructure, priority will be placed on selecting the most advanced GPUS. However, it is worth noting that Nvidia’s A100 chip — considered to be the most cutting edge foraiapplications—costsaround $10,000,meaningthatadatacentreof10,000suchgpuscouldcost at least $100 million (roughly Rs 8,000 crore).

Computing capacity, or compute, is among the most important elements of building a large AI system apart from algorithmic innovation and datasets. It is also one of the most difficult elements to procure for smaller businesses looking to train and build such AI systems.

The government will also finance deeptech startups at various levels of growth.