India eyes to have 50,000 GPUs in 3 years to storm into the AI league

India, a late entrant to the AI celebration, is making efforts to make up for the misplaced floor to be within the high league. Abhishek Singh, Further Secretary within the Union Ministry of Electronics and Info Know-how (MeitY), mentioned that the nation would have at the very least 50,000 GPUs within the subsequent 2-3 years, giving AI start-ups the much-needed firepower. GPUs are on the coronary heart of the computation spine for AI purposes.

“We’re lagging in computing energy and analysis papers. The federal government itself commits for 10,000 GPUs, whereas Yotta introduced plans to construct a 25,000-GPU clusters and Tatas and Jios had introduced plans to construct 50,000 GPUs. We don’t know their timelines however we needs to be reaching a minimal of fifty,000 GPUs within the subsequent 2-3 years,” he mentioned. “The response to the RFPs for establishing computing infrastructure has been excellent. As many as 50 firms confirmed curiosity. That ought to tackle the computing infrastructure challenge,” he added.

Speaking to reporters on the sidelines of the two-day International AI Summit that started right here on Thursday, he mentioned that the Mission launched an software growth initiative the place 14 drawback statements had been recognized throughout the 5 themes of training, agri, healthcare, flood forecast, local weather change, and governance.

“For every drawback assertion, we introduced a prize cash of₹1 crore. The AI Mission has been accepted. It backs all of the seven key pillars together with computing, foundational fashions, datasets, and start-ups. As many as eight out of 200 start-ups that submitted concepts to deal with some key challenges have been recognized for funding help,” he mentioned.

Seven pillars

Highlighting the seven pillars of the nation’s AI Imaginative and prescient, he mentioned a key focus of the mission is to enhance the nation’s laptop infrastructure and supply inexpensive computing assets to start-ups, researchers, and college students.

Although the nation has about 5 lakh datasets, not a lot of them are usable (by AI firms), he mentioned. “Many of those datasets aren’t AI-usable because of challenges associated to information standardisation, privateness, and accessibility. To handle this, an information administration framework is being labored out. We shall be working with varied departments to make related information usable and out there,” he mentioned.

He additionally touched on India’s function within the International Partnership on AI (GPAI), highlighting the nation’s efforts to signify the International South in AI governance discussions.  He mentioned AI-powered options might assist tackle societal challenges and bridge the digital divide, significantly for the five hundred million unconnected people within the nation. They need to have the ability to entry the app-based providers utilizing audio options.

Moreover providing computing energy at subsidised prices, there’s a must construct use circumstances in sectors reminiscent of agriculture, training and healthcare.