Centre to fund up to 50% of AI compute infra to be set up in partnership with industry

The government will fund up to 50 per cent of AI compute infrastructure that will be created in partnership with the private sector, a senior official said on Friday.

While speaking at a CII Annual Business Summit, Ministry of Electronics and IT (Meity) Secretary S Krishnan said the government is looking to set up at least 10,000 graphics processing unit (GPU) based compute capacity in the country soon.

"The government is prepared to provide for at least up to about 50 per cent of the cost of creating this infrastructure. In order to speed up the entire process. The idea is to partner with a variety of private institutions so that we are able to work with the private sector in making this capacity available quickly," Krishnan said.

Meity has set a target to make available GPU-based servers approved under the IndiaAI Mission by March 2026. The Cabinet has approved the India AI Mission with an outlay of Rs 10,372 crore for five years to encourage AI development in the country.

The rapid development of AI across the globe has led to an increase in demand for GPU-based servers as they can process data at a higher speed compared to CPU-based servers.

Krishnan said there is an independent capacity for the government's own use, which is being attempted through the National supercomputing mission and the GPU capacity under India AI mission will be set up separately.

"We want to make it (AI compute) available both on a VGF (viability gap funding) basis where new capacity will be created, and there would be viability gap funding from the government or alternatively," he said.

Krishnan said a voucher-based model is also being considered. He said the other key element is to create a foundation model specific to India. "We are all aware of the issues with the foundation models. We are also aware that even in the last 12 months, the way people are thinking about foundation models is changing.

"Now people are talking about large language models, large multimodal models. Some of them are becoming smaller and more efficient. A number of those efforts are going on in our conversations with the industry, private sector, and with a variety of experts," Krishnan said. Asked about apprehension around AI on jobs, Krishnan said there is a concern but there is no clear answer to it.

"Many people feel that the newer AI jobs which will come, will get created in India for two or three reasons. One, we have a large number of STEM-trained graduates, many of whom have exposure to AI far more than any other country in the world.

"The second aspect is that we yet don't have as many white collar jobs in the services sector that many of the other advanced countries have which will get more readily displaced.

"Both of these, I think, give rise to some hope and expectation that in the job loss scenario it may not be as bad a situation as it is in certain other economies," he said.