IGI Logo

We hosted Dr. Pratyush, Co-Founder, Sarvam AI, as part of our CEO track session at AGIC 2026. Here are the key takeaways of the session

Full-stack approach is central to Sarvam's strategy

Sarvam is building a complete AI stack, from infrastructure and foundation models to enterprise applications. The company believes owning every layer allows it to capture a larger share of value rather than becoming only a low-margin AI infrastructure provider. It has already built a 100b-parameter foundation model from scratch and plans to continue scaling up its models while selectively using open-model distillation where it makes sense

Sovereign AI can lower costs and improve data security

Sarvam believes enterprises increasingly want greater control over their AI infrastructure, data and costs. Running AI workloads on domestic infrastructure reduces dependence on global model providers while keeping sensitive enterprise data inside the country. The company cited examples where its models offered materially lower inference costs than global alternatives, making AI deployment more economical for enterprises, financial institutions and government agencies.

India needs risk-capital, R&D DNA and long-term investment

According to Sarvam, building India's AI ecosystem will require nearly USD10b of focused investment across infrastructure, models, research and talent development. The company believes India does not need to match the capital spent by global hyperscalers, as model training costs have declined significantly over the last few years. However, stronger R&D spending, risk capital, government support and talent development will be essential to create globally competitive AI companies. As part of this effort, Sarvam plans to launch ‘Sarvam Forward’ on 5 th Sep’26 to expand AI education and training across campuses in India.

Enterprise AI adoption is already gaining traction

The current AI wave began with ChatGPT in 2022, making this only the fourth year of the large language model era. While the US and China hold an early lead, Sarvam believes India still has ample opportunity to build globally competitive AI companies over the next decade. Sarvam's enterprise AI offerings, including Voice AI, Document AI, Sarvam Code and enterprise copilots, are seeing growing adoption. Voice AI processed over 325m call minutes in the past year, while Sarvam APIs handled around 400m API calls. AI is also reducing customer acquisition costs, improving software development productivity, and automating customer support.

For More Research Reports : Click Here 

For More Motilal Oswal Securities Ltd Disclaimer
http://www.motilaloswal.com/MOSLdisclaimer/disclaimer.html
SEBI Registration number is INH000000412