The Indian government is making sure it has a stake in the key home-bred players of the AI revolution.
An Indian government body will take an equity stake in Sarvam AI, which is the first company selected under the country’s AI mission to build a foundational AI model. This was clarified by Sarvam AI co-founder Pratyush Kumar on X after several sources had reported that the government would provide compute as a “grant”.
“So you’re telling me that Deepseek with private funds can release an open source model, but govt awarding Rs 220 crores of public funds to Sarvam isn’t asking for the same? This is tax payers money, so the full pipeline ought to be open source!” Wingify founder Paras Chopra had said on X.
After the post gained some traction, with several people wondering why Sarvam AI’s model wouldn’t be open source, its co-founder came up with a clarification. “This is not a grant,” said Pratyush Kumar. “A government body will take equity in Sarvam for the compute we receive. And we are committed to building public interest use-cases and enabling the ecosystem in various ways such as hyper-optimising the inferencing costs in India. BTW, we as a country have a long way to go in AI. This is the time to strengthen each other. Let’s go!” he added.
The Indian government had earlier put out calls for proposals to build an indigenous foundational AI model from India. The government had said it would provide the GPUs to train the model. Access to these GPUs can be expensive and difficult, so the government has procured over 10,000 GPUs which it intends to provide to startups to help train their models. The government will provide Sarvam — the first company selected under the program — access to 4,096 high-end Nvidia H100 GPUs for six months to train its model.
And it could make sense for the government to own a slice of foundational models coming out of India. AI is poised to be an extremely disruptive technology, which could not only upend white-collar jobs, but also national security sectors like warfare and defense. The US already has AI companies that are collaborating with the government in these sectors, and in China the coupling between companies and the government is even greater. Both these countries have foundational AI models that are rapidly growing in capabilities. As such, it might be a strategic imperative for India to have its own foundational model, and one that the government is closely involved in. And the government taking equity in exchange for providing compute might be a good structure for building foundational Indian AI companies which could be critical for national strategic interests in the years to come.