In Short:
Industry experts suggest that the government should adopt a marketplace model for AI computing infrastructure, allowing private players to compete in offering hardware and cloud solutions. The government’s role should be that of an enabler, acquiring compute capacity from hyperscalers and offering it at subsidized rates to startups and researchers. Private investments in data centers and GPU infrastructure should be encouraged for a more efficient and competitive market.
Government’s AI Computing Initiative to be Driven by Private Players
The government’s initiative to make significant AI (artificial intelligence) computing capacity available to innovators can be best served through a marketplace model where private players compete to offer hardware, cloud solutions, and other services, while the government becomes an enabler, industry experts said.
Cabinet Approval for India AI Mission
In early March, the union cabinet approved a Rs 10,371.92-crore India AI Mission that aims to build AI computing infrastructure of at least 10,000 graphics processing unit (GPUs) through public-private partnerships in five years.
Instead of spending top dollars in acquiring and installing GPUs, experts suggest that the government could acquire compute capacity from hyperscalers and offer it at subsidised rates to startups and researchers.
Private Sector Participation
Tanuj Bhojwani, from people+ai, suggested that the government should think like an investor and conduct pooled procurement negotiations. This would allow private players to buy compute capacity at negotiated rates set by the government.
Data center operator and cloud services company Yotta expressed its willingness to participate in the India AI Mission by offering GPU usage at globally competitive pricing.
Private Sector Should Lead in GPU Infrastructure
Experts emphasize that private players are better positioned to compete and invest in GPU infrastructure deployment. They suggest that the government should provide subsidies and enable the private sector rather than setting up GPU centers.
Enabling Usage from Different Sources
Vishnu Vardhan, CEO of Seetha Mahalaxmi Healthcare (SML), highlighted the importance of enabling GPU usage from various sources to avoid monopolies and high costs.
Government’s Role as an Enabler
Anushree Verma from Gartner suggested that the government should support startups and research institutions by offering AI compute capacity or managed services at subsidized pricing.
Optimizing Tax Dollars Investment
Experts recommend that the best return on tax dollars would be facilitating private investments in AI computing infrastructure.
Future Recommendations
Experts recommend creating a big, pooled capacity for data centers and maintaining an interoperable layer for better utilization of resources.
