Sarvam Hires Google Cloud India MD as President in Enterprise Push

AI unicorn Sarvam has appointed Sashikumar Sreedharan, until recently managing director of Google Cloud India, as its president and chief business officer. The hire signals that a company valued at $1.5 billion now intends to sell as seriously as it researches.

Sreedharan will run Sarvam's business across enterprises and government institutions and expand its engagement with digital-native companies, developers and partners. His remit also covers marketing, finance, forward-deployed engineering, service delivery and commercial operations, according to a company blog post. He joins the board of directors as well.

His tenure at Google Cloud India lasted just over a year. The wider career spans roughly 35 years, with stops at Microsoft India, IBM, Wipro and Happiest Mind Technologies. That is a conventional enterprise-software pedigree attached to a very unconventional company.

Why does a research lab need a chief business officer?

The appointment comes as Sarvam tries to move beyond building indigenous AI models and construct a durable enterprise business around them. The startup says it already provides full-stack AI offerings to more than 100 enterprises and government institutions. A hundred logos is a start; a hundred renewals is a business.

Sarvam framed the move in expansive terms. "The next phase of Sarvam is about advancing our research, developing products loved by people around the world, and building a company that can operate at a much greater scale," the company said. Scale, in this reading, is a commercial problem before it is a technical one.

Founded in 2023 by Pratyush Kumar and Vivek Raghavan, Sarvam develops foundational AI models, computing infrastructure and enterprise applications. Its products target businesses and government agencies that want tools tailored to Indian languages, Indian data-security requirements and local use cases. It is, in effect, betting that India deserves its own AI stack rather than a translated one.

What has Sarvam actually shipped?

The commercial buildout rests on a widening portfolio. Most recently, Sarvam launched Prism, a "System One" decision model built on top of its Sarvam 105B foundation model. Unlike a conventional large language model that generates open-ended text token by token, Prism evaluates a context, a question and a predefined list of options in a single processing pass, returning calibrated probabilities for each choice. For structured enterprise decisions, that is a meaningful departure from chatbot mechanics.

Other releases have arrived in steady succession:

• Sarvam Vision 2.1, a vision-language model for document processing and text recognition in Indian languages.
• Saaras V4, the company's speech-to-text model.
• Bulbul V4 Flash, a text-to-speech model launched in October for generating natural speech across Indian languages.

The pattern is consistent: cover the input and output surfaces of Indian-language computing, then wrap services around them. Sreedharan's job is the wrapping.

How much money is behind the push?

Earlier this year, Sarvam raised $234 million in the first close of a $300 million Series B round that valued the company at $1.5 billion. The round drew participation from HCLTech, Bessemer Venture Partners, Khosla Ventures and Peak XV Partners. That round minted the unicorn.

The fundraise may not be finished. Sarvam is reportedly in talks with Nasdaq-listed IT services company Cognizant Technology Solutions to bring in further funds for the round. The reported target is around $300 million at a valuation of $3.5 billion to $4 billion. A doubling of valuation inside a year would be aggressive by any standard, and these figures remain report-stage numbers rather than confirmed terms.

What comes next?

The appointment places a veteran of India's enterprise-software trade in charge of converting a model catalog into contracted revenue. If the Cognizant talks conclude near the reported valuation, Sreedharan will have both the capital and the customer relationships to test whether the India AI stack can carry enterprise weight.