India unveiled three Sovereign AI models at the AI Impact Summit | Current Affairs | Vision IAS

Upgrade to Premium Today

Start Now
MENU
Home
Quick Links

High-quality MCQs and Mains Answer Writing to sharpen skills and reinforce learning every day.

Watch explainer and thematic concept-building videos under initiatives like Deep Dive, Master Classes, etc., on important UPSC topics.

A short, intensive, and exam-focused programme, insights from the Economic Survey, Union Budget, and UPSC current affairs.

ESC

In Summary

  • India launched Sovereign AI models and plans to add 20,000 GPUs under the IndiaAI Mission, aiming for self-reliance in AI development.
  • The IndiaAI Mission, launched in 2024 with ₹10,371.92 crore, focuses on compute, application development, datasets, skills, financing, and safe AI.
  • New models include Sarvam AI's LLMs, Gnani.ai's Vachana text-to-speech model, and BharatGen's Param2 17B multilingual foundational model.

In Summary

Launched in the backdrop of IndiaAI Mission, these models highlight towards a clear push for India-first, Sovereign AI. 

  • Further, announcement on adding 20, 000 Graphics Processing Units (GPUs), beyond existing 38, 000 GPUs under the IndiaAI mission was also made. 

Sovereign AI 

  • Meaning: Ability to develop AI technologies on domestically based infrastructure, using local data and human talent, free from dependency on foreign platforms or corporations. 
    • It implies control over full AI lifecycle, data collection and training of models to deployment and governance of AI systems.

About IndiaAI Mission

  • Launched: In 2024 with budget outlay of ₹10,371.92 crore over five years.
  • Target: Initial target of 10,000 GPUs, India has achieved 38,000 GPUs. 
  • Ministry: Ministry of Electronics and Information Technology (MEITY)
  • Key Pillars: IndiaAI Compute Capacity, IndiaAI Application Development Initiative, IndiaAI Datasets Platform (AIKosh), IndiaAI FutureSkills, IndiaAI Startup Financing, and Safe & Trusted AI.

Models Launched

  • Two Large Language Models (LLMs): By Sarvam AI that could outperform global systems like DeepSeek R1 and  Gemini Flash on several benchmarks. 
    • LLMs are advanced AI systems that can generate natural language, or human-like text, using trained data through machine learning techniques.
  • Text to Speech Model, Vachana: By  Gnani.ai for cloning human voices across 12 Indian languages. 
  • Param2 17B is a Mixture-of-Experts (MoE): Multilingual foundational model by BharatGen. 
    • BharatGen is first government supported national initiative to develop sovereign foundational AI models tailored to Indian languages and societal contexts. 
Watch Video News Today

Explore Related Content

Discover more articles, videos, and terms related to this topic

RELATED TERMS

3

Foundational AI models

Large-scale artificial intelligence models that are trained on vast amounts of data and can be adapted for a wide range of downstream AI tasks. They serve as a base for developing more specific AI applications.

Mixture-of-Experts (MoE)

A type of neural network architecture where multiple specialized 'expert' models are combined to process different parts of a problem. This approach can lead to more efficient and performant AI models, especially for multilingual tasks.

Large Language Models (LLMs)

Advanced AI models trained on massive amounts of text data, capable of understanding, generating, and manipulating human language for various tasks like translation, summarization, and content creation.

Title is required. Maximum 500 characters.

Search Notes

Filter Notes

Loading your notes...
Searching your notes...
Loading more notes...
You've reached the end of your notes

No notes yet

Create your first note to get started.

No notes found

Try adjusting your search criteria or clear the search.

Saving...
Saved

Please select a subject.

Referenced Articles

linked

No references added yet