India is rapidly strengthening its artificial intelligence ecosystem with the launch of AIKosh, a sovereign AI repository and innovation platform developed under the IndiaAI Mission. Designed to provide developers, startups, researchers, and institutions with access to high-quality datasets, AI models, and computing infrastructure, AIKosh aims to solve one of India’s biggest AI challenges — the lack of localized, secure, and India-centric training data.
Developed by the Ministry of Electronics and Information Technology (MeitY), AIKosh functions as a centralized AI infrastructure platform supporting India’s digital transformation and AI innovation goals.
What is AIKosh?
AIKosh, formally known as AIKosha: IndiaAI Datasets Platform, serves as a sovereign AI repository designed specifically for India’s artificial intelligence ecosystem. The platform operates as a centralized hub for:
- AI-ready datasets
- Pre-trained AI models
- Development tools
- Compute infrastructure
- Research and innovation resources
The initiative is part of the ₹10,300+ crore IndiaAI Mission launched by the Government of India.
Industry experts describe AIKosh as India’s answer to global AI model repositories and data-sharing platforms, but with a stronger focus on Indian languages, local datasets, data privacy, and sovereign AI infrastructure.
AIKosh Hosts Thousands of AI Datasets and Models
According to the platform details, AIKosh currently supports:
- 11,800+ datasets
- 300+ pre-trained models
- 25,000+ registered users
- 1.9 crore+ visitor impressions
The platform aggregates data from over 470 verified organizations across sectors including:
- Healthcare
- Agriculture
- Education
- Governance
- Climate change
- Public services
The goal is to make AI development more accessible for Indian startups, enterprises, researchers, and independent developers.
AI Sandbox and Compute Infrastructure
One of AIKosh’s most important features is its integrated AI Sandbox environment powered by India’s AIRAWAT supercomputing infrastructure.
The platform provides multiple compute access options:
CPU Access
Developers receive free access to lightweight computing environments with:
- 3.2 processing cores
- 7GB RAM
This allows students, developers, and researchers to experiment with machine learning models without expensive hardware investments.
GPU Access
AIKosh also offers GPU-powered development access through NVIDIA A100 infrastructure.
The platform includes:
- Free 4-hour GPU sessions for moderate AI experiments
- Advanced 22-hour GPU access for large-scale training workloads (approval-based)
Industry analysts believe this significantly lowers entry barriers for AI startups and researchers in India.
Focus on Indian Languages and Sovereign AI Models
A major objective of AIKosh is to reduce India’s dependence on foreign datasets and Western AI models that often fail to represent India’s linguistic and cultural diversity.
The platform hosts multiple indigenous AI models developed by Indian startups and institutions, including:
- Sarvam AI
- BharatGen
- Soket AI
- Gnani AI
Key India-Centric AI Assets
Some of the notable AI assets available on the platform include:
Sarvam-30B
A large conversational AI model optimized for Indian regional languages and multilingual deployment.
IndicVoices
A multilingual speech dataset featuring:
- 12,000 hours of speech data
- 22 Indian languages
- Coverage across 208 districts
The dataset was developed through collaborations involving IIT Madras, AI4Bharat, and Sarvam AI.
BhasaAnuvaad
One of the world’s largest speech translation datasets containing:
- 44,400 hours of audio data
- 13 Indian languages
Experts believe such datasets are critical for building India-focused AI assistants, language models, voice systems, and digital public services.
AIKosh Uses Tiered User Access System
To ensure data protection and governance, AIKosh uses a structured access management framework supported by National Single Sign-On systems such as:
- MeriPehchaan
- Parichay
The platform defines three major user categories:
Explorers
General users who can browse public datasets and download open-access resources.
Contributors
Organizations and users who can upload and manage AI datasets and models.
Organization Administrators
Institutional managers responsible for permissions, moderation, and access governance.
Strong Focus on Data Privacy and Ethical AI
AIKosh includes strict safeguards to maintain:
- Data privacy
- Anonymization standards
- Intellectual property protection
- Ethical AI compliance
Datasets are classified under:
- Open access
- Restricted access
- Private access
The platform also uses automated onboarding protocols to identify:
- Systemic biases
- Privacy risks
- Sensitive personal information
Industry observers note that ethical AI and sovereign data governance are becoming increasingly important as countries build domestic AI ecosystems.
Why AIKosh Matters for India’s AI Future
Experts believe AIKosh could play a major role in accelerating India’s AI innovation ecosystem by:
- Lowering infrastructure costs
- Providing affordable compute access
- Improving data availability
- Supporting Indian language AI development
- Enabling startups to scale faster
The platform also aligns with India’s broader ambitions around:
- Digital public infrastructure
- AI sovereignty
- Deep-tech innovation
- Semiconductor and compute ecosystem growth
By democratizing access to AI infrastructure and datasets, AIKosh could help create a stronger pipeline of India-focused AI applications across sectors such as healthcare, education, agriculture, governance, and financial services.
Conclusion
AIKosh represents a significant step in India’s efforts to build sovereign AI infrastructure and strengthen its domestic artificial intelligence ecosystem.
As AI adoption accelerates globally, platforms offering localized datasets, affordable compute access, and ethical AI governance are expected to become increasingly important for countries seeking technological independence and innovation leadership.
With growing support for indigenous AI models and India-centric datasets, AIKosh could emerge as a foundational pillar of India’s long-term AI and digital economy strategy.
Investors Double Down on Asian Startups in AI, Solar and Enterprise Tech
Ruchi Kumar is the associate editor at Entrepreneur News Network and TVW News India, where she leads editorial strategy, brand storytelling, and startup ecosystem coverage. With a strong focus on innovation, business, and marketing insights, he curates impactful narratives that spotlight India’s evolving entrepreneurial landscape. She has written extensively on fintech, AI and emerging startups.