Gnani.ai, one of the country's established artificial intelligence firms founded in 2017, has announced the launch of Gnani Artha. This end-to-end sovereign AI stack is built for Indian enterprises and public institutions and is backed by the government's India AI Mission.
The newly launched platform consists of two primary components. The first is Gnani Evon v3.3, a 30-billion-parameter open-weights model trained natively across 11 Indian languages. The second component is Gnani Plexus, which serves as the company's agentic AI platform. According to the company, the model uses tokens significantly more efficiently for languages like Gujarati and Malayalam compared to global alternatives.
In an interview regarding the launch, Ganesh Gopalan, Co-Founder and Chief Executive Officer of Gnani.ai, stated that the company plans to expand its family of models in the future. Future iterations include planned launches of 70-billion-parameter and over 100-billion-parameter models tailored for different use cases. Gopalan also noted that the company's Prisma speech-to-text model continues to process multiple Indian dialects and languages using a dedicated data set.
Addressing market competition and the entry of global players offering free models, Gopalan remarked that the current market is expanding rapidly at 'day zero' of AI adoption. He noted that the company occasionally turns down business requests due to capacity constraints in sales and support teams, emphasizing a focused approach over monitoring competitors.
On the topic of upcoming government regulations for AI, Gopalan expressed support for balanced policies. He stated that while regulations are necessary to address issues such as deepfakes in a large democracy, the framework should be implemented carefully to avoid stifling the domestic technology sector.
"The launch of sovereign AI stacks like Gnani Artha highlights the growing maturity of India's domestic technology ecosystem. Building models natively trained in multiple regional languages addresses a critical gap left by global alternatives, which often struggle with local dialects and token efficiency. For Indian enterprises and public institutions, localized infrastructure ensures data security and relevance. As the AI sector scales rapidly, companies that focus on domain-specific execution rather than short-term market noise will secure long-term viability." — Dr. Shishir Gupta, Founder & CEO, StartupLanes
Recent StartupLanes Articles
Browse through our 30 latest publications on venture capital, startups, and angel investing.