Indian startup Sarvam AI raises $75 million in Series B funding, with Nvidia among backers, to build a pioneering trillion-parameter AI model in India, signalling a significant step towards indigenous frontier AI development.
Sarvam AI has raised $75 million in Series B funding, with Nvidia among the backers, in a move that comes only days after the Indian startup said it planned to build a 1 trillion-parameter foundation model from the ground up in India. The deal values the company at about $1.5 billion and signals that one of the world’s most important chipmakers is willing to put money behind India’s bid to build frontier AI, not just consume it.
The timing matters because large language models at this scale are expensive to train and even more expensive to run. Sarvam says its planned model will be trained on 10,000 Nvidia Blackwell GPUs, the latest high-end chips designed for advanced AI workloads. That makes access to Nvidia’s hardware and software ecosystem more than a nice-to-have: it is central to whether the project is commercially and technically viable.
Pratyush Kumar, Sarvam’s co-founder, told The Economic Times that the Nvidia relationship has been building for years, with the startup working on the company’s AI stack and contributing improvements back to its software. He described the funding as a natural extension of that technical partnership. According to Sarvam, this is the first investment Nvidia has made in an Indian company of its kind.
For Indian businesses and developers, the bigger takeaway is what Sarvam says it wants to build around the model. The company is not just chasing a headline-grabbing parameter count; it is trying to create a full AI ecosystem covering foundation models, computing infrastructure and tools for coding, cybersecurity, scientific research, defence and enterprise use. It is also part of the government’s IndiaAI Mission, where it was chosen earlier this year to develop the country’s first sovereign foundation model, with an emphasis on Indian languages and local use cases. If it succeeds, the effort could matter for firms that want AI systems trained closer to Indian needs, rather than adapted from overseas products.
Disclaimer: This article is intended to inform and educate, not to recommend or endorse any financial product, investment or strategy. Please consider your own financial circumstances and seek professional advice where appropriate before making financial decisions.
