As the race to build sovereign AI capabilities intensifies, India is seeking to move beyond exporting talent to building its own AI ecosystem. The IndiaAI Mission, which backs domestic compute infrastructure and indigenous foundation models, reflects that ambition.
But if the reported proposal takes shape, what it could mean for Sarvam’s competition, industry innovation and the future of India’s AI ecosystem?
Why Sarvam?
Sarvam is among a handful of companies selected by the Centre to develop indigenous, multilingual and domain-specific foundation models under the IndiaAI Mission, as part of its strategy to strengthen the country’s AI capabilities. Under the programme, the government is providing the selected companies access to graphics processing unit (GPU) compute at subsidised rates, covering up to 40 per cent of the cost.
Founded in 2023, Sarvam is building a full-stack AI business spanning foundation model development, inference infrastructure and enterprise applications, with a focus on Indian languages and domestic use cases. The company says its products are being deployed across sectors such as banking, insurance, government services, and defence.
Why could the government take equity?
What makes the proposal unusual is not the government’s support for Sarvam, but the form that support could take. Unlike conventional government support programmes, the proposed investment is not a direct cash infusion. The Centre’s stake would arise from compute infrastructure and other resources provided under the IndiaAI Mission, with the equity likely to be acquired through convertible instruments, The Economic Times had reported.
Experts say the move reflects a broader shift in how governments are beginning to view AI.
“An investment by the government into Sarvam AI could signify a paradigm shift in how India supports the creation and commercialisation of AI technologies,” said Kumar Rajagopalan, vice-president, strategic initiatives and India country head at Dexian. “It suggests the government wants to move from merely providing AI infrastructure to becoming a stakeholder in the country’s AI ecosystem.”
Rajagopalan said the proposal also reflects the government’s commitment to developing sovereign AI capabilities, which will support additional investments into cloud-based AI infrastructure.
What would this mean for India’s AI ecosystem?
Rajagopalan argued that equity investments should be reserved for projects with strategic national importance, while the government’s primary focus should remain on expanding shared AI infrastructure, such as compute capacity, high-quality datasets, cloud resources and research.
“Rationalising equity investments should only be given priority where there are strategic national implications and only if there are clear, fair and transparent policies governing their use,” he said.
Jaspreet Bindra, co-founder and chief executive of AI & Beyond, said government equity can be justified where scarce compute materially de-risks a frontier AI company, “but it should be the exception, not the model”.
India’s priority, he said, should remain building shared AI infrastructure rather than picking national champions.
Ritwik Batabyal, chief technology and innovation officer at Mastek, an IT services and digital engineering company, said, “AI development, especially at the foundation model level, requires significant investments in computing power, talent and research. If government support helps reduce these barriers, it could accelerate innovation.”
“At the same time, such initiatives should be backed by clear and transparent guidelines so that the wider startup community also has access to similar opportunities,” Batabyal said. “A healthy AI ecosystem is one where innovation is encouraged across the board, rather than concentrated in a few companies.”
