The IndiaAI mission is further bolstering its efforts to establish indigenous foundation artificial intelligence (AI) models with the inclusion of Gan.ai, adding to the previously announced collaborations with Sarvam.ai, Soket AI, and Gnani.ai. This announcement highlights the mission’s commitment to fostering a robust and self-reliant AI ecosystem within the country.
The IndiaAI mission, a flagship initiative of the Indian government, aims to develop and deploy foundational AI models trained on India-specific data, catering to the unique linguistic and cultural diversity of the nation. The selection of Gan.ai, following Sarvam.ai, Soket AI, and Gnani.ai, signifies a significant step forward in achieving this goal.
Details of the New Collaborations:
- Soket AI: Will develop India’s first open-source 120-billion parameter foundation model. This model will be optimized for the country’s linguistic diversity and target critical sectors such as defense, healthcare, and education.
- Gnani.ai: Is tasked with building a 14-billion parameter multilingual Voice AI foundation model. This model aims to deliver real-time speech processing with advanced reasoning capabilities across multiple Indian languages.
- Gan.ai: Joins this cohort to create a 70-billion parameter multilingual foundation model. Gan.ai’s focus will be on achieving superhuman text-to-speech (TTS) model capabilities, setting a high benchmark for indigenous AI in this domain.
- Sarvam.ai: Was previously announced as the first startup to build an indigenous foundational model. Their multi-modal, multi-scale model will also have 70 billion parameters, focusing on reasoning, voice capabilities, and fluency in Indian languages for population-scale deployment.
This multi-pronged approach, as emphasized by the IndiaAI mission, aims to ensure that India develops AI models that are not only technologically advanced but also deeply rooted in the Indian context. By supporting multiple teams with distinct focuses, the mission intends to cover a wider spectrum of AI capabilities, from large language models to voice AI and text-to-speech technologies.
In parallel to these developments in foundation models, the IndiaAI mission has also significantly scaled up the nation’s AI compute infrastructure. The IndiaAI Compute portal now boasts over 34,000 GPUs, sourced from leading providers like Nvidia, AMD, AWS, and Intel. This enhanced compute power will be crucial in training these large AI models and making AI resources more accessible to startups, researchers, and enterprises across India.
Union Minister for Electronics and Information Technology, Ashwini Vaishnaw, has highlighted the importance of democratizing access to AI and ensuring that India plays a leading role in the global AI landscape. The selection of these startups and the expansion of compute capacity are key steps in realizing this vision.
Key Highlights:
- The IndiaAI Mission has added Gan.ai to its initiative to build indigenous foundation AI models, alongside Sarvam.ai, Soket AI, and Gnani.ai.
- These startups will focus on developing multilingual large language models, voice AI, and advanced text-to-speech capabilities tailored for India.
- The IndiaAI Compute portal now offers access to over 34,000 GPUs to support the training and deployment of these AI models.
- This effort aims to create a self-reliant AI ecosystem in India, aligned with the nation’s linguistic and cultural diversity.