Bengaluru-based Generative AI (GenAI) startup Gnani.ai has announced the launch of its speech-to-speech large language model (LLM), built in collaboration with NVIDIA, capable of handling over 10 million voice interactions daily. The breakthrough model, which supports 14 languages, marks a new era in multilingual voice processing technology.
The innovative speech-to-speech LLM was developed on NVIDIA’s AI-accelerated computing platform, using NVIDIA Hopper GPUs and the NVIDIA NeMo platform. The model has been trained on more than 14 million hours of proprietary multilingual conversational data. By combining NVIDIA’s advanced technology with Gnani.ai’s expertise in conversational AI, the startup aims to redefine enterprise customer interactions.
“The speech-to-speech LLM is transforming the way enterprises handle their most critical customer interactions. By combining NVIDIA accelerated computing with our conversational AI expertise, we are helping businesses achieve scale and efficiency with a business impact of over $6 Bn,” said Ganesh Gopalan, Co-Founder of Gnani.ai.
Since its inception in 2017 by founders Ganesh Gopalan and Ananth Nagaraj, Gnani.ai has developed speech engines for 10 Indian languages and built a diverse client base including TVS Credit, Bajaj Group, Fibe, and Muthoot Finance. The startup has also filed over 12 patents, establishing itself as a leader in India’s growing AI ecosystem.
Gnani.ai’s model integrates NVIDIA software components such as TensorRT-LLM, Triton, and Riva, enabling real-time performance and enhanced interaction capabilities. Currently, Gnani.ai works with more than 150 enterprises in India and the United States, primarily in the banking, financial services, and insurance sectors. Its Automate365 virtual assistant platform also enables customer interaction automation across voice, email, and WhatsApp.
At the recent NVIDIA AI Summit in Mumbai, NVIDIA CEO Jensen Huang emphasized the importance of India’s AI ecosystem and the strategic value of developing large language models for Indian languages. “If India can figure out how to make a large Hindi language model then you can build it in any part of the world,” Huang said, reinforcing the nation’s role in NVIDIA’s global strategy.
India has rapidly emerged as a leader in AI adoption, with a recent analysis by Inc42 projecting the Indian GenAI market to reach $17 billion by 2030. The growth is fueled by initiatives such as the IndiaAI Mission, approved by the union cabinet, with a five-year investment of INR 10,372 crore. The IndiaAI Datasets Platform is also slated for launch in January 2025, providing vital support to AI innovation in the country.
With over 100 GenAI startups raising $600 million since 2019, India’s focus on artificial intelligence has positioned the country as a global hub for AI development. Gnani.ai’s partnership with NVIDIA adds to this momentum, showcasing India’s potential to lead advancements in multilingual AI applications.