Indian AI startup Sarvam AI has unveiled its latest large language model optimized for India's diverse linguistic landscape, prompting questions about adoption by local startups.
Founders highlight where the Sarvam LLM excels in handling Indian accents, languages, and cultural nuances but falls short in some advanced global benchmarks.
Sarvam AI's Rise in India's Sovereign AI Push
Founded in 2023 by Pratyush Kumar and Vivek Raghavan, Sarvam AI quickly raised $41 million in Series A funding from Lightspeed, Peak XV Partners, and Khosla Ventures.
The company has developed foundation models supporting 22 Indian languages, including Hindi, Tamil, and Bengali, marking a milestone in localized AI.
Recent Launches and Partnerships Boosting Impact
In February 2026, Sarvam launched 30-billion and 105-billion parameter models, along with text-to-speech, speech-to-text, and vision capabilities trained from scratch.
Key partnerships include Swiggy for multilingual voice-led ordering and the launch of the Indus AI chat app for web and mobile users.
Government backing via the IndiaAI Mission provided compute resources and viability gap funding, accelerating Sarvam's sovereign AI ambitions.
The impact extends to public sector applications, like voice agents making millions of calls for policy renewals, demonstrating real-world scale.
However, challenges persist as Indian startups weigh Sarvam against global giants like ChatGPT in terms of cost, performance, and seamless integration.
Founders emphasize that adoption will hinge on factors like affordability, data sovereignty, low-latency edge computing, and developer-friendly APIs.
Looking ahead, Sarvam is in talks for $200-250 million funding from investors including Nvidia, signaling ambitious scaling plans for 2026.
As India's AI ecosystem matures, Sarvam's success could inspire a wave of homegrown models, reducing dependency on foreign tech.
Maria Lourdes
